Localized hetero-ion modulation engineering of nickel-based catalysts for electrochemical urea oxidation

Abstract

The electrochemical urea oxidation reaction (UOR) is an effective alternative to the oxygen evolution reaction (OER) for the electrochemical production of hydrogen; however, nickel (Ni)-based catalysts show inferior reaction kinetics. To overcome the performance limitations, localized hetero-ion modulation of the intrinsic properties of Ni-based catalysts has evolved to meet practical expectations. The modulation of the local coordination environment by incorporating hetero-ions is imperative for enhancing the intrinsic catalytic activity and efficiency of nickel-based catalysts. Guided by this insight, a systematic comprehension of localized hetero-ion modulation in the UOR is presented in this review to provide atomic-level insights into the catalytic mechanism and catalyst design for Ni-based catalysts. This review begins with an introduction to the electrochemical reaction pathways of the UOR, followed by the physiochemical properties of Ni(OH)₂ and universal scaling relations of reaction intermediates, providing a fundamental understanding of the structure-derived electrochemical behavior in the UOR. Subsequently, the specific functions of hetero-ion species for the enhancement of the UOR in Ni-based catalysts are discussed, delving into the close relation between the localized reaction site formats and the reaction pathways. In the conclusion of this review, data-driven catalysis for the electrochemical UOR via hetero-ion modulation over the Ni catalyst based on a closed-loop framework is proposed. We believe that this review will attract pronounced attention for the advancement of the electrochemical UOR.

---

1. Introduction

The ever-increasing global consumption of fossil fuels raises environmental concerns, posing challenges to the rapid advancement of sustainable energy resources.^1–6 To overcome the severe environmental crisis, tremendous attempts have been devoted to novel energy systems including green hydrogen, nitrogen cycle, and carbon dioxide cycle.^7–11 Notably, electrochemical water splitting is renowned as a flexible system that is able to couple with renewable solar, wind, and tidal power sources.^12–18 However, originating from the four-electron transfer, the sluggish kinetics of the anodic oxygen evolution reaction (OER) largely limits the hydrogen production efficiency, representing a significant bottleneck.^19–21 Recently, the electrochemical oxidation of various organic molecules, including alcohols, amines, and hydrazine, has been discovered to replace the energy-burdened water-splitting process.^22–26 Among these promising candidates, the electrochemical urea oxidation reaction (UOR) has garnered enhanced attention as an effective solution to this drawback in small-organic molecule oxidation due to the availability of feedstock from wastewater and animal metabolism.^27,28 More importantly, owing to the low thermodynamic equilibrium potential, the UOR can remarkably decrease the electrolytic voltage and energy consumption compared with alcohol oxidation. Meanwhile, domestic and industrial wastewater contains roughly 80% urea, and the UOR achieves simultaneous wastewater denitrification and hydrogen production, alleviating the toxic ammonia influence that contributes to acid rain, thus facilitating a beneficial cycle in the agricultural ecosystem.^29–31

In principle, the thermodynamic equilibrium potential of the UOR is assigned to be 0.37 V versus reversible hydrogen electrode (vs. RHE), corresponding to the chemical conversion of urea into harmless products such as H₂O, CO₂ and N₂, which is obviously lower than that required to drive the OER of 1.23 V (Fig. 1a).^32–36 Therefore, the replacement of the OER by the UOR can decrease the cost by ∼36% in terms of hydrogen production. Furthermore, benefiting from the unique gravimetric hydrogen content (6.7 wt%), urea acts as the hydrogen carrier that enables its application in direct urea fuel cells (Fig. 1b).^37–39 With the direct coupling between the UOR and hydrogen evolution, the utilization of urea electrolysis further provides an advanced energy conversion framework, benefiting the simultaneous approach of hydrogen production and purification of urea-abundant wastewater. However, the practical application of the UOR within these frameworks is still deterred by the sluggish six-electron transfer kinetics, side reactions for the release of pollutants (e.g. NO₂⁻ and CNO⁻), and the detrimentally irreversible phase transformation of nickel oxides under the operando oxidative conditions, thus resulting in the urgent requirement of designing and discovering highly effective catalysts.^40–42

Fig. 1 (a) Qualitative display of the polarization curves for electrochemical urea oxidation, oxygen evolution, and hydrogen evolution. (b) Schematic of direct urea fuel cell and urea electrolysis systems. (c) Molecular property of urea, including electrostatic potential distribution and Mulliken population analysis, optimized by the CP2K code at the RPBE-D3(BJ)/DZVP-MOLOPT-SR-GTH level, where the iso-surface value is set to be 0.15 (a.u.).^58–67 (d) Timescale for the development of localized modulation of the Ni catalysts in the UOR.

With respect to the urea molecule, the molecular connection between the carbonyl group (C=O) and the amino group (–NH₂) contributes to the polarized property, where the C=O group area mainly acts as the electron-negative accumulation while the amino group area behaves as the electron-positive depletion (Fig. 1c). Benefiting from such topological polarization, the specific chemical adsorption of the urea molecule by electrostatic interaction discloses the non-identical splitting pathways during the electrochemical oxidation process, depending on the local electrophilic and nucleophilic property of the catalyst surface sites.^43–45 As the most promising non-noble candidates, nickel-based catalysts have emerged for the UOR owing to the adjustable active sites and low cost, since the initiation by Botte et al. in 2009, where the employment of Ni(OH)₂ reveals the performance under alkaline conditions (Fig. 1d).^46–49 However, limited by the insufficient reaction rate and catalytic degradation under practical application conditions, the intrinsic performance of Ni(OH)₂ should still be modulated through external strategies, including the incorporation of defects, construction of active phases, and crystallinity evolution, as concluded in several recent reviews and works.^50–57 Despite this focus on the modulation strategies over nickel-based catalysts, the localized hetero-ion modulation provides atomic precision targeting the effective enhancement of the UOR and a straightforward model for understanding the reaction mechanisms based on the active unit site simultaneously. More specifically, the atomic-precision doping of hetero-ion on Ni-based materials can modulate the coordination environment of Ni active sites effectively. Meanwhile, interfacial electron transfer and orbital hybridization between hetero-ion and Ni sites can also regulate the d-band center and valence state configuration of Ni species, thereby decreasing the adsorption strength of UOR intermediates and significantly boosting the catalytic activity. Therefore, localized hetero-ion modulation exhibits unique advantages in Ni-based materials and exhibits important research significance. However, for this, a dedicated review is urgently required exclusively on localized hetero-ion modulation in light of structure-derived catalytic mechanism, which motivates a deep comprehension of the rational design of highly performing UOR candidates.

