Tailoring d–p orbital hybridization of single-atom Cu-N2S sites for enhanced photo-Fenton reaction Authors: Minjia Yan, Xiaoxin Shao, Yu Li, Zhixing Wang, Xi-Lin Wu, Linguo Lu, Guo-Ping Sheng Published: 2026 ABSTRACT Although single-atom catalysts (SACs) are emerging as advanced heterogeneous catalysts for Fenton-like reactions, enhancing their performance by precisely tailoring d–p orbital hybridization remains challenging. Herein, single Cu atoms with a unique Cu-N2S asymmetric coordination structure are designed and fabricated to catalyze photo-Fenton reactions. Density functional theory (DFT) calculation reveals that strong d–p orbital hybridization elevates the HOMO energy level and promotes splitting of d-orbital energy levels, thereby regulating the adsorption affinity of single-atom Cu sites toward H2O2 via a unique anti-d-band-center principle. Consequently, the Cu-N2S single sites exhibit unique photo-switching behavior, transforming the inactive sites into Fenton-active ones under light irradiation, thereby enabling sustained and enhanced generation of hydroxyl radicals for efficient degradation of various organic micropollutants. The developed photo-Fenton system achieves 98.9% removal of sulfamethoxazole, along with broad pH applicability (pH 3-11), high tolerance for complex water matrices, and exceptional durability for long-term operation. This work highlights the critical role of single-atom coordination symmetry in modulating electronic orbital structures, providing an avenue for the rational design of advanced Fenton-like catalysts. Introduction Single-atom catalysts (SACs) have emerged as promising advanced heterogeneous catalysts for a variety of applications1,2. The catalytic activity and selectivity of single-atom metal sites anchored on a support can be precisely regulated through tunable metal-support interactions and well-defined coordination structures3,4. There has been intensive research on the design and synthesis of SACs for advanced oxidation processes (AOPs), making it a highly active area in environmental science. To date, SACs have been employed to activate a range of oxidants, including hydrogen peroxide (H2O2)5, persulfate (HS2O8–)6, peroxymonosulfate (PMS, HSO5–)7, periodate (IO4–)8, and peroxyacetic acid (CH3COOOH)9. These SAC-mediated AOP systems generally involve multiple distinct reaction pathways, leading to the production of diverse radical and non-radical reactive species. For instance, single-atom sites such as MnN410, FeN411, Co‒N212, Ni‒N413, Cu‒O314, are capable of generating multiple reactive species, including sulfate radical (SO4•–), hydroxyl radical (•OH), singlet oxygen (1O2), and high-valent metal-oxo species. Despite these considerable advances, the fundamental trade-off between activity and selectivity in radical production remains a critical challenge for SAC-mediated AOPs. While SAC design has primarily focused on coordination chemistry and metal-support interactions, the strategic control of atomic-level electronic structures via orbital hybridization remains a critically underexplored avenue to unlock unprecedented catalytic performance. Heterogeneous Fenton and Fenton-like catalysis are driven by the Mn+1/Mn+ redox cycle for peroxide activation, while this process is still limited by the sluggish electron-transfer. By leveraging light irradiation to enhance charge transfer and accelerate the Mn+1/Mn+ redox cycle, photo-Fenton reactions thereby efficiently produce reactive radical species for the oxidation of organic pollutants5,15. Thus far, the study of SAC-driven photo-Fenton reactions remains in its infancy. Current SACs design for photo-Fenton reactions has focused on coordination regulation, defect engineering, morphology regulation, heteroatom doping, and modulation of metal-support interactions to enhance the production of reactive species5,16–18. Engineering the electronic structures of single-atom metal sites offers a promising route to control the catalytic activity and selectivity of SACs19. By engineering the d-orbital electronic structure, it is possible to adjust the d-band center, thus promoting electron transfer and optimizing reactant adsorption affinity in Fenton-like systems20,21. However, the precise regulation of the d-orbital electronic structure via d–p orbital hybridization remains unexplored for the design of high-performance SACs in photo-Fenton reactions. In this study, we address this critical gap by constructing a unique asymmetric Cu-N2S coordination architecture that enables precise manipulation of the d-orbital electronic structure in single-atom Cu sites. This asymmetric configuration induces significant d–p orbital hybridization and results in dramatically enhanced intrinsic catalytic activity compared to conventional symmetric Cu-N2 sites. Notably, we reveal an unconventional anti-d-band-center phenomenon in the Cu-N2S sites, which fundamentally promotes H2O2 activation and accelerates reaction kinetics of the photo-Fenton system. The superior performance of this photo-Fenton system is validated through efficient degradation of various refractory organic micropollutants in complex water matrices. Furthermore, using femtosecond transient absorption (fs-TA) spectroscopy, we reveal accelerated charge-transfer dynamics enabled by this electronic-structure modulation. This work pioneers a revolutionary approach for designing high-performance SACs through d–p