NH₃ is a fundamental inorganic compound essential in the synthesis of a wide range of nitrogen-based products and plays a pivotal role in agriculture, medicine, energy, military applications, metallurgy, chemistry, and other industries [1–8]. Industrial-scale Haber–Bosch process was applied to produced NH₃ using N₂ and H₂. Nitric acid (HNO₃), another vital industrial chemical, is mainly used to produce ammonium nitrate, calcium ammonium nitrate, and other complex fertilizers produced via a large-scale Ostwald ammonia oxidation process [9,10]. Nowadays, both the Haber–Bosch process and the Ostwald ammonia oxidation process require high energy consumption and emit CO₂, which giving rise environmental risk and high-cost problem [11,12]. Achieving “overall N₂ fixation”, which integrates the Haber–Bosch and Ostwald process, is of economic and ecological gains and application prospects.

Photocatalysis, in particular, holds significant potential for artificial NH₃ synthesis, enabling the N₂ reduction reaction (NRR) under mild conditions with N₂ and H₂O as inputs [13–16]. By promoting the NRR pathway to produce NH₃ and subsequently oxidizing NH₃ within a photocatalytic system to yield HNO₃, an overall N₂ fixation process not only improves economic returns but also provides an environmentally friendly pathway for HNO₃ synthesis. For example, CdS/WO₃ heterojuntion with a built-in electric field and Coulombic attraction effect achieved the overall N₂ photofixation [17]. A suitable band structure that satisfies the redox potential is of necessary to drive the occurrence of overall N₂ photofixation reactions.

Achieving overall N₂ photofixation is challenging due to the sluggish kinetics of NH₃ oxidation. NH₃, being highly soluble in water, readily dissolves and separates from the catalyst surface, making further oxidation to NO₃⁻ difficult. Therefore, controlling NH₃ desorption while promoting oxidation kinetics are critical strategies for advancing overall N₂ photofixation. A FeCo₂O₄ photocatalyst with nanoconfined structure confined N₂ mass transfer and thus promoted the NRR performance [18]. This kind of nanoconfined catalyst structure was widely confirmed to lower the reaction energy barrier [19–21]. Therefore, the nanoconfined photocatalyst structure can facilitate NH₃ accumulation, improving interactions between NH₃ and active sites within the catalyst, thereby accelerating NH₃ oxidation process kinetics.

Single-atom catalysts (SACs) provide unique opportunities to explore these structure-activity relationships at the atomic level due to their well-defined active sites. Research about Ru single-atom over amorphous Fe₂O₃ nanosheets photocatalyst revealed that the photo-electron was transferred from amorphous Fe₂O₃ semiconductor support to Ru single-atom [22]. Single-atom Nb anchored on g-C₃N₄ photocatalyst revealed similar photo-electron transfer mechanism. The photo-electron was accumulated near Nb single-atom site through Nb-C [23]. Research revealed that anchoring single-atom Mo on TiO₂ effectively promoted electron separation and migration, leading to a substantial weakening of the N≡N bond [24]. Electronic metal-support interactions exist between the single-atom sites and the support semiconductor, which integrate active sites and electron acceptors [25–27]. The application of metallic semiconductor nanoparticles in the traditional Haber-Bosch process mainly includes Fe, Ru, Os, Ni, Co, etc. [28]. Nickel oxide (NiO), a p-type semiconductor material, was widely studied in the field of photocatalytic researches [29–32]. NiO was modified by Ni metal and GO for faster carriers’ separation efficiency with better NRR performance [33]. Studies have shown that electron-rich transition metal catalysts are particularly favorable for N₂ adsorption and activation, presenting a promising avenue for enhancing N₂ photofixation performance [34]. The d-orbital of Co element can act as active site for absorbing and activating the N₂ molecule. More important, Co²⁺ has similar ionic radius with Ni²⁺, which can be anchored on NiO semiconductor supporter and difficult to be agglomerated.