Herein, we mainly concentrate on the most advancement of localized modulation via the incorporation of hetero-ions into Ni-based catalysts that form the unique unit active site for the electrochemical UOR. To lead the insightful atomic comprehension of the localized modulation framework, the basic electrochemical mechanism of the UOR over Ni-based catalysts is first discussed briefly including the physiochemical properties of Ni(OH)₂ and the universal scaling relations of the free energy of reaction intermediates. Afterwards, detailed catalyst-driven reactions are discussed in terms of hetero-ion coordination modes, UOR performance, and oxidation pathways, highlighting the evolution of hetero-ions from main group and transition-metal species to rare-earth elements. Finally, in the conclusion of this review, an outlook for the future development of data-driven catalysis over the UOR based on a closed-loop framework is proposed, aiming at the comprehensive advancement of novel and high performing nickel-based catalysts. The chemical insights obtained from this review are hopeful to guide the future novel discovery of high-performance nickel-based catalysts to tackle the fundamental bottlenecks of electrochemical urea splitting.

2. Mechanical insights into electrochemical UOR

2.1. Typical reaction pathways

Similar to the OER, the UOR in aqueous solutions occurs at the anode in urea-driven electrochemical systems, which experiences the complex multiple H⁺/e⁻ transfer steps along with specific chemical steps. The typical format of UOR follows CO(NH₂)₂ + H₂O → N₂ + 3H₂ + CO₂, requiring a thermodynamic potential of 0.37 V, that is lower than that of the OER.^68,69 However, due to the transfer of a greater number of electrons with the complex chemisorption behavior of various intermediates, the slower reaction kinetics and complicated side reactions still necessitate the determination and discovery of highly effective pathways for urea splitting, which is vital for reducing the cost of green hydrogen production and the degradation of agricultural pollutants in an economical way. Presently, several reaction pathways have been proposed for potential-driven splitting of urea by Ni-based materials through sophisticated physical characterizations and theoretical investigations. The representative mechanisms of the UOR are discussed over typical nickel-based (notably the nickel hydroxide) catalysts briefly in this section.

In principle, the UOR should proceed through the chemical adsorption of urea molecules at the active sites of catalysts, where the consecutive release of protons and electrons drives the cleavage of C–N and the coupling of the N–N bond, till the creation of final products including NH₃, N₂, NO₂⁻, NO₃⁻, CO₂, and CO₃²⁻, depending on the oxidation condition of the Ni-based catalyst surface.^70,71 Since the UOR appears in the oxidative potential window, the oxidation of Ni-based catalysts is unavoidable. Initially, it is demonstrated by Botte et al. that Ni(OH)₂ in basic media can be oxidized into NiOOH, where the chemical transition from Ni²⁺ into Ni³⁺ is noted as the active state.⁷² Afterwards, Botte et al. also performed the theoretical investigation for the dissociation rate of urea over NiOOH catalysts.⁷³ It is illustrated that the absence of OH⁻ can promote the transformation of urea into NH₃ and HNCO with the formation of NH₃ as the rate-limiting step (RDS) at a rate of 1.5 × 10⁻⁶ s⁻¹, while the presence of hydroxide assigns the creation of NH₃ as the RDS at a rate of 1.4 × 10⁻²⁶ s⁻¹. The deactivation of UOR performance due to surface blockage by *CO₂ confirms the catalytic role of OH⁻ attack in urea splitting, providing an initial model for the nucleophilic attack mechanism on NiOOH. As shown in Fig. 2a, the nucleophilic OH⁻ attack for the UOR over the nickel catalyst was first proposed, where the Ni site mainly experiences the splitting of urea molecules in the first stage.⁷⁴ In detail, the urea molecule is first adsorbed at the active Ni site through oxygen coordination, and afterwards, the consecutive OH⁻ attack towards the C–N bond leads to the formation of *OCONH₂ and the release of NH₃. In the second stage, the OH⁻ attack further transforms the NH₃ into *NH₂, and the consecutive intermolecular coupling between two adjacent *NH₂ species guides the formation of N₂ through the release of H⁺/e⁻ over the active Fe sites via the electrochemical oxidation of *NH₂ (*NH₂NH₂ → *NHNH₂ → *NNH₂ → *NNH → N₂).

Fig. 2 (a and b) Nucleophilic attack mechanism and intramolecular coupling mechanism. (c and d) Adsorbate evolution mechanism and lattice oxygen-assisted mechanism. The green, blue, grey, red, and orange balls represent catalyst surface, nitrogen, carbon, oxygen, and hydrogen atoms, respectively. The inset plot shows the schematic of the anodic surface distribution of urea and solvents, where the proceeding reaction is proposed to experience dynamic surface state changes with solvent reorganization and configuration shift of adsorbates, which can be projected onto the potential energy surface during the intramolecular N–N coupling process.

It is observed in the splitting process of urea molecules that the cleavage of the C–N bond plays a central role in terms of the formation of final products such as CO₂ and N₂. However, such cleavage typically involves the reconstruction of the –CONH₂– group, including the intramolecular and intermolecular N–N coupling over Ni(OH)₂ (Fig. 2b).⁷⁵ In terms of the intramolecular N–N coupling mechanism, the urea molecule experiences the adsorption and the consecutive release of H⁺/e⁻ to form the transformation of *CONHNH₂ into *CONHNH. The subsequent coupling between two –NH groups within *CONHNH leads to the formation of *CO and N₂, where the adsorbed *CO is further oxidized to CO₂ (CO₃²⁻) in alkaline media under oxidative condition. Practically, the over-oxidation of NiOOH derived from the Ni(OH)₂ phase often deviates the innocuous N₂ product into oxidative species such as N₂O and NO₂⁻ (NO₃⁻), which is the intermolecular N–N coupling after the C–N cleavage by OH⁻ attack with the assistance of additional urea molecules. Specifically, with the nucleophilic attack from the OH⁻ species in an alkaline electrolyte, the C–N cleavage of *CONHNH₂ forms *NH₂ and the intermediate product of NH₂–COOH, where the further attack by OH⁻ species towards *NH₂ is able to release products such as NO₂⁻ and NO₃⁻. The proceeding of intermolecular N–N coupling after the C–N cleavage can be facilitated by increasing the concentration of urea, which promotes the formation of N₂O. The existence of the competitive intermolecular N–N coupling can be attributed to the high energy barrier when the two N atoms approach each other that contributes to the intermolecular N–N coupling, where the kinetics of C–N cleavage by OH⁻ attack is more favorable than the formation of *CON₂.