orbital hybridization, establishing a paradigm for electronic-level precision engineering in environmental catalysis. Results Characterizations of the catalysts The synthesis begins with the self-assembly of trithiocyanuric acid and melamine in aqueous solution, followed by the adsorption of Cu²⁺ onto the precursor, and concludes with a final calcination step (Fig. 1a). Scanning electron microscopy (SEM) and transmission electron microscopy (TEM) images reveal a two-dimensional sheet-like morphology of SA-Cu-SCN with abundant surface folding (Supplementary Fig. 1). The high-resolution TEM (HRTEM) image shows the graphitic structure of SA-Cu-SCN (Fig. 1b), with no visible Cu nanoparticles. This observation is consistent with the X-ray diffraction patterns (Supplementary Fig. 2)22,23. The high-angle annular dark-field scanning TEM (HAADF-STEM) image confirms the presence of atomically dispersed Cu, which appears as uniformly distributed bright spots (circled) on the SCN substrate (Fig. 1c).Fig. 1Characterizations of the SA-Cu-SCN.a Synthesis scheme of the SA-Cu-SCN catalyst. b HRTEM image and c HAADF-STEM image of SA-Cu-SCN (Yellow circles highlight the bright spots). d S 2p XPS spectra of the SCN and SA-Cu-SCN. e XANES spectra of the SA-Cu-SCN and reference samples. f Correlation between valence states and k-edge energies of Cu in the various samples. g FT-EXAFS spectra of the various samples. (h) EXAFS fitting curves of SA-Cu-SCN at R-space. i Wavelet transform of the EXAFS spectra. Source data for Fig. 1 are provided in the Source Data file. Characterizations of the SA-Cu-SCN. a Synthesis scheme of the SA-Cu-SCN catalyst. b HRTEM image and c HAADF-STEM image of SA-Cu-SCN (Yellow circles highlight the bright spots). d S 2p XPS spectra of the SCN and SA-Cu-SCN. e XANES spectra of the SA-Cu-SCN and reference samples. f Correlation between valence states and k-edge energies of Cu in the various samples. g FT-EXAFS spectra of the various samples. (h) EXAFS fitting curves of SA-Cu-SCN at R-space. i Wavelet transform of the EXAFS spectra. Source data for Fig. 1 are provided in the Source Data file. The chemical states and coordination environment of the Cu sites in SA-Cu-SCN were analyzed by X-ray photoelectron spectroscopy (XPS). Notably, compared to pristine SCN, the S 2p XPS spectrum of SA-Cu-SCN exhibits a new peak at about 160.6 eV (Fig. 1d), attributable to the Cu–S bond. This finding provides conclusive evidence for the successful construction of the S-Cu coordination on SCN substrate24,25. The strong d–p hybridization of the Cu–S bond establishes additional electron-transfer channels for heterogeneous catalysis. The Cu K-edge XANES spectrum of SA-Cu-SCN differs significantly from those of the reference samples (Fig. 1e). The first derivative of the absorption edge was used to quantitatively estimate the average valence state of Cu in SA-Cu-SCN (Fig. 1f), yielding a value of approximately +1.326. This result is fully consistent with the predominant Cu(I) species identified by XPS analysis. The Fourier-transformed EXAFS spectrum exhibits a primary peak at ∼1.4 Å (Fig. 1g), attributed to the first-shell Cu-N/S coordination. The absence of Cu-Cu bonds (∼2.2 Å) precludes the existence of metallic Cu clusters or particles in SA-Cu-SCN27. To resolve the specific coordination structure, EXAFS fitting was performed in R-space (Fig. 1h). The corresponding fitting parameters are listed in Supplementary Table 1. The optimal EXAFS fitting path for SA-Cu-SCN corresponds to a Cu-N2S coordination structure, yielding coordination numbers of 2.0 for Cu-N at 2.1 Å and 1.0 for Cu-S at 2.3 Å. This configuration is further corroborated by wavelet-transform (WT) EXAFS analysis (Fig. 1i). For comparison, parallel characterizations were conducted to identify the single-atom Cu–N2 site in SA-Cu-CN (Supplementary Fig. 3). We analyzed the textural properties of the SA-Cu-SCN and SA-Cu-CN catalysts using N2 adsorption-desorption isotherms (Supplementary Fig. 4) to determine the Brunauer–Emmett–Teller (BET) surface areas and pore size distributions. Additionally, the Cu content in both catalysts was quantified via inductively coupled plasma optical emission spectrometry (ICP-OES). As summarized in Supplementary Table 2, the two catalysts show comparable textural properties and Cu loadings, with nearly identical specific surface areas, pore volumes, and pore size distributions. Photo-fenton performance of the SA-Cu-SCN Sulfamethoxazole (SMX), a common antibiotic, was chosen as a representative organic pollutant for evaluating the performance of the catalysts. Among the various transition metal (TM = Mn, Fe, Co, Ni, Cu) catalysts evaluated (Supplementary Fig. 5), SA-Cu-SCN exhibited superior catalytic activity in the photo-Fenton-like reaction. An optimum metal loading of 20% was identified for SA-Cu-SCN, exhibiting the best performance for the photo-Fenton reactions (Supplementary Fig. 6). To ensure the efficient degradation and mineralization of pollutants, the catalyst dosage and H2O2 concentration were optimized to be 0.2 g/L and 10 mM, respectively (Supplementary Fig. 7). Compared to the control groups, the SA-Cu-SCN + H2O2 + Vis system exhibits superior performance for SMX degradation, achieving enhanced removal efficiency and faster reaction kinetics (Fig. 2a). More specifically, the SA-Cu-SCN + H2O2 + Vis system achieved 98.9% SMX degradation, surpassing that of the