In this work, we employed a direct metal atomization method to anchor Co atoms onto NiO@C nanosheets with a nanoconfined pore structure (Co-SAs/NiO@C) for overall N₂ photofixation in pure water. The ink-bottle pores in the carbon layer with nanoconfined structure facilitate NH₃ confinement, thus enhancing NO₃⁻ production and improving overall photocatalytic efficiency. X-ray absorption fine structure (XAFS) analysis confirmed the single-atom anchoring of Co on NiO@C. In-situ XPS tracked photogenerated electron transfer, showing electron flow from NiO@C to Co-SAs sites. DFT calculations further demonstrated an efficient NRR pathway in this photocatalyst, with Co-SAs/NiO@C showing superior capacity for N₂ activation. This work introduces a novel approach for enhancing N₂ photofixation efficiency.

The Co-SAs/NiO@C photocatalyst was synthesized via a direct metal atomization method. Nickel(Ⅱ) acetylacetonate and cobalt(Ⅱ) acetylacetonate were used as the main precursors, with KBr serving as a salt template. Specifically, Ni(acac)₂ (5 g), KBr (25 g), and Co(acac)₂ (0.3 g) were thoroughly homogenized for 4 h using a planetary ball mill at 700 rpm. The resulting powder was then calcined at 300 ℃ for 2 h in an air atmosphere, followed by washing with deionized water and ethanol. The Co-SAs/NiO@C nanosheets were obtained after drying overnight at 60 ℃.

To produce a series of Co-SAs/NiO@C materials with different Co loadings, the amount of Co(acac)₂ was varied, and these products were labeled Co-SAs/NiO@C-x (where x corresponds to the amount of Co(acac)₂ added). The NiO@C sample was synthesized following the same procedure without adding Co(acac)₂. Further, extending the calcination time of Co-SAs/NiO@C-0.3 resulted in Co-SAs/NiO@C-xh materials, where “xh” indicates specific calcination durations.

Previous studies have shown that during high-temperature calcination of metal acetylacetonates, organic ligands first separate from the complex, vaporize, and subsequently condense onto the material's surface as liquid droplets, forming a carbon layer [35]. Accordingly, the carbon component in these photocatalysts exists as a layer on top of the nanosheets.

Co K-edge analysis was conducted at the BL14W1 beamline of the Shanghai Synchrotron Radiation Facility (SSRF) using Si (111) crystal monochromators, and extended X-ray absorption fine structure (EXAFS) spectra were recorded in fluorescence mode. The morphology and structure were observed using transmission electron microscopy (TEM) (TECNAI G2 TF20) and scanning electron microscope (SEM) (Zeiss Sigma 300). Brunauer-Emmett-Teller (BET) surface area was analyzed using TriStar Ⅱ 3020. Powder X-ray diffraction (SmartLab) and XPS (Thermo Scientific K-Alpha) determined the phase composition and chemical properties. UV–vis diffuse reflectance spectra (UV–vis DRS) were collected with a UV-2550 spectrophotometer, using BaSO₄ as the reference material. Raman spectra were acquired using a Horiba (LabRam HR-800) spectrometer. Time-resolved photoluminescence (TRPL) test was recorded on an FLS1000 spectrometer with EPL 375 laser, and electron spin resonance measurements were carried out under light irradiation on a Bruker A300 instrument. The Ni and Co content was determined by ICP-OES on an Agilent 5110 ICP-OES system. The ¹H NMR spectra were acquired using a Bruker Avance Ⅲ HD 600 MHz WB spectrometer at 298 K. More details can be found in Sections S1.1 and S1.2 (Supporting information).