Additionally, it witnesses that the competitive OER possesses a similar potential window as with the UOR to some extent. Thus, the participation of H₂O and the nucleophilic attack by OH⁻ in the above UOR mechanism may also rationalize the typical pathways as from the OER, notably the adsorbate evolution mechanism (AEM) and lattice oxygen mechanism (LOM) over Ni(OH)₂. In terms of the AEM pathway (Fig. 2c), similar to the above N–N coupling, the previous steps involve the adsorption of urea molecules and dehydrogenation for the formation of the *CONH₂NH₂ intermediate. The key step in the AEM pathway lies in the intramolecular N–N coupling of *CON₂ after the dehydrogenation of *CONH₂NH₂, where the subsequent formation of *COOH from *CO is available to guide electrooxidation into CO₂. Among these elementary steps, the desorption of *CO₂ is assumed to be the most rate-determining step (RDS) that limits the overall UOR process. Distinct from the AEM pathway, the LOM pathway is associated with expediated kinetics, where lattice oxygen species act as the nucleophilic attack site (Fig. 2d). The activated lattice oxygen site is able to combine with urea molecule that forms oxygen vacancies during the UOR process, thus completing the whole catalytic loop. After the formation of *CO₂NNH₂, the C–N cleavage with dehydrogenation allows for the creation of *NH₂ and *N on the catalyst surface with the release of CO₂ as the product. The final coupling between *NH₂ and *N leads to the formation of N₂ as the product.

Furthermore, distinctive UOR processes can occur in neutral or acidic media, where NaCl or HCl acts as the solute. However, it should be pointed out that the oxidation of Cl⁻ is also unavoidable during the electrooxidation process that forms side products such as HClO and Cl₂ with oxidative power. The disadvantages of electrolytes beyond alkaline media include low faradaic efficiency due to side reactions and shortened lifespan of Ni-based materials caused by harsh H⁺ electrolytes and highly oxidative conditions. As a result, few works are concentrated over acidic UOR processes compared with neutral and alkaline media. Undoubtedly, the deep insight into the UOR will largely provide novel guidance for the design of efficient transformation systems.

2.2. Electrochemical property of the active Ni(OH)₂ phase

Nickel hydroxide typically exhibits a prototype layered structure characteristic of common transition-metal-based oxides, where the direct exposure of surface –OH sites provides abundant H⁺/e⁻ transfer kinetics. With the interlayer insertion of water molecules and alkaline metal cations, different bulk phases are available, comprising α-Ni(OH)₂ and β-Ni(OH)₂.⁷⁶ When serving as electrocatalysts, Ni(OH)₂ exposed to harsh conditions, including high electrolyte pH and oxidative bias potential, undergoes phase transition into NiOOH and NiO₂ with the depletion of surface H sites (Fig. 3a). The increased oxidation of Ni²⁺ into Ni³⁺ in NiOOH and even Ni⁴⁺ in NiO₂ equips the Ni site with oxidation power that drives electrochemical oxidation reactions such as the OER, alcohol oxidation, and UOR. Notably, under certain pH and bias potential conditions, the water molecule will experience electrochemical equilibrium with the Ni(OH)₂ surface through the coverage effect of various intermediates, which further affects the reaction progress. As shown in Fig. 3b, the ab initio energy of various *H coverage states over the NiOOH surface was used to construct the surface Pourbaix diagram.⁷⁷ The horizontal line was taken as the basic reference, denoting the fully deprotonated surface at pH = 14 for both the surface and the subsurface *H coverage. It should be noted that at potentials below U < 0.7 V (vs. SHE), most state is expected to be fully covered with *H species on both the surface and subsurface regions. A further increase in the bias potential will promote the deprotonation with the decrease in the surface coverage of *H that drives the deep phase transformation from NiOOH into NiO₂. Concerning the potential window for the UOR, the electrochemical stable phase assigns the transition between Ni(OH)₂ and NiOOH, where the Ni³⁺ site with a higher oxidation number is promising to act as the active center as is typically different from the competitive OER process.

Fig. 3 (a and b) Schematic of the phase transformation of Ni(OH)₂ and the corresponding surface Pourbaix diagram with *H adsorbates.⁷⁷ Copyright 2024, the American Chemical Society. (c and d) Comparison of Pourbaix diagrams for Ni(OH)₂ with and without Fe dopants, along with the AIMD simulation results for the dynamic evolution of Ni(OH)₂ with surface Fe species with various models: (i) 3UP model, (ii) 2UP model, (iii) BRIDGE model, (iv) 2IN6 model, (v) 3IN6 model, and (vi) TILT model. The accompanied plots show the coordination numbers and oxidation states changing with the simulation time series.⁷⁸ Copyright 2020, the American Chemical Society.

It is observed in the electrochemically driven phase transition to the active Ni(OH)₂ phase, the change in the coordination environment around Ni is found to modulate the intrinsic electrochemical property. Taking Fe as the model, Zhou et al. unveiled the dynamic surface state change under oxidative conditions in terms of the formation of several metastable NiO_xH_y phases and Fe doping effects.⁷⁸ As shown in Fig. 3c, with the incorporation of Fe, the phase equilibrium potentials all decrease compared with the undoped condition under the same pH window. Such phenomenon states that the incorporation of Fe into Ni(OH)₂ facilitates the oxidation of Ni²⁺ into a higher oxidation state, which is vital to drive the electrochemical oxidation. Notably, the pH window for the Ni(OH)₂ phase area is extended after Fe doping, implying enhanced tolerance for the pH change to promote the phase stability towards harsh electrolytes. In terms of the exfoliated gel-like structure, Fe ions are able to be inserted into the interlayer with more complex coordination with water molecules or layered hydroxyls of the NiO_xH_y surface, which is determined by the favorable adsorption catalyst of the octahedral Fe²⁺ structure. To further testify the dynamic change in water intercalation with adsorbed Fe species, various configurations of the Fe–(H₂O)_x structure were further investigated within ab initio molecular dynamics (AIMD) simulations (Fig. 3d). To understand the Fe-induced dynamic change of the local structure, various models are referred including 2UP, 3UP, 2UP-TETRA, 3UP-TETRA, 2IN6, 3IN6, BRIDGE, and TILT models. Among these models, the 2UP and 3UP models are more active that the coordinated H₂O with Fe²⁺ (2UP model) and Fe³⁺ (3UP model), which is easier to provide the surface oxygen of NiOOH with released proton to lead the dynamic phase equilibrium. Compared with Fe²⁺, coordinated H₂O with the Fe³⁺ center favors splitting to attract more OH⁻ species, implying the enhanced electrified field induced by the highly positive charge of Fe³⁺. With the increased coordination of O around Fe³⁺ such as in the BRIDGE model, high oxidation state of Fe⁴⁺ can be obtained, where a similar phenomenon can also be achieved in the 2IN6 model as of the dynamic charge transfer between Fe²⁺ and Fe⁴⁺. By noting about such dynamic change of oxidation state, the local polarons induced by the high oxidation state can further influence the reactivity such as CO reduction.