SA-Cu-SCN + Vis (38.7%), SA-Cu-SCN + H2O2 (30.6%), and SA-Cu-CN + H2O2 + Vis (55.2%) systems. This indicates modest catalytic activity of Cu-N2 single sites, underscoring the critical role of Cu-S coordination in H2O2 activation. The comparably poor performances of the H2O2 + Vis, SA-Cu-SCN + Vis, and SA-Cu-SCN + H2O2 systems indicate the inactivity of SA-Cu-SCN in the dark. This finding reveals a photo-switching effect, where light irradiation transforms the inactive Cu-N2S sites of SA-Cu-SCN into Fenton-active ones (Fig. 2b).Fig. 2Performances of the SA-Cu-SCN for photo-Fenton reactions.a Degradation of SMX in the various systems. b Schematic illustration of the photo-switching Fenton-like catalysis on SA-Cu-SCN. c Effect of solution pH on SMX degradation in the SA-Cu-SCN + H2O2 + Vis system. d Normalized K-values for pollutant degradation with state-of-the-art photo-Fenton-like catalysts (Detailed comparisons and relevant references are listed in Supplementary Table 5). e Quenching experiments for SMX degradation. f DMPO spin-trapping EPR spectra for •OH. g Quantitative determination of •OH production in the various systems. (Reaction conditions: [SMX] = [NOR] = [CIP] = [LEV] = [TCH] = 10 mg L–1, [H2O2] = 10 mM, [Catalyst] = 0.2 g L–1, and pH = 7.0). Data are presented as mean ± s.d. (standard deviation) from n = 3 independent experiments. The error bars are defined as the standard deviation, and the center of each error bar represents the mean value of the corresponding three independent experiments. Source data for Fig. 2 are provided in the Source Data file. Performances of the SA-Cu-SCN for photo-Fenton reactions. a Degradation of SMX in the various systems. b Schematic illustration of the photo-switching Fenton-like catalysis on SA-Cu-SCN. c Effect of solution pH on SMX degradation in the SA-Cu-SCN + H2O2 + Vis system. d Normalized K-values for pollutant degradation with state-of-the-art photo-Fenton-like catalysts (Detailed comparisons and relevant references are listed in Supplementary Table 5). e Quenching experiments for SMX degradation. f DMPO spin-trapping EPR spectra for •OH. g Quantitative determination of •OH production in the various systems. (Reaction conditions: [SMX] = [NOR] = [CIP] = [LEV] = [TCH] = 10 mg L–1, [H2O2] = 10 mM, [Catalyst] = 0.2 g L–1, and pH = 7.0). Data are presented as mean ± s.d. (standard deviation) from n = 3 independent experiments. The error bars are defined as the standard deviation, and the center of each error bar represents the mean value of the corresponding three independent experiments. Source data for Fig. 2 are provided in the Source Data file. The degradation kinetics were further analyzed using a pseudo-first-order model (Supplementary Figs. 8 and Supplementary Table 3). As expected, the observed rate constant (kobs) for the SA-Cu-SCN + H2O2 + Vis system reached 0.11 min‒1, exceeding those of the other systems by approximately 13 to 29 times. These findings further corroborate that the Cu–S bond effectively bridges photocatalysis and Fenton-like catalysis, thereby enhancing the photo-Fenton reaction for pollutants degradation. Unlike the conventional Fenton system, which is only effective under acidic conditions, the SA-Cu-SCN + H2O2 + Vis system performed effectively over a broad pH range (3–11) while maintaining high SMX removal efficiencies ( > 90%) (Fig. 2c and Supplementary Fig. 9). The leaching of Cu ions from the SA-Cu-SCN + H2O2 + Vis system was negligible, accounting for less than 2.0% at pH 3 and less than 0.1% in the pH range of 5–11 (Supplementary Table 4), demonstrating the excellent stability of the catalyst. Furthermore, the SA-Cu-SCN + H2O2 + Vis system also effectively eliminated various antibiotic micropollutants, including sulfisoxazole (SIZ, 97.6%), norfloxacin (NOR, 97.4%), ciprofloxacin (CIP, 89.9%), levofloxacin (LEV, 99.9%) and tetracycline hydrochloride (TCH, 100%) (Supplementary Fig. 10). To benchmark against state-of-the-art photo-Fenton systems, a normalized kinetic rate (K-value = (kobs×Cpollut.)/(CH2O2×Ccat.) was introduced to evaluate catalytic activity28. Based on the reaction conditions and calculated K-values (Supplementary Table 5 and Fig. 2d), SA-Cu-SCN exhibits a significantly higher K-value for pollutant degradation, confirming its superior intrinsic catalytic activity in Fenton-like reactions. Collectively, these results substantiate the SA-Cu-SCN-mediated photo-Fenton process as an effective strategy for advanced water purification. The reactive oxygen species (ROS) produced in the SA-Cu-SCN + H2O2 + Vis system were identified and quantified. Quenching experiments were performed using tert-butanol (TBA) for •OH, L-histidine (L-his) for 1O2, nitro blue tetrazolium (NBT) for O2•– and methyl phenyl sulfoxide (PMSO) for high valent metal species29–32. As shown in Fig. 2e and Supplementary Fig. 11, the presence of TBA significantly suppressed SMX degradation, revealing the dominant role of radical species (•OH) in the photo-Fenton system. Scavenging experiments with L-his and NBT quenched SMX degradation by approximately 4.7% and 20.9%, respectively, indicating that 1O2 and O2•– are not the primary ROS responsible for pollutant oxidation. To further evaluate the involvement of 1O2, degradation experiments were conducted in deuterated water (D2O), a solvent known to extend the lifetime of 1O2. As shown in Supplementary Fig. 12, the kinetic profile of SMX degradation