Photoelectrochemical measurements were performed using a CHI 760E electrochemical workstation equipped with a three-electrode system. An Ag/AgCl electrode and platinum wire were used as the reference and counter electrodes, respectively. The working electrode was fabricated on ITO glass (2 × 1 cm²). A 0.1 mol/L Na₂SO₄ aqueous solution was used as the electrolyte, with a 300 W Xe lamp (PLS-SXE300) as the light source. Measurements were converted to the normal hydrogen electrode (NHE) potential scale according to the Nernst equation (Eq. 1):

$ E_{\mathrm{NHE}}(\mathrm{~V})=E_{\mathrm{Ag} / \mathrm{AgCl}}+E_{\mathrm{Ag} / \mathrm{AgCl}}^0 $

(1)

where E_NHE(V) is the converted potential scale vs. NHE, E_Ag/AgCl is the measured potential and E⁰_Ag/Agcl= 0.1976 at 25 ℃.

Carrier density N_c can be calculated from the plot slope using Eq. 2.

$ N_{\mathrm{c}}=2 \mathrm{e}_0^{-1} \varepsilon^{-1} \varepsilon_0^{-1}\left|\mathrm{~d}\left(C^{-2}\right) / \mathrm{dv}\right|^{-1} $

(2)

where e₀ is the electron charge, ε₀ is the vacuum permittivity, ε = 25 is the dielectric constant of NiO [36].

The photocatalytic reaction was evaluated by sonicating 50 mg of the photocatalyst in 100 mL of deionized water, then transferring the suspension to a quartz reactor equipped with a water-cooling system. Before irradiation, the suspension was stirred in darkness for 30 min and purged with high-purity N₂ gas. A 300 W Xe lamp (400 mW/cm²) served as the simulated sunlight source, with continuous N₂ bubbling at 25 ℃.

During irradiation, a 6 mL aliquot was extracted hourly, filtered through a 0.22 µm MCE membrane, and analyzed to determine NH₄⁺ and NO₃⁻ concentrations using Nessler's reagent and the national standard method, respectively. Further details are provided in Sections S1.3 and S1.4 (Supporting information).

DFT calculations were performed using the Quantum Espresso software with the generalized gradient approximation (GGA) and the PBE functional [37,38]. The frozen-core projector augmented-wave (PAW) method described interactions between atomic cores and valence electrons [39]. Wavefunction and charge density cutoff energies were set to 50 and 400 Ry, respectively, and all structural optimizations involved spin-polarization and DFT+U computations. Further details can be found in Sections S1.5 (Supporting information).

The electron exchange process during N₂ adsorption and activation typically involves σ-donation and π-back-donation [32]. Specifically, N₂ adsorption occurs as the metal d-orbital accepts electron density from the N₂ binding orbital, followed by electron density back-donation from the metal to the N₂ anti-bonding orbital. This back-donation process effectively weakens the N≡N bond and strengthens the M-N bond, facilitating N₂ activation, as shown in Figs. S1 and S2 (Supporting information). Once activated, N₂ molecules readily undergo hydrogenation with H⁺ and photoelectrons, making N₂ adsorption and activation on the catalyst surface essential for the NRR. A robust adsorption environment is therefore favorable for N≡N bond weakening, promoting subsequent hydrogenation reactions [40].

NiO, a typical wide-bandgap semiconductor, thermodynamically meets the requirements for overall N₂ photofixation. Integrating single atoms into NiO enhances catalytic performance by providing well-defined active sites. The N₂ absorption site models were discussed in Table S1 and Fig. S3 (Supporting information). We constructed simulation models of NiO and Co-SAs/NiO to assess improvements in N₂ adsorption and activation ability, as shown in Fig. S4 (Supporting information). Electron localization function (ELF) diagrams confirmed the formation of a polar M-N covalent bond, with N₂ adsorbed via chemisorption (Fig. S5 in Supporting information). The N₂ adsorption energy (ΔE_ad) of Co-SAs/NiO (−0.513 eV) was more negative than that of NiO (−0.487 eV), suggesting that Co-SAs/NiO has a stronger adsorption capacity for N₂, as depicted in Figs. 1a and b. Charge population analysis showed a higher degree of electron transfer from the catalyst to N₂ on Co-SAs/NiO compared with NiO, as seen in electron density difference diagrams (Figs. 1c and d) and charge population analysis (Figs. 1a and b). The N≡N bond length increased from 1.126 Å to 1.149 Å, representing enhanced adsorption and activation performance [41–43]. In additional, the introduction of Co atom onto NiO decease the band gap of NiO (Figs. 1e and f, Figs. S6-S8 in Supporting information). Narrowed band gap facilitates the conductivity of the material. Consequently, introducing Co atoms imparts a stronger N₂ adsorption and activation capacity, creating effective active sites.