2.3. Construction of universal scaling relations

Heterogeneous catalysis includes the complex adsorption of various adsorbates over the solid surface, where the transformation of both stable species and transition states describe the catalyst's surface states. As initialized by Hammer and Nørskov et al., the proposal of the d-band theory reduced by the Newns–Anderson–Grimley model provides the basic prototype of linear correlation between the adsorption energies of distinct adsorbates on metal surfaces.^79,80 Consolidated by the d-band theory, Abild-Pedersen and Nørskov further disclose the term of scaling relation for describing the energies of hydrogenation or dehydrogenation reactions for organic molecules over transition-metal surfaces with the assistance of density functional theory calculations.^81,82 To be specific, the adsorption of the molecule AH_x is determined to be linearly correlated with the adsorption energy of atom A in the simple format of ΔE^AHx = γΔE^A + ξ, where the slope γ is assumed to be affected by the valency of the adsorbate while the intercept ξ denotes the energy deviation. Beyond the constructed scaling relation based on the chemisorption energy, the Brønsted–Evans–Polanyi (BEP) relationships are constructed for surface reactions that describe the function between the transition state energy (E_TS) and the reaction energy (ΔE), E_TS = γΔE + ξ. As a result, with these universal scaling relations, the coupled ordinary differential equations based on the rate theory from Langmuir–Hinshelwood–Hougen–Watson (LHHW) expressions can be solved to obtain the microkinetic property of the catalytic system. That is, the construction of universal scaling relations allows for the simple physical descriptor to describe the whole microkinetic property, largely decreasing the computation cost of DFT for complex heterogeneous catalysis systems. However, when it comes into the electrochemical reactions, owing to the driven effect from external bias potential, the redox behaviors induce the complex electron transfer within the Helmholtz layer that is hard to be treated in a constant charge model.^83,84 Therefore, the incorporation of electronic potential energy should be further considered beyond the intrinsic chemical reaction energy.

In terms of the UOR process, the universal scaling relations were constructed by Lu et al. over transition-metal (TM) pair atoms loaded on a C₂N substrate (Fig. 4a) and Zhan et al. over metal-doped CoS nanosheets (Fig. 4b), respectively.^85,86 In detail, for the transition-metal pair atoms supported on a C₂N substrate, two different pathways are considered including the O_ter pathway and N_ter pathway, where the former pathway involves the adsorption of intermediates with O–TM interaction, while the later pathway experiences the N–TM interaction (Fig. 4c). It is noted that the intermediate *CONHN was chosen as the key descriptor for interaction investigation between the catalytic site and adsorbates. As shown in Fig. 4d, the adsorption of *CONHH presents the even end-on mode over the Ni₂/C₂N substrate, where the symmetric metal atom pair allows the homogeneous adsorption mode with the medium transfer compared with the CrNi/C₂N and CuNi/C₂N system. The linear regression in Fig. 4e states that the adsorption free energy of *CONHN provides strong correlation with *CONH₂N with the format of ΔG(*CONH₂N) = 0.95ΔG(*CONHN) + 0.95, where the average R² value exhibits 0.90. For *CONN, the *CONHN also reveals the strong correlation of ΔG(*CONN) = 0.79ΔG(*CONHN) + 1.32, where the average R² value is assigned to be 0.97 (Fig. 4f). Besides the adsorption energy as the descriptor in the UOR, the integral crystal orbital Hamilton population (ICOHP) below the Fermi level can be used. As shown in Fig. 4g, the RDS in the UOR is the cleavage of the C–N bond in urea, and ΔG_RDS obtained by the subtraction of the free energy of *NH from *urea-H shows a strong linear correlation with ΔICOHP defined by metal-nitrogen pair interaction, where the R² value is determined to be 0.996.

Fig. 4 (a) Schematic of the layer model for TM atomic pair loaded on a C₂N substrate.⁸⁵ Copyright 2025, the Royal Society of Chemistry. (b) Schematic of the pristine CoS and metal-doped CoS model.⁸⁶ Copyright 2025, the American Chemical Society. (c) Process of the O_ter and N_ter pathways for the UOR.⁸⁵ Copyright 2025, the Royal Society of Chemistry. (d) Charge density difference for the adsorption of *CONHN with Bader charge population.⁸⁵ Copyright 2025, the Royal Society of Chemistry. (e and f) Scaling relations of ΔG(*CONH₂N) and ΔG(*CONN) with ΔG(*CONHN).⁸⁵ Copyright 2025, the Royal Society of Chemistry. (g) Scaling relation constructed by ΔICOHP and ΔG_RDS over the metal-doped CoS system.⁸⁶ Copyright 2025, the American Chemical Society.