in D2O was nearly identical to that in H2O. This observation confirms the negligible contribution of 1O2 to the photo-Fenton-like process. Likewise, the negligible inhibition observed with PMSO suggests that high-valent copper species are not involved in the photo-Fenton system. The ROSs were further identified by electron paramagnetic resonance (EPR) spectroscopy using 5,5-dimethyl-1-pyrroline N-oxide (DMPO) and 2,2,6,6-tetramethyl-4-piperidinol (TEMP) as spin-trapping agents. As shown in Fig. 2f, the SA-Cu-SCN + H2O2 + Vis system generated significantly stronger DMPO-•OH signals than the SCN + H2O2 + Vis, SA-Cu-CN + H2O2 + Vis and SA-Cu-SCN + H2O2 systems. Besides, much stronger DMPO-O2•– and TEMP-1O2 signals were detected in the SA-Cu-SCN + H2O2 + Vis system than in the above control systems (Supplementary Figs. 13 and 14). These findings demonstrate that the Cu-N2S coordination structure is conducive to the generation of the various ROSs in the photo-Fenton reaction, thereby boosting the performance of SA-Cu-SCN for pollutant oxidation. Quantitative analysis of ROS concentrations in the various systems was conducted (Fig. 2g and Supplementary Fig. 15). The results confirm that •OH is the predominant reactive oxygen species (ROS) generated in the SA-Cu-SCN + H2O2 + Vis system, with its concentration far surpassing that in the control systems. This further corroborates the photo-switching phenomenon of the SA-Cu-SCN catalyst, which transforms inactive Cu-N2S sites into Fenton-active ones under light irradiation, thereby boosting •OH generation. Compared to state-of-the-art catalysts (Supplementary Table 6), SA-Cu-SCN shows much higher efficiency for H2O2 to •OH conversion. Consequently, the developed photo-Fenton system mitigates H2O2 consumption while augmenting radical generation to facilitate pollutant degradation and mineralization, thereby underscoring its potential for industrial wastewater treatment. Enhanced electron transfer process To understand the superior activity of SA-Cu-SCN, we systematically investigated its optical properties and charge carrier dynamics33. As evidenced by UV-Vis diffuse reflectance spectra (DRS), SA-Cu-SCN exhibits a significant enhancement in light absorption compared to pristine SCN (Fig. 3a). This is evidenced by the color change from light yellow for SCN to brown for SA-Cu-SCN. Tauc plot analysis indicates a substantial bandgap narrowing from 2.04 eV for SCN to 1.64 eV for SA-Cu-SCN (Supplementary Fig. 16), thereby facilitating more efficient production of photoinduced charge carriers. The implanted SA-Cu dopants create impurity energy levels between the valence band (VB) and conduction band (CB) of SCN, facilitating the energy transitions for the generation of photoexcited electrons. Photoluminescence (PL) spectra (Supplementary Fig. 17) demonstrate substantially quenched emission for SA-Cu-SCN relative to SCN, indicating suppressed recombination of photogenerated charge carriers. This suggests that the incorporated Cu sites create effective trapping states that promote charge separation via strong d–p orbital coupling. Mott-Schottky (M–S) measurements were performed to assess the semiconductor properties. Both SCN and SA-Cu-SCN exhibit positive slopes in their M–S plots (Supplementary Fig. 18), confirming their n-type character. The charge carrier density (ND) can be quantitatively described by the following equation34:1\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${{\mbox{N}}}_{{\mbox{D}}}=\frac{2}{{\mbox{q}}\upvarepsilon {\upvarepsilon }_{0}}\frac{1}{{\mbox{slope}}}$$\end{document}ND=2qεε01slopewhere q, ε and ε0 are fundamental physical constants. Consequently, ND is inversely proportional to the slope of the M–S plot. The notably smaller slope for SA-Cu-SCN indicates a substantially higher charge carrier density compared to SCN, attributed to enhanced electron transfer driven by d–p orbital hybridization. In situ XPS was performed to monitor the migration of photoexcited electrons in photocatalysts (Fig. 3b, c and Supplementary Fig. 19). Under light irradiation, the Cu 2p XPS peaks shifted to lower binding energies, whereas the S 2p, C 1s, and N 1s XPS peaks all shifted to higher energies. These results visually capture the directional charge transfer from the SCN support to the SA-Cu sites via the Cu–S electron-transfer channels35, which in turn accelerates the Cu(I)/Cu(II) cycle for the photo-Fenton reaction.Fig. 3Optical properties and charge-transfer dynamics of the SA-Cu-SCN.a UV-Vis diffuse reflectance spectra (insets show the digital pictures) of the samples. b S 2p XPS spectra, and c Cu 2p XPS spectra of the SA-Cu-SCN (red and blue arrows indicate the peak-shift directions). d TA spectra of SCN and e the corresponding TA kinetics at 451 nm. f TA spectra of SA-Cu-SCN and g the corresponding TA kinetics at 451 nm. Schematic depictions of the electron transfer process in h SCN and i SA-Cu-SCN. j The proposed charge transfer process for photo-Fenton catalysis. Source data for Fig. 3 are provided in the Source Data file. Optical properties and charge-transfer dynamics of the SA-Cu-SCN. a UV-Vis diffuse reflectance spectra (insets show the digital pictures) of the samples. b S 2p XPS spectra, and c Cu 2p XPS spectra of the SA-Cu-SCN (red and blue arrows indicate the peak-shift directions). d TA spectra of SCN and e the corresponding TA kinetics at 451 nm. f TA spectra of SA-Cu-SCN and g