The preparation process of Co-SAs/NiO@C was depicted in Fig. 2a. The microstructure and morphology of Co-SAs/NiO@C were examined via TEM and SEM, as shown in Fig. 2b and Fig. S9 (Supporting information). The main structure of Co-SAs/NiO@C consists of large carbon multilayer nanosheets (35 µm), with lattice spacings of 0.209 and 0.240 nm corresponding to the (012) and (101) planes of NiO, respectively (Fig. 2c). NiO nanoparticles were uniformly distributed within the carbon nanosheets, and EDX analysis verified the presence of Ni, Co, O, and C elements (Fig. S10 in Supporting information). Raman spectrum (Fig. 2d and Fig. S11 in Supporting information) displayed broad D (∼1340 cm⁻¹) and G (∼1585 cm⁻¹) bands, with an I_D/I_G ratio below 1, indicating a predominantly graphitic carbon structure over Co-SAs/NiO@C and NiO@C [44]. Raman peaks at 479, 766, and 1057 cm⁻¹ were attributed to NiO vibrational modes [45]. XANES at the Co K-edge (Fig. 2e) showed that the oxidation state of Co in Co-SAs/NiO@C lies between those of Co-foil and Co₃O₄. Fourier-transformed EXAFS spectra in R-space (Fig. 2f, Fig. S12 and Table S2 in Supporting information) showed a broad peak at 1–2 Å, attributed to light atom (O) backscattering, indicating Co-O bonding. Similarly, a peak at 2–3 Å corresponded to heavy atom (Co) backscattering, representing a Co-Co coordination environment [46]. Wavelet-transformed EXAFS spectra (Figs. 2g-i) provided a detailed view of Co's chemical environment, further confirming the atomic dispersion of Co in the material. Together, these XAS results validate the successful synthesis of a Co single-atom photocatalyst with well-dispersed Co atoms in NiO@C.

The structure of a series of Co-SAs/NiO@C-xh samples under different treatment conditions was determined through XRD analysis, as shown in Fig. 3a and Fig. S13 (Supporting information). The diffraction peak observed at around 18.1° corresponds to the presence of carbon layer, this provides additional evidence that this peak cannot be attributed to Ni(acac)₂ and confirms that the original structure of Ni(acac)₂ is completely destroyed after undergoing calcination treatment (Fig. S14 in Supporting information) [46]. The broad and weak diffraction peaks at 37.2° and 43.3° demonstrate the presence of NiO with low crystallinity (PDF #44–1159) [47,48]. The low temperature and short heating time during the calcination processes is the key factor for the appearance of carbon layer. By prolonging the calcination time, the carbon layer was gradually disappeared.

The BET analysis of NiO@C and Co-SAs/NiO@C-xh treated with different calcination time all display the Ⅰ/Ⅳ-type adsorption-desorption isothermal curve, as shown in Figs. 3b and c [49,50]. The average pore size in the all samples falls within the mesoporous range (2–50 nm), as shown in Table S3 (Supporting information). With the prolonged calcination time and the disappearance of the carbon layer, there is a significant decrease in specific surface area from 28.1 m²/g to 9.1 m²/g. Additionally, the Co-SAs/NiO@C-2h and NiO@C, which equipped with carbon layer, both exhibit a non-closed loop N₂ absorption-desorption isotherms due to hindered N₂ desorption at relatively low pressures caused by ink-bottle pores [50–52]. Conversely, Co-SAs/NiO@C-4h and Co-SAs/NiO@C-6h without carbon layer display a loop-locked adsorption-desorption hysteresis loop. Therefore, the structure of inner cavity nanoconfined photocatalyst with ink-bottle pore is a crucial factor for confining the gas adsorbed inside the pores of catalyst (Fig. S15 in Supporting information).