3. Modulation types over the nickel catalyst

3.1. Localized modulation of hetero-anion incorporation

As the typical two-dimensional material, Ni(OH)₂ usually exposes the low-coordinated surface sites to lead the electrocatalytic UOR process, while the strong on-site coulombic interaction derived from the Ni 3d shell suffers from the inferior electrical conductivity, limiting the expectation of large current density for the electrochemical UOR. Taking β-Ni(OH)₂ as the example, Carter et al. performed the insightful investigation over the structural and electronic properties of β-Ni(OH)₂ and β-NiOOH based on first principles, where the formation of β-NiOOH is denoted as the deprotonation of β-Ni(OH)₂.⁸⁷ Using the calculation of single-shot Green function (G₀W₀) for the correction of notorious self-interaction and derivative discontinuity error, the band gaps of β-Ni(OH)₂ and β-NiOOH are determined to be 5.83 eV and 1.96 eV, respectively, stating that β-Ni(OH)₂ is a wide-gap insulator and β-NiOOH possesses an indirect gap as the semiconductor that guides the inferior electrical conductivity. Besides, the lowest energy state for β-NiOOH assumes the low-spin and antiferromagnetic configuration, where the low-spin d⁷ state of Ni sites favors the Jahn–Teller distortion with the formation of two types of Ni–O bond length. Such distortion is the basis of heteroatom incorporation for β-NiOOH to modulate the intrinsic property for meeting the operando condition of the electrochemical UOR.

To improve the electrical conductivity and UOR performance of β-Ni(OH)₂, the surface sulfur incorporation in the replace of O site presents enhanced metallic property with effective electron transport by heating β-Ni(OH)₂ nanosheets under H₂S (Fig. 5a).⁸⁸ The sulfur incorporation into the β-Ni(OH)₂ nanosheet can be viewed as the exchange between H₂O derived from the surface –OH group and the H₂S molecule with a –SH group. As shown in Fig. 5b, the overall formation of metallic Ni(OH)₂ experiences an exothermic energy change of 8.49 eV, where the protonation of the surface –OH group into the release of H₂O is assigned to the larger downshift of energy along the reaction progress. After the sulfur incorporation, the metallic Ni(OH)₂ noted as M-Ni(OH)₂ exhibits the zero band contributed by the spin-down electronic states in comparison with pristine Ni(OH)₂ as P-Ni(OH)₂, demonstrating the metallic property of M-Ni(OH)₂ (Fig. 5c). Consistently, the M-Ni(OH)₂ nanosheets deliver increased electrical resistivity with the enhancement of temperature, where such type of metallic behavior is determined to possesses a resistivity of 3.13 × 10⁻⁴ Ω m under the room temperature with ultrahigh electrical conductivity (Fig. 5d). Additionally, M-Ni(OH)₂ also shows an electron concentration of ∼10²⁰ cm⁻³ with negative values of Hall coefficients, confirming the promotive role of sulfur incorporation for the electrical transport of P-Ni(OH)₂ (Fig. 5e). Benefiting from the metallic property of M-Ni(OH)₂, the higher peak current density in 1 M KOH + 0.33 M urea is achieved with the contribution from urea compared with P-Ni(OH)₂, confirming the enhanced UOR performance of M-Ni(OH)₂ (Fig. 5f). Meanwhile, M-Ni(OH)₂ exhibits durable stability in charge transport and active site regeneration. With constant voltage analysis, M-Ni(OH)₂ delivers constant and high current density, along with a stable current response, while P-Ni(OH)₂ suffers obvious current attenuation. These results confirm that sulfur incorporation enhances the urea oxidation activity and operational stability of β-Ni(OH)₂, which is consistent with the above-mentioned mechanism analysis. It should be noted that the poison effect by the adsorption of CO₂ can also affect the catalytic performance beyond the high conductivity property. When NiOOH acts as the catalytic phase, the RDS during the oxidation of urea is the desorption of CO₂ with a high energy barrier of 1.48 eV over P-Ni(OH)₂. In comparison, the energy barrier of RDS over M-Ni(OH)₂ states a lower value of 1.17 eV, suggesting that the sulfur incorporation can effectively alleviate the poison effect by CO₂ for the favorable UOR performance (Fig. 5g).

Fig. 5 (a) Synthetic process of M-Ni(OH)₂ with the S doping via heating H₂S. (b) Schematic of the reaction between H₂S and P-Ni(OH)₂ nanosheets to create M-Ni(OH)₂. (c) Density of states (DOS) of P-Ni(OH)₂ and M-Ni(OH)₂. (d) Electrical resistivity of the M-Ni(OH)₂ sample with the change in temperature. (e) Carrier concentration and Hall coefficient of the M-Ni(OH)₂ sample. (f) Linear sweep voltammetry (LSV) curves for P-Ni(OH)₂ and M-Ni(OH)₂ samples in 1 M KOH + 0.33 M urea. (g) Energy profile for CO₂ desorption during the UOR over P-Ni(OH)₂ and M-Ni(OH)₂.⁸⁸ Copyright 2016, Wiley-VCH GmbH.