the corresponding TA kinetics at 451 nm. Schematic depictions of the electron transfer process in h SCN and i SA-Cu-SCN. j The proposed charge transfer process for photo-Fenton catalysis. Source data for Fig. 3 are provided in the Source Data file. As shown in Supplementary Fig. 20, SA-Cu-SCN exhibits a significantly stronger photocurrent response than pristine SCN, suggesting the improved charge separation and transport. This finding is consistent with the electrochemical impedance spectra (EIS) measurements showing that SA-Cu-SCN possesses a much smaller arc radius than SCN (Supplementary Fig. 21), which is indicative of the lower charge transfer resistance. Such facilitated charge transport can be attributed to the optimized d–p orbital hybridization in SA-Cu-SCN, which promotes electron delocalization and interfacial transfer via the Cu–S electron-transfer channels. To specifically probe the electron transfer between the catalysts and H2O2, we monitored the open circuit potential (OCP) and amperometric current–time (i–t) curves in the photo-Fenton systems. The introduction of H2O2 induced a pronounced positive OCP shift in the SA-Cu-SCN + Vis system (Supplementary Fig. 22), which reflects an elevated surface electron density and enhanced electron-donating capacity. Concurrently, a rapid current response was observed in the i–t curve upon H2O2 addition (Supplementary Fig. 23), directly evidencing efficient electron transfer from the SA-Cu-SCN catalyst to H2O2. These results collectively demonstrate that the strong d–p orbital hybridization in SA-Cu-SCN drives efficient interfacial charge transfer, thereby enhancing H2O2 activation in the photo-Fenton reaction. The charge carrier dynamics of SA-Cu-SCN and SCN were further probed using femtosecond transient absorption (fs-TA) spectroscopy under 330 nm excitation. The fs-TA spectra of SCN display a broad negative absorption band between 425–670 nm (Fig. 3d), characteristic of ground-state bleach and stimulated emission. In contrast, SA-Cu-SCN shows a distinct spectral profile with negative absorption in the range of 425–600 nm and a positive absorption band above 600 nm (Fig. 3f), the latter attributable to hole trapping, a process conducive to charge separation36,37. The overall higher fs-TA intensity of SA-Cu-SCN compared to SCN also implies a larger population of photogenerated charge carriers. Kinetic profiles at 451 nm were fitted with a tri-exponential function (Fig. 3e, g). For SCN, the fitted lifetimes are τ1 = 2 ps, τ2 = 500 ps, τ3 = 4359 ps, which were assigned to electron trapping in shallow surface state, deep trapping state, and eventual electron-hole recombination, respectively38. In contrast, SA-Cu-SCN exhibited an improved charge-trapping behavior, with lifetimes of τ1 = 1 ps, τ2 = 357 ps and τ3 = 7080 ps. The corresponding electron-trapping pathways are depicted in Fig. 3h, i. The ultrafast τ1 component (1 ps) in SA-Cu-SCN is attributed to efficient electron transfer to the single-atom Cu trapping sites, while τ2 (357 ps) represents subsequent trapping into deeper states. The significantly extended τ3 (7080 ps) reflects suppressed charge recombination in SA-Cu-SCN, facilitating more efficient utilization of photogenerated carriers for surface redox reactions. The optimized charge trapping and transport accelerate the Cu(I)/Cu(II) redox cycle by the photo-induced hot electrons, thereby endowing SA-Cu-SCN with enhanced Fenton-like activity (Fig. 3j). This finding aligns with the results from the in situ irradiation XPS measurements, corroborating the accelerated and directional charge-transfer process in the SA-Cu-SCN photocatalyst. Consequently, the photogenerated electrons effectively transfer from the SCN substrate to the single-atom Cu sites via the Cu–S electron-transfer channels, thereby boosting H2O2 activation and •OH production in the photo-Fenton system. The synergistic combination of photocatalysis and heterogeneous Fenton-like reaction provides a cost-effective approach for efficient degradation of organic pollutants. This work elucidates the fundamental mechanism by which d–p orbital hybridization establishes additional electron-transfer pathways, thereby activating dormant catalysts in Fenton reactions and accelerating surface redox cycles for environmental applications. Photo-fenton system for practical water purification The influences of various water matrices and common anions on SMX degradation in the SA-Cu-SCN + H2O2 + Vis system were investigated. As shown in Fig. 4a and Supplementary Fig. 24, the photo-Fenon system effectively removed SMX ( > 90%) from various real waters, including tap water, river water, lake water, piggery effluent and pharmaceutical effluent. Given the coexistence of inorganic anions and natural organic matter in actual wastewater, their impacts on pollutant degradation were examined by introducing Cl−, NO3−, HCO3−, SO42− and humic acid (HA) into the reaction system. The photo-Fenton system effectively eliminated SMX in the presence of various competing anions (Fig. 4b and Supplementary Fig. 25), confirming its excellent tolerance to complex water matrices. In addition, the SA-Cu-SCN maintains 90% of SMX removal over ten successive reaction cycles (Fig. 4c), underscoring its excellent stability and reusability.Fig. 4Performance of the SA-Cu-SCN + H2O2 + Vis for real wastewater treatment.Effects of the a various water matrix, b anions and c