The chemical composition and oxidation state of the materials were investigated through XPS analysis. The XPS survey spectrogram reveals that Co-SAs/NiO@C consists of Ni, Co, O, and C without any presence of N element while the Co signal is absented in NiO@C (Fig. S16 in Supporting information). The high-resolution spectrogram of C 1s exhibits four peaks at binding energies of 284.8, 286.6, 288.3, and 292.4 eV corresponding to C—C, C—O, C═O, and π-π* bonds, respectively (Fig. 3d) [53,54]. According to Fig. 3e, the O 1s high-resolution spectrum at the binding energy of 529.7, 531.7 and 533.3 eV can detect the coordination environment of M—O, C═O, and C—O, respectively [55,56]. The M-O bond consists of Ni-O and Co-O bonds, which correspond to the characterization results obtained from XAS analysis. Fig. 3f displays the spectrum of the Ni 2p high-resolution XPS with two peaks at 856.3 and 854.7 eV ascribe to Ni³⁺ and Ni²⁺, while the Ni³⁺ signal peak was attribute to defects caused by low crystallinity NiO [57–59]. The Co high-resolution result confirms the presence of Co element in Co-SAs/NiO@C as shown in Fig. S17 (Supporting information). Due to the low loading amount of Co, the XPS signal for Co high-resolution appears chaotic. However, it is still possible to recognize peaks corresponding to Co 2p_3/2 and 2p_1/2 levels. The ICP-OES results further proved precisely the introduction of Co single-atom (Table S4 in Supporting information).

To evaluate the photocatalytic performance, a series of N₂ fixation experiments were conducted (Figs. 4a-c and Fig. S18 in Supporting information). During light irradiation, the NH₄⁺ concentration displayed an initial increase, followed by a sharp decrease, resulting in a volcano-like trend due to further oxidation of NH₄⁺ to NO₃⁻. As shown in Fig. S19 (Supporting information), N₂H₄ was not detected in the photocatalytic system after 4 h of irradiation, which effectively exclude its involvement in the photocatalytic process. Moreover, NH₃ oxidation is more likely to occur than the pathway of direct N₂ oxidation to NO₃⁻ [60]. The NO₃⁻ concentration continuously increased, confirming the N₂ → NH₃ → HNO₃ pathway within the reaction system. The NiO@C photocatalyst demonstrated relatively low NH₄⁺ and NO₃⁻ generation rates of 9.56 and 13.97 µmol g_cat⁻¹ h⁻¹, respectively. With the increase of Co loading content, the photocatalytic performance increased, and then decreased, in which the performance of Co-SAs/NiO@C-0.3 reached the higher yields of 67.97 and 104.28 µmol g_cat⁻¹ h⁻¹ for NH₄⁺ and NO₃⁻, respectively, within the first hour. Due to the increase of Co single-atom active sites, the photocatalytic performance increases. Since excessive Co content may lead to the aggregation of single-atom and the reduction of the effective active sites, the performance of the catalyst is reduced. Overall N₂ conversion (NH₄⁺ + NO₃⁻) peaked at 282.32 µmol/g_cat after 4 h of illumination, with a NO₃⁻ yield of 60.54% in the first hour (Fig. 4d). Comparison between this work and recent works can be found in Table S5 (Supporting information), which indicates that this catalyst has a potential overall N₂ photofixation performance.