Except for the strategy based on heteroatom incorporation, the inducement of anion vacancy also provides modulation over nickel-based materials for the electrochemical UOR. By referring to the NiMoO₄ nanosheets (p-NiMoO₄) loaded on a nickel foam (NF), Tong et al. reported a defect engineering strategy for abundant oxygen vacancies (r-NiMoO₄), serving as the highly effective platform for the UOR performance (Fig. 6a).⁸⁹ Compared with the Ni foam, the growing of NiMoO₄ significantly enhances the UOR performance, indicating the catalytic active phase. With the formation of oxygen vacancies in NiMoO₄, a lower onset potential and a higher current density (249.5 mA cm⁻²) are observed for r-NiMoO₄ than p-NiMoO₄ (130.5 mA cm⁻²) (Fig. 6b). The DOS plot of r-NiMoO₄ (Fig. 6c) shows that the existence of oxygen vacancies downshifts the conduction band edge towards the Fermi surface related with p-NiMoO₄, illustrating the increased carrier concentration for promising transfer during the oxidation process. Correspondingly, the charge transfer resistance (R_ct) measured by electrochemical impedance spectroscopy (EIS) determines the trend of p-NiMoO₄ (4.8 Ω) > r-NiMoO₄ (2.6 Ω), confirming the facilitated electron transfer for a larger UOR current density of the r-NiMoO₄ sample. Using the S–S bond as the electron transfer bridge, Ji et al. reported Co, V co-doped NiS₂ as the ternary synergistic system (NCVS) for the electrochemical UOR.⁹⁰ Within the ternary system, it is hypothesized that the valence configurations of Ni²⁺, Co²⁺, V⁴⁺, and V⁵⁺ are with 3d⁸, 3d⁷, 3d¹, and 3d⁰, where the corresponding crystal field configurations are attached with t_2g⁶e_g², t_2g⁵e_g², t_2g¹e_g⁰, and t_2g⁰e_g⁰, respectively. By noting about the partial occupancy t_2g/e_g orbitals, the Ni–S–S–V model in the ternary system equips Ni²⁺ with the fully occupied t_2g state, which promotes the electron hopping towards the V site with partial occupied t_2g state through the S₂²⁻ bridge, defining the extended super-exchange interaction scheme for the modulation of the intrinsic property of Ni sites (Fig. 6d). In reference with the current density normalized by electrochemically active surface area (ECSA), the best NCVS-3 sample exhibits the largest current density, outperforming both NCS-6 and NVS-1, thus confirming the vital role of ternary synergism effect. Within the NCVS system, the UOR pathway mainly follows the intermolecular N–N coupling with the captured ¹⁴N¹⁵N in 1 M KOH and 0.33 M urea [CO(¹⁴NH₂)₂ + CO(¹⁵NH₂)₂]. The proposed reaction route is revealed in Fig. 6e, where the V site acts as the electron extractor from the Ni–Co pair to lead to the electrooxidation of the UOR. With the co-adsorption of urea molecules, the subsequent splitting of the C–N bond leads to the formation of CO and N₂ over the Co and Ni site, respectively, in which the further oxidation of CO into CO₃²⁻ produces the final products within the ternary system (Fig. 6f). Besides, with the incorporation of N into Ni₂P, Zhao et al. reported the interfacial modulated N–Ni₂P as the active phase for the electrochemical UOR.⁹¹ It is shown in Fig. 6g that N–Ni₂P requires the lowest potential of 1.347 V_RHE to achieve a current density of 100 mA cm⁻² compared with Ni₂P (1.375 V_RHE) and Pt/C (1.41 V_RHE), highlighting the most favorable kinetics. Notably, the NiOOH catalyst also presents a similar potential of 1.361 V_RHE as with N–Ni₂P despite the lower current density, indicating that the oxidative reconstruction of Ni₂P into NiOOH is responsible for the active UOR performance. Using NiOOH as the active model, the free energy change in Fig. 6h states that the γ-NiOOH phase possesses a stronger thermodynamic driving feasibility than β-NiOOH in the splitting of urea molecule into CO₂, where the RDS is assigned to be the desorption of *CO₂ derived from the energetics of ∼1242.2 kJ mol⁻¹ for the strong binding of the Ni³⁺ site reported. Particularly, due to the dual adsorption of *CONHNH₂ over γ-NiOOH rather than the O-terminal mode over β-NiOOH, the susceptible nucleophilic attack by OH⁻ facilitates the C–N cleavage, where the subsequent N modulation towards the lower adsorption of *CO₂ underscores the kinetic promotion of γ-NiOOH for the highly active UOR.

Fig. 6 (a) Schematic of the oxygen vacancy in r-NiMoO₄. (b) Polarized cyclic voltammetry (CV) curves for NiMoO₄ samples. (c) DOS curves of oxygen-defect NiMoO₄.⁸⁹ Copyright 2017, the American Chemical Society. (d) Schematic of the electronic coupling within the Ni, Co, and V system. (e) Polarization curves for NCVS-based samples; (f) proposed UOR mechanism over NCVS.⁹⁰ Copyright 2022, the American Chemical Society. (g) LSV curves of the Ni₂P system for the UOR. (h) Free energy profile of urea splitting over β-NiOOH and γ-NiOOH.⁹¹ Copyright 2026, Elsevier.

With the notation about the unique role of anion modulation, Geng et al. discovered the highly effective Ni₂Fe(CN)₆ catalyst by using –CN– as the bridge that connects the Ni and Fe sites for splitting the urea molecule.⁹² At a current density of 100 mA cm⁻², a lower potential of 1.35 V_RHE is observed for Ni₂Fe(CN)₆, much smaller than the OER of 1.68 V_RHE. The satisfactory UOR performance of Ni₂Fe(CN)₆ is attributed to the stable Ni²⁺, with the avoidance of the formation of NiOOH phase. Compared with Ni₃[Co(CN)₆]₂ and Fe₄[Fe(CN)₆]₃, the higher current density and lower potential are still obtained on Ni₂Fe(CN)₆, demonstrating the vital role of cooperative interaction within the Ni–CN–Fe site (Fig. 7a). By changing the urea concentrations in the KOH electrolyte, Ni₂Fe(CN)₆ shows the urea-independent behavior similar to most of the Ni-based catalysts, where the slope of scaling relation between log j and log C_urea is calculated to be 0.18 (Fig. 7b). Differently, with the changing KOH concentrations, Ni₂Fe(CN)₆ discloses the strong dependence (slope of 1.10) with respect to the OH⁻ concentration, where such pH dependence versus onset potential is determined to be ∼59 mV pH⁻¹, denoting the typical Nernstian-type behavior for the single H⁺/e⁻-coupled RDS (Fig. 7c). The UOR pathway is proposed to include two stages of chemical splitting of urea into CO₂ + NH₃ (CO(NH₂)₂ + H₂O → CO₂ + 2NH₃) and the oxidation of NH₃ into N₂ (2NH₃ + 6OH⁻ → N₂ + 6H₂O + 6e⁻), where the Ni and Fe sites undergo different catalytic processes. As shown in Fig. 7d, in the first stage of urea splitting, the Ni site exhibits more favorable transformation of *OCONH₂ into *NH₂ and CO₂ with a lower free energy change (0.90 eV) compared with the Fe site (1.02 eV). The reaction dynamics in Fig. 7e also supports the active trend for the C–N cleavage of Ni > Fe, where the Ni site presents a lower activation energy of 1.38 eV. In terms of the second stage, the Fe site facilitates the deprotonation of *NH₂ into *NH as the RDS, where a free energy change of 0.92 eV is assigned to the Fe site, which is lower than the 1.20 eV assigned to the Ni site. The corresponding activation energy also suggests a lower value of 1.29 eV (Fe site) related to the Ni site (1.33 eV) (Fig. 7f), confirming the promoted deprotonation kinetics for subsequent intermolecular N–N coupling at the Fe site. Similarly, Zhao et al. also reported the cyanide-bridged Ni–Fe bimetallic catalyst as Ni₂Fe(CN)₆ for the efficient UOR.⁹³ In their work, it is noted that the cyanide bridge serves as the electron transfer medium, supplying the stable Ni³⁺/Ni²⁺ redox shuttle with linked Ni and Fe sites. The synergism with the unique Ni–CN–Fe site assisted by the adsorption of *OH undergoes the advantageous consecutive dehydrogenation, thus promoting the multi-pathway for urea splitting.