cyclic performance of the SA-Cu-SCN + H2O2 + Vis system for SMX degradation. d Digital picture and schematic diagram of the continuous-flow reactor (Size of the flow-device panel: 25 cm × 20 cm × 3.6 cm). e SEM images of the melamine sponge before and after loading of the SA-Cu-SCN catalyst. f The changes in solar radiation intensity and degradation of SMX in the SA-Cu-SCN + H2O2 + Vis system under sunlight using the continuous-flow reactor. g The proposed degradation pathways of SMX (Reaction conditions: [SMX] = 10 mg L–1, [H2O2] = 10 mM, [Catalyst] = 0.2 g L–1, and pH = 7.0). Data are presented as mean ± s.d. (standard deviation) from n = 3 independent experiments.The error bars are defined as the standard deviation, and the center of each error bar represents the mean value of the corresponding three independent experiments. Source data for Fig. 4 are provided in the Source Data file. Performance of the SA-Cu-SCN + H2O2 + Vis for real wastewater treatment. Effects of the a various water matrix, b anions and c cyclic performance of the SA-Cu-SCN + H2O2 + Vis system for SMX degradation. d Digital picture and schematic diagram of the continuous-flow reactor (Size of the flow-device panel: 25 cm × 20 cm × 3.6 cm). e SEM images of the melamine sponge before and after loading of the SA-Cu-SCN catalyst. f The changes in solar radiation intensity and degradation of SMX in the SA-Cu-SCN + H2O2 + Vis system under sunlight using the continuous-flow reactor. g The proposed degradation pathways of SMX (Reaction conditions: [SMX] = 10 mg L–1, [H2O2] = 10 mM, [Catalyst] = 0.2 g L–1, and pH = 7.0). Data are presented as mean ± s.d. (standard deviation) from n = 3 independent experiments.The error bars are defined as the standard deviation, and the center of each error bar represents the mean value of the corresponding three independent experiments. Source data for Fig. 4 are provided in the Source Data file. A pilot-scale test successfully demonstrated the practical application of the SA-Cu-SCN + H2O2 + Vis system. Using a continuous-flow microreactor with SA-Cu-SCN on a melamine sponge support (Fig. 4d, e), the system treated 10 L of SMX wastewater under natural sunlight, maintaining high removal efficiency over 12 h continuous operation (Fig. 4f). Furthermore, we investigated the applicability of this photo-Fenton system for indoor and night-time scenarios, substituting natural light with artificial illumination. In a controlled indoor setup under low-power LED irradiation, the system achieved a high SMX removal efficiency of 90.6% within 90 min (Supplementary Fig. 26). Additionally, a continuous-flow pilot test conducted outdoors at night, using LED arrays as the sole light source (Supplementary Fig. 27), yielded a SMX removal rate of approximately 80% (Supplementary Fig. 28). These results demonstrate that the photo-Fenton process can be effectively driven by energy-efficient LED sources, ensuring reliable and continuous operation under all-weather conditions. This underscores the promise of this photo-Fenton system as a viable technology for practical water treatment. The total organic carbon (TOC) analysis demonstrates a high removal efficiency of 71.6% for SMX degradation within the SA-Cu-SCN + H2O2 + Vis system (Supplementary Fig. 29). This indicates that a significant portion of the SMX molecules was effectively mineralized into inorganic end products, such as CO2 and H2O. Furthermore, the intermediates formed during SMX degradation were identified using liquid chromatography-mass spectrometry (LC-MS) (Supplementary Fig. 30 and Fig. 4g). Based on the LC-MS data, plausible degradation pathways are proposed, revealing the gradual decomposition and mineralization of SMX through hydroxylation and ring-opening processes. To evaluate the environmental impacts and sustainability, the SA-Cu-SCN + H2O2 + Vis process was benchmarked against the SA-Cu-CN + H2O2 + Vis system using life cycle assessment (LCA) (Supplementary Fig. 31). This assessment encompasses the entire treatment chain, from catalyst synthesis to pollutant degradation, with 18 relevant impact descriptors at the laboratory scale (Supplementary Tables 7‒8)39,40. As depicted in Supplementary Figs. 32‒34, the SA-Cu-SCN + H2O2 + Vis system exhibited a markedly lower environmental impact across most indicators, due to its superior intrinsic activity, which reduces metal and H2O2 consumption, energy demand, and emissions throughout the life cycle. Theoretical analysis of d–p orbital hybridization Conventional d-band center theory provides only an approximate measure of adsorption strength and may fail to capture the orbital-specific nature of catalytic interactions41–43. A more precise and effective strategy is to engineer the hybridization between specific metal d-orbitals and p-orbitals of the heteroatoms44,45, thereby reconfiguring the local electronic environment of the active site to improve its interaction with reactants. Here, we employ density functional theory (DFT) calculations to unravel how S-coordination replacement modulates the H2O2 activation on a single-atom Cu-N2S site compared to a Cu-N2 site. The configurations of Cu-N2S site supported on sulfur-rich carbon nitride (SA-Cu-SCN) and Cu-N2 site supported on carbon nitride (SA-Cu-CN) were constructed (Supplementary Fig. 35). As shown in Fig. 5a, the projected density of states (PDOS) of SA-Cu-SCN exhibits a pronounced overlap between Cu 