The effect of the carbon layer on performance was further evaluated by comparing Co-SAs/NiO@C-2h, Co-SAs/NiO@C-4h, and Co-SAs/NiO@C-6h samples. Results showed that removing the carbon layer decreased NO₃⁻ yield, as shown in Fig. 4d, indicating the layer's role in NH₃ confinement and oxidation. Stability testing across four cycles showed only a 9% performance decline, demonstrating the catalyst's stability (Fig. 4e). XPS analysis post-reaction indicated negligible compositional changes (Fig. S20 in Supporting information). Isotope tracing experiments with ¹⁴N₂ and ¹⁵N₂ feedstocks, analyzed via ¹H NMR, showed peaks at ¹⁵NH₄⁺ (¹⁵J_NH = 72 Hz) and ¹⁴NH₄⁺ (¹⁴J_NH = 52 Hz), confirming N₂ as the N element source (Fig. 4f).

To achieve the overall N₂ fixation reaction, it is essential to satisfy the reaction potential from a thermodynamic perspective. We analyzed the Co-SAs/NiO@C using linear potential scans to determine the energy levels of the conduction band minimum (CBM) and valence band maximum (VBM) [61,62]. As illustrated in Figs. 5a and b, the cathodic and anodic scan results indicate that the CBM and VBM are positioned at −1.50 V and 2.00 V, respectively. Notably, NiO is classified as a p-type semiconductor, with its flat-band potential (E_fb) being associated with its valence band level. The Mott-Schottky plots reveal that the flat-band potential of Co-SAs/NiO@C is 1.74 V (Fig. 5c). The carrier densities of Co-SAs/NiO@C and NiO@C are measured at 2.8 × 10¹⁸ and 5.6 × 10¹⁷ cm⁻³, respectively. This indicates that the introduction of Co electron traps in NiO@C enhances the photocarrier density, as shown in Fig. 5c and Fig. S21 (Supporting information). Fig. S22 (Supporting information) confirmed that Co-SAs/NiO@C has higher photo-electron response degree compared with NiO@C. Carrier separation efficiency was investigated using TRPL test. According to the exponential decay fit (Table S6 in Supporting information), the prolonged the PL lifetimes of Co-SAs/NiO@C indicate a better carrier separation efficiency for Co-SAs/NiO@C than NiO@C, which could be attributed to the introduction of Co electron trap (Fig. S23 in Supporting information). Moreover, EIS and CV tests further confirmed that Co-SAs/NiO@C have better charge transfer performance and more active sites (Figs. S24 and S25 in Supporting information).

We employed electron paramagnetic resonance (EPR) under illumination conditions, as depicted in Figs. 5d and e. Following a light irradiation period of 10 min, two types of characteristic signal peaks were observed with height ratios of 1:2:2:1 and 1:1:1:1, suggesting the presence of hydroxyl radicals (^•OH) and superoxide radicals (^•O₂⁻) [63,64]. The generation of these free radicals under light exposure indicates that Co-SAs/NiO@C can overcome thermodynamic barriers. This is due to its more negative CBM relative to the superoxide radical (E(O₂/^•O₂⁻) = −0.33 V vs. NHE) and its more positive VBM compared to the hydroxyl radical (E(OH⁻/^•OH) = 1.99 V vs. NHE), which further confirms the reliability of the linear potential scan results. Co-SAs/NiO@C exhibits a band structure with a more positive valence band position compared to the oxidation potential of NH₃/NO₃⁻ (0.36 V vs. NHE, pH 7), and a more negative conduction band position than the potential for NH₃ production (E(N₂/NH₃) = −0.28 V vs. NHE, pH 7). This band structure fulfills the potential requirements for the overall N₂ fixation process, as diagrammed in Fig. 5f.