Fig. 7 (a) Polarized LSV curves of distinctive nickel ferrocyanide catalysts in 1.0 M KOH that contains 0.33 M urea. (b) LSV curves of Ni₂Fe(CN)₆ in 1.0 M KOH with different urea concentrations, where the inset shows the fitted linear scaling for the UOR current density upon the concentration of urea. (c) Comparison of the UOR and OER performance, attached with the fitted scaling relation for the UOR current density upon the concentration of KOH. (d) Gibbs free energy profile of the UOR in the first stage. (e) Activation energy profile of C–N cleavage. (f) Activation energy profile of *NH₂ deprotonation.⁹² Copyright 2021, Springer Nature.

3.2. Localized modulation from a metal cation with a closed shell

Assuming Ni(OH)₂ or NiOOH as the active phase for the electrochemical UOR, the modulation for the Ni site also plays a vital role as the changes in the geometric structure contributes to both electronic structure and topology effects, which can further tune the adsorption behavior of intermediates. As indicated in the electronic structure of Ni(OH)₂, the localized 3d shell possesses the strong on-site coulomb interaction within the coordination with lattice O, thus necessitating the re-comprehension over the band structure theory. From the chemist's view, the band structure of six-fold metal–oxygen (M–O) coordination comprises a simple 3d-2p orbital overlap, leading to the formation of a bonding (M–O) band and an antibonding (M–O)* band separately (Fig. 8a).⁹⁴ Depending on the electronegativity difference between M and O, the energy gap between (M–O) and (M–O)* bands is defined as the charge transfer (Δ_CT). In terms of the Li-rich materials, the negligible overlap between Li 2s and O 2p orbitals leads to an increased degree of involvement of lattice oxygen atoms in thermal motion, which creates the additional non-bonding O states, also known as the state in the C_2v symmetry. The Lewis configuration of O²⁻ ions is composed of one 2s (|O_2s) and three 2p (|O_2p) orbitals, with the hybridization typically in an sp³ format with countable electron lone pairs as the nucleophilic site. Cationic substitution provides an effective strategy to increase the number of |O_2p electron lone pairs. With the incorporation of metal species with a closed valence shell such as alkali, alkali earth or 3d¹⁰ series into transition-metal-based oxides (TMOs), the degree of covalency involvement for the TM site can be largely reduced, as indicated by the accumulated electron localization function distribution around the Li and Na sites, which is contributed by the |O_2s and |O_2p orbitals. With the cationic vacancies derived from the leaching of Li⁺, Na⁺, and Mg²⁺ species, the anionic redox can be activated, which can be confirmed by the electron localization function (ELF) around the O lattice O sites (Fig. 8b).⁹⁵ During the charging process such as the oxidation reaction, the electron loss from the lattice O 2p states creates oxygen holes (h^O), resulting in the topological disordering at a high number of h^O for the formation of dimerization beyond the interaction within the 3d-2p orbital overlap (Fig. 8c), where the location of O–O dimerization is assigned to O–O π* above the Fermi level.⁹⁶

Fig. 8 (a) Quantitative molecular orbital diagram with the conventional band-structure of transition-metal oxides.⁹⁴ Copyright 2018, Springer Nature. (b) ELF diagrams obtained from the first-principles calculations for Li[Li_1/3M_2/3]O₂ (M = Mn, Ru, and Sn).⁹⁵ Copyright 2019, Springer Nature. (c) Schematic of the charging process for cations with strong TM–O π hybridization.⁹⁶ Copyright 2022, Springer Nature.

However, besides the contribution of non-bonding O 2p states, the removal of electrons from the mixed non-bonding O 2p and (M–O) bands can avoid the risk of structural destabilization, denoting the complex components within the (M–O)* band. To understand such behavior, the d–d coulomb interaction term U should be considered beyond the Δ_CT term, where the insight into the Mott–Hubbard splitting for correcting the failure of one-band view is typically not sketched by chemists but by solid-state physicists. In the Mott–Hubbard scheme, the splitting of the (M–O)* band forms the occupied lower Hubbard band (LHB) and the empty upper Hubbard band (UHB). Depending on the competition between U and Δ_CT terms, the electron loss during the charging process determines the ion redox property including the reversible structural distortion with O–O dimerization (U/2 ≈ Δ_CT) and irreversible release of O₂ with completely structural distortion (U >> Δ_CT) (Fig. 9a). One thing should be noted that the band position of non-bonding O 2p states largely affects the ion redox behavior within the Mott–Hubbard splitting scheme. The accumulation of O 2p states can be captured by adding the number of metal cations with a closed shell (e.g. Li⁺), in which the increased DOS of O 2p around the energy level of −1.5 eV in Fig. 9b for the Li–O–Li configuration confirms the non-bonding property that activates the degree of covalency involvement (Fig. 9b).⁹⁷ Similarly, such activation of non-bonding states can also be determined from the π[Mn–O] system (e.g. Na₄Mn₆O₁₄), as analyzed by Kitchaev et al.⁹⁸ As shown in Fig. 9c, the π system presents the D_3d symmetry, comprising two subsystems contributed by the Mn-d and O-p orbitals, where the single d orbital in the Mn site reveals the non-bonding t_2g level in the octahedral MnO₆ structure, while the p orbital in the O site provides the indirect σ bonding. With the symmetry-adapted linear combination of the D_3d group orbital sets, the total energy level diagram splits into a_1g, e_u, e_g, a_2u, a_1u and their corresponding antibonding states, in which the non-bonding states are composed of a_2u from the O site and a_1u from the metal site. By the projection of DOS over Mn and O sites, the predicted π states originate from non-bonding t_2g and e_g* states in the MnO₆ environments, where the highest occupied molecular orbital (HOMO) is assigned to be the anti-bonding e_g* state. As a result, the delocalized π[Mn–O] bonding condition creates the high-energy e_g* level, being electrochemically active during the charging process for structural distortions derived from the formation of oxygen dimers.