3d and S 3p states near the Fermi level, indicating strong d–p hybridization. Importantly, S coordination both shifts the HOMO of the Cu center to higher energy and induces a specific delocalization of the Cu dxz orbital near the Fermi level (Supplementary Fig. 36). Such orbital reconstruction significantly enhances H2O2 adsorption, leading to the much larger adsorption energy on SA-Cu-SCN than on SA-Cu-CN (Fig. 5b).Fig. 5DFT calculations of the d–p orbital hybridization.a PDOS of SA-Cu-SCN and SA-Cu-SCN. b Adsorption energy. c PDOS of SA-Cu-CN + H2O2 and SA-Cu-CN + H2O2. d Energy level diagram depicting d–p orbital hybridization of SA-Cu-CN and SA-Cu-SCN. e Schematic depiction of d–p orbital hybridization. Bader charge of f H2O2 onto CuN2S1 sites of SA-Cu-SCN and g H2O2 onto CuN2 sites of SA-Cu-CN. Gray, blue, yellow, red, white and orange spheres represent C, N, S, O, H and Cu atoms, respectively. The red and green electron clouds correspond to regions of charge accumulation and depletion, respectively. Red and blue arrows indicate the directions of electron transfer. Source data for Fig. 5 are provided in the Source Data file. DFT calculations of the d–p orbital hybridization. a PDOS of SA-Cu-SCN and SA-Cu-SCN. b Adsorption energy. c PDOS of SA-Cu-CN + H2O2 and SA-Cu-CN + H2O2. d Energy level diagram depicting d–p orbital hybridization of SA-Cu-CN and SA-Cu-SCN. e Schematic depiction of d–p orbital hybridization. Bader charge of f H2O2 onto CuN2S1 sites of SA-Cu-SCN and g H2O2 onto CuN2 sites of SA-Cu-CN. Gray, blue, yellow, red, white and orange spheres represent C, N, S, O, H and Cu atoms, respectively. The red and green electron clouds correspond to regions of charge accumulation and depletion, respectively. Red and blue arrows indicate the directions of electron transfer. Source data for Fig. 5 are provided in the Source Data file. Furthermore, we calculated the PDOS of H2O2 adsorbed onto both Cu-N2S and Cu-N2 sites in SA-Cu-SCN and SA-Cu-CN, respectively. (Fig. 5c). Compared to SA-Cu-CN, SA-Cu-SCN shows more orbital overlap between the O p-orbital of H2O2 and Cu d-orbitals. Notably, this overlap resides above the Cu original orbital energies but below the Fermi level (Supplementary Fig. 37), revealing an anti-d-band-center phenomenon that a more negative d-band center results in stronger adsorption46,47. This finding challenges the conventional d-band center theory and thus highlights d–p hybridization as a powerful strategy for the rational design of SACs. Moreover, orbital coupling analysis was employed to elucidate H2O2 adsorption on the Cu-N2S and Cu-N2 sites (Fig. 5d, e). In SA-Cu-SCN, the strong d–p hybridization induced by S doping significantly enhances the overlap integral and electronic coupling matrix element between the dx- and O pz-orbitals. Consistent with the Newns-Anderson model, this strengthened orbital coupling facilitates charge transfer and stabilizes the adsorption configuration48. Consistently, Bader charge analysis (Fig. 5f, g) confirms enhanced electron transfer from SA-Cu-SCN to H2O2, underscoring the role of S-mediated d–p orbital hybridization in boosting the heterogeneous Fenton reaction. The d–p orbital hybridization provides an electron-transfer pathway that initiates the photo-switching of Cu-N2S sites, significantly promoting •OH generation in the photo-Fenton-like process. Discussion In summary, this work realized precise modulation of the d-orbital electronic structure in an asymmetric Cu-N2S coordination structure on SCN through d–p orbital hybridization. A unique anti-d-band center phenomenon was discovered in the Cu-N2S single-atom sites, responsible for enhancing H2O2 adsorption and activation on SA-Cu-SCN. The SA-Cu-SCN catalyst demonstrated superior catalytic activity in the photo-Fenton system relative to the SA-Cu-CN benchmark. Charge-transfer mechanisms probed by fs-TA reveal that the Cu-S channels accelerated the electron transfer from the SCN substrate to the single-atom Cu sites, thereby accelerating the Cu(I)/Cu(II) redox cycle for the photo-Fenton reaction. The improved d-orbital electronic structure promotes the electron-transfer processes for H2O2 activation to produce the various radical species. The SA-Cu-SCN-mediated photo-Fenton system demonstrated excellent performance in degrading various organic micropollutants, exhibiting broad applicability over a wide pH range and in diverse real water matrices. This work presents a strategy for manipulating the electronic structure of SACs, offering a cost-efficient approach to develop advanced SAC-mediated Fenton-like systems. Methods Chemicals and characterizations Detailed information is provided in Supplementary Text S1–S5. Catalysts synthesis For synthesis of SA-Cu-SCN, a polymer precursor was first prepared by assembling trithiocyanuric acid (0.73 g) and melamine (0.52 g) in aqueous solution. Subsequently, an ethanol solution containing CuCl2·2H2O was added dropwise to 1.25 g of the polymer precursor. The resulting mixture was thoroughly ground and subjected to calcination at 550 °C for 2 h under an N2 atmosphere. The obtained product was then ground again, washed with deionized water to remove impurities, and dried to yield a brown powder, denoted as SA-Cu-SCN. Using the same procedure without the addition of metal precursor, a yellow powder labeled SCN was obtained. For synthesis of SA-Cu-CN, a polymer precursor was synthesized by combining