The migration process of photogenerated carriers was investigated using in-situ XPS, as shown in Fig. 6. Under light illumination, the signal peaks associated with M-O and π-π* bonds shifted toward higher binding energies. Simultaneously, the position of the Ni²⁺ signal peak shifted toward lower binding energies decreased, as illustrated in Figs. 6a-c. These in-situ XPS results indicate that photogenerated electrons transfer from the oxygen atoms and carbon layer to the Ni atoms when NiO@C is exposed to light. Upon the introduction of Co atoms, the Ni 2p signal peak position and the M-O signal peak shifted toward higher binding energies, while the signal associated with Co 2p_3/2 occurred negative movement under light conditions (Figs. 6d-f) [65–67]. This observation suggests that the photogenerated electrons became localized around the Co atoms rather than the Ni atoms. Consequently, after the introduction of the Co electron trap, photogenerated electrons flow from the support to the Co atom and localize around it. This directional photoelectron transfer mechanism, combined with DFT results indicating a strong electron exchange phenomenon between Co and N, highlights that the electron trap and the reaction center are integrated within the Co single atom.

In-situ DRIFTS serves as a powerful tool for unraveling the overall process of N₂ photofixation on the surface and exploring its possible mechanisms. After achieving surface absorption equilibrium in a N₂/H₂O atmosphere, as shown in Figs. 7a and b, and Fig. S26 (Supporting information), various vibrational band signals intensified upon the introduction of an external light source. The peak Ⅰ at 3392 cm⁻¹ is attributed to N—H stretching vibrations, while peaks Ⅱ and Ⅶ at 2846 cm⁻¹ and 1380 cm⁻¹ correspond to NH₄⁺ stretching vibrations. Peak Ⅲ, assigned to the bending mode of σ(N—H), appears at 1691 cm⁻¹. Peak Ⅳ, corresponding to adsorbed N₂ (*N₂), is observed at 1625 cm⁻¹. The characteristic absorption peak of NH₃ is identified at 1557 cm⁻¹ (peak Ⅴ), and peak Ⅵ at 1463 cm⁻¹ indicates the generation of NO₃ [60,68–70]. All peaks gradually increased with the duration of irradiation, demonstrating the continuous generation of NH₄⁺ and NO₃⁻. These results suggest a consecutive reaction process of N₂ → NH₃ → HNO₃ on the Co-SAs/NiO@C catalyst, as illustrated in Fig. S27 (Supporting information).

The NRR pathway is typically divided into distal and alternate pathways. To determine the dominant NRR pathway, we calculated and compared the energy barriers for both pathways (Fig. 7c). According to the free energy profile of the reaction pathway, Co-SAs/NiO@C demonstrates a stronger N₂ adsorption capacity, as discussed previously. The rate-determining step (RDS) for Co-SAs/NiO@C involves hydrogenation on the distantly adsorbed N atom, with an energy barrier of 1.43 eV, while the RDS for NiO@C has a higher energy barrier of 1.78 eV. A lower RDS is advantageous for the NRR, aligning with the observed catalytic performance. The distinctions between the distal and alternate pathways lie in the intermediate products formed from the second hydrogenation site leading to the *NH₂ intermediate. Notably, Co-SAs/NiO@C and NiO@C exhibit different bonding configurations of N atoms in their alternate pathways. The results of geometry optimization, as shown in Fig. 7d and Fig. S28 (Supporting information), indicate that the two N atoms tend to be vertically adsorbed on the surface of the NiO@C catalyst, whereas they are adsorbed in parallel on the surface of Co-SAs/NiO@C, with both N atoms forming bonds with the catalyst. This parallel adsorption mode facilitates the activation and polarization of the N≡N bond, thereby reducing the energy barrier for the reaction. Consequently, the results indicate that the alternate pathway is the dominant pathway for Co-SAs/NiO@C, while the distal pathway tends to prevail in NiO@C.

Reaction kinetics play a crucial role in enhancing chemical reactions. The introduction of Co electron traps promotes the localization and enhancement of photoelectron density. As the active site for N₂ absorption and activity, the Co single-atom facilitates the transfer of photoelectrons to the adsorbed N₂, effectively improving the kinetics of the NRR. Furthermore, the nanoconfined structure of the carbon layer, featuring ink-bottle pores, directly prevents the desorption and diffusion of NH₃ in water. This nanoconfined structure creates a sufficient reaction environment for both the generated NH₃ and the catalyst. The thermodynamics of the overall N₂ photofixation reaction is synergistically enhanced by the nanoconfined structure and single-atom active sites. The overall reaction pathway for N₂ photofixation can be summarized as follows (Fig. 8). (1) Adsorption: N₂ gas is initially adsorbed onto the surface of the Co-SAs/NiO@C nanosheets. It is then reduced to NH₃ gas through an alternate pathway facilitated by photogenerated electrons localized around the Co active sites. (2) Formation of NH₄⁺ and confined NH₃: A portion of the produced NH₃ dissolves in water, forming NH₄⁺, while another portion remains confined within the ink-bottle pores, continuing to adsorb onto the surface of Co-SAs/NiO@C. (3) Oxidation to NO₃⁻: The NH₃ absorbed on the surface pores is preferentially oxidized to NO₃⁻ by the Co-SAs/NiO@C. This surface in-situ overall N₂ photofixation reaction (N₂ → NH₃ → HNO₃), facilitated by the nanoconfined structure, is essential for increasing the proportion of NO₃⁻ among the N₂ fixation products.

In summary, encouraged by the DFT calculations, we developed a novel Co single-atom catalyst supported by NiO@C nanosheets with a nanoconfined structure for overall N₂ photofixation. Both DFT and experimental results confirmed that the Co-SAs/NiO@C achieve overall N₂ photofixation effectively. Co single-atom sites integrated N₂ achieve site and photo-electron trap, which enhance the NRR kinetics. In additional, the introduction of Co single-atom reduce the NRR energy barrier. Focus on NH₃ oxidation reaction, the nanoconfined structure of ink-bottle pores hinder the desorption of NH₃ and thus improve NH₃ oxidation kinetics. In short, the Co single-atom modification over a nanoconfined structural support facilitates a deeper N₂ conversion pathway with a high percentage of NO₃⁻ production.

Declaration of competing interest

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

CRediT authorship contribution statement

Longjian Li: Writing – original draft, Visualization, Validation, Software, Resources, Methodology, Investigation, Formal analysis, Data curation, Conceptualization. Ping Zhang: Writing – review & editing, Visualization, Validation, Supervision, Software, Resources, Project administration, Methodology, Investigation, Funding acquisition, Formal analysis, Data curation, Conceptualization. Yongchong Yu: Writing – original draft, Software, Resources, Methodology, Investigation, Formal analysis, Data curation. Reyila Tuerhong: Writing – original draft, Software, Resources, Methodology, Investigation, Formal analysis, Data curation. Xiaoping Su: Writing – original draft, Validation, Software, Resources, Methodology, Investigation, Formal analysis. Lijuan Han: Writing – original draft, Visualization, Validation, Software, Resources, Methodology, Investigation, Formal analysis, Data curation, Conceptualization. Enzhou Liu: Writing – original draft, Visualization, Validation, Software, Resources, Methodology, Investigation, Formal analysis, Data curation. Jizhou Jiang: Writing – review & editing, Visualization, Validation, Supervision, Software, Resources, Project administration, Methodology, Investigation, Funding acquisition, Formal analysis, Data curation, Conceptualization.

Acknowledgments

This work was supported by the National Natural Science Foundation of China (No. 62004143), the Key Research and Development Program of Gansu Province - Industrial Project under Grant (No. 25YFGA058), the Key Talent Project Foundation of Gansu Province (No. 2025RCXM066), the Gansu Provincial Department of Education: Industrial Support Plan Project (No. 2025CYZC-005), the Key R&D Program of Hubei Province (No. 2022BAA084), the Science and Technology Project of Lanzhou (No. 2024-3-42), the Fundamental Research Funds for the Central Universities (No. 331920240059).

Supplementary materials

Supplementary material associated with this article can be found, in the online version, at doi:10.1016/j.cclet.2025.111118.