cyanuric acid (0.52 g), melamine (0.32 g), and polyethyleneimine (PEI) in aqueous solution. Following the same protocol described above, a black powder was obtained, designated as SA-Cu-CN. Computational methodology Density functional theory (DFT) computations were performed with the Vienna ab initio simulation package (VASP)49,50. The interactions between core and valence electrons were treated using the projector-augmented wave (PAW) method51. Exchange–correlation effects were described within the generalized gradient approximation (GGA) using the Perdew–Burke–Ernzerhof (PBE) functional52. To enhance modeling accuracy, we applied spin polarization and integrated an effective Hubbard U value of 3.87 for Cu53. For Brillouin zone integration, we applied a Gamma-centered k-points mesh of 6 × 6 × 154. The plane-wave kinetic energy cutoff was set as 550 eV, while the convergence threshold of electronic energy and force on each atom was set as 1 × 10−5 eV and 0.01 eV/Å, respectively. To avoid periodic interactions, the lattice parameter along the c direction was fixed at 20 Å. The dispersion interactions between adsorbates and substrates were incorporated using Grimme’s DFT-D3 scheme with Becke–Johnson damping55,56, which has been widely used in recent catalysis studies57–60. In all calculations, the accurate precision setting was employed to minimize wrap-around errors. The Gibbs free energies of each intermediate were calculated as:2\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${\mbox{G}}={{\mbox{E}}}_{{\mbox{DFT}}}+{\mbox{ZPE}}-{\mbox{TS}}+{\mbox{PV}}+{{\mbox{E}}}_{{\mbox{vib}}}+{{\mbox{E}}}_{{\mbox{tran}}}+{{\mbox{E}}}_{{\mbox{rot}}}$$\end{document}G=EDFT+ZPE−TS+PV+Evib+Etran+Erotwhere EDFT is the energy obtained from DFT calculations, and ZPE denotes the zero-point energy. The PV contribution converts internal energy into enthalpy for free molecules. The TS term was computed at room temperature (298.15 K), and S represents entropy, which converts enthalpy to Gibbs free energy. \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${{\mbox{E}}}_{{\mbox{vib}}}$$\end{document}Evib, \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${{\mbox{E}}}_{{\mbox{tran}}}$$\end{document}Etran, and \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${{\mbox{E}}}_{{\mbox{rot}}}$$\end{document}Erot correspond to the vibrational, translational, and rotational contributions for free molecules; for adsorbed species, only the vibrational term was considered. These thermal contributions (second to seventh terms) were obtained using the VASP toolkit (split)61. The free-energy changes along the reduction pathway were determined within the framework of the computational hydrogen electrode (CHE) model proposed by Nørskov et al.62. Charge transfer between adsorbates and substrates was calculated by Bader charge analysis63–65, and charge-density differences were visualized by the VESTA package66. Experimental procedures A typical reaction mixture was prepared by dispersing an appropriate amount of the catalyst and SMX in 30 mL of aqueous solution. Initial pH was precisely adjusted to target values using 0.5 M HCl or NaOH solutions. The mixture was subjected to magnetic stirring for 60 min to establish adsorption–desorption equilibrium between SMX and the catalyst prior to photocatalytic tests. The catalytic oxidation process was initiated by adding 31 μL of H2O2 to the suspension under simulated visible-light irradiation (300 W xenon lamp with a CEL-UVCUT filter, λ > 420 nm). Samples were withdrawn at designated time intervals, filtered through 0.22 μm PTFE membranes, and quenched with excess methanol prior to analysis. The concentration of SMX was quantified using high-performance liquid chromatography (HPLC, Shimadzu 20AV) equipped with a C18 column and a UV/Vis detector set at 265 nm. The mobile phase consisting of methanol and water (50:50, v/v) was delivered at a flow rate of 1.0 mL/min. Detailed HPLC conditions for the determination of various antibiotics are summarized in Supplementary Table 9. The kinetics of SMX degradation were fitted to a pseudo-first-order model, and the kobs value was calculated according to the following equation:3\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${{\mathrm{ln}}}\frac{{{\mbox{C}}}_{{\mbox{t}}}}{{{\mbox{C}}}_{0}}=-{k}_{{\mathrm{obs}}}\times {\mbox{t}}$$\end{document}lnCtC0=−kobs×twhere C0 and Ct represent the pollutant concentrations at time zero and at a given reaction time t, respectively. Electrochemical measurements To prepare the catalyst solution, 5 mg of catalyst was dispersed in a 2 mL centrifuge tube containing a mixed solution (500 μL EtOH, 500 μL water, and 15 μL of 5 wt% Nafion solution) by sonication for 30 min. Open-circuit potential (OCP), transient photocurrent responses, and amperometric i-t curve measurements were performed by uniformly coating the mixed solution onto a glassy carbon electrode (GCE), then allowing it to dry naturally. The electrolyte consisted of 0.05 M Na2SO4 aqueous solution (pH = 7.0). All electrochemical measurements were carried out in a standard three-electrode configuration using a catalyst-modified working electrode, a platinum counter electrode, and a saturated calomel reference electrode (SCE). Supplementary information Supplementary Information Transparent Peer Review file Supplementary Information Transparent Peer Review file Source data Source Data Source Data