Fig. 9 (a) Schematic of the Mott-Hubbard splitting in the ion redox behavior during the charging process.⁹⁴ Copyright 2018, Springer Nature. (b) Gradual changing of O 2p states in the LiNiO₂ system, with the labile O 2p observed in the particular Li–O–Li configuration, where the inset displays the corresponding atomic models with different Li contents.⁹⁷ Copyright 2016, Springer Nature. (c) Symmetry analysis for D_3d with the π[Mn–O] redox center, the derived molecular orbital levels, and the projection of the Na₄Mn₆O₁₄ G₀W₀ DOS diagram.⁹⁸ Copyright 2021, the American Chemical Society.

Concerning the role of nucleophilic attack from the lattice O site, Han et al. reported the cationic vacancy controllable LiNiO₂ (LNO) model catalyst for the electrochemical UOR.⁹⁹ By a chemical delithiation method, various LNO-x catalysts were achieved with the specific ratio between Li and NO₂BF₄. As shown in Fig. 10a, the octahedral (O_h) NiO₆ structure presents the orbital components of a_1g, t_1u, e_g, t_2g and their anti-bonding states, in which the O t_2g states contributed by O sites guide the non-bonding area. Due to the absence of the interaction between Li and O, the electron density is accumulated onto the lattice O site for LNO-2 with a higher ratio of delithiation, denoting the activation of lattice oxygen (Fig. 10b). To be specific, the Ni²⁺ 3d⁸L configuration denotes the t_2g⁶e_g² term, while Ni³⁺ 3d⁷L configuration denotes the t_2g⁶e_g¹ term. Benefiting from the Mott–Hubbard splitting, the charge imbalance emerging from delithiation creates the rearrangement of electron density around the band edge below the Fermi level, leading to the formation of new non-bonding O bands (Fig. 10c). As a result, the downshifted LHB induced by the hybridization between Ni⁴⁺-t_2g orbitals and O 2p band provides mixing with non-bonding O 2p states, delivering the two-band redox behavior during oxidation conditions. With the addition of urea into a 1 M KOH electrolyte, the LNO-2 sample presents the sharply increased current density, which confirms the catalytic activity for the response of the electrochemical UOR behavior (Fig. 10d). After the 1000 cycles, the increase in anodic current density of 1.25 mA cm⁻² with a positive shift in the peak potential is displayed in comparison with the initial cycle, confirming the well-behaved durability of the LNO-2 sample, as shown in Fig. 10e. Concerning the activation of lattice oxygen of LNO, the most stable adsorption configuration of urea molecule is determined to be the bridge coordination between C and lattice O, where the lattice O behaves as the nucleophilic site. As shown in Fig. 10f, the distinction between the AEM and LOM pathways for LNO-0 is assigned to the coupling between *CONNH₂ and lattice oxygen to form the oxygen vacancy. Compared with the deprotonation of *CONNH₂ into *CONNH via the AEM pathway, such deprotonation with the formation of oxygen vacancies is exothermic via the LOM pathway, denoting its thermodynamic advantage. With the delithiation, the LNO-2 sample shows the facilitated electrochemical deprotonation for the urea molecule along with the accelerated electron transfer. Within the LOM pathway, the LNO-2 sample reveals the favorable intramolecular N–N coupling with the release of CO₂ and N₂. After the electrochemical transformation of urea with the sacrifice of oxygen vacancies, the recovery of OH⁻ into the vacancy completes the catalytic loop for LNO-2.

Fig. 10 (a) Qualitative molecular orbital diagram for the octahedral NiO₆ structure. (b) Charge density difference of LNO-0 and LNO-2, respectively. (c) Diagram of the Mott–Hubbard splitting in the LNO system with PDOS profiles of LNO-0 and LNO-2. (d) LSV curves of the UOR for LNO-2. (e) Polarization curves before and after 1000 cycles for LNO-2, where the inset delivers the magnified polarization curves for the localized potential window. (f) Comparison of the free energy profiles between the AEM and LOM pathways during the UOR for LNO-0, attached with localized change of intermediates.⁹⁹ Copyright 2022, Wiley-VCH GmbH.

3.3. Localized modulation from transition metal

The insight obtained from the above-mentioned band structure theory indicates that the unique interaction within d electrons plays a vital role in the modulation of the electrochemical UOR. With the incorporation of tungsten (W) into the nickel catalyst, Wang et al. reported the highly effective Ni–WO_x catalyst for the electrochemical UOR.¹⁰⁰ As shown in Fig. 11a, the Ni–WO_x sample requires a lower potential of 1.40 V to reach a larger current density of 100 mA cm⁻², presenting the downshifted 273 mV potential related to the OER. Concerning the advantage of the UOR in replace of the OER, the electrochemical performance was further applied in a flow electrolyzer (Fig. 11b). With the assistance of cathodic CO₂ reduction, the electrochemical performance only needs a cell potential of 2.16 V to operate at a current density of 100 mA cm⁻² for the formation of CO and urea spitting, which is significantly lower than the use of the OER as the anodic reaction (a downshift of 370 mV) (Fig. 11c). It is concluded that the well-behaved UOR performance originates from the unique electron transfer between W and Ni atoms within Ni–WO_x. With the occurrence of the UOR over the directly exposed Ni site, the adsorption of the C=O group mainly locates at the negatively charged region (Ni site), while the amino group –NH₂ prefers the adsorption over the W site, thus favoring the splitting of urea molecules through polarization (Fig. 11d). Similarly, Liu et al. proposed the self-supported W-doped Ni–C₃S₃N₃ coordination polymer (W–NT) as the highly effective catalyst for the electrochemical UOR.¹⁰¹ With the tailoring of active Ni³⁺ sites via ligand anchoring and high-valence metal doping, the lower potentials for driving the UOR require only 1.39 V and 1.43 V at current densities of 50 mA cm⁻² and 100 mA cm⁻², compared with the OER outcome of 1.63 V and 1.71 V. The lowering of the d-band center for the Ni site with the doping of W confirms the chemical reduction tendency, which helps the stabilization of the appropriate valence state of the Ni³⁺ site and the optimization of the adsorption strength towards urea molecules (Fig. 11e). Benefiting from the W doping, Ni³⁺ reveals a preferable adsorption energy of −0.70 eV for the urea molecule compared with H₂O of −0.29 eV, denoting the favorable capture of urea for consecutive transformation (Fig. 11f). During the electrochemical UOR progress, the potential-determining step (PDS) is the deprotonation of *NH₂NH₂ into *NHNH₂ with an energy gap of 0.83 eV, significantly lower than the PDS of formation of *OHOH in the OER (1.71 eV). Besides, the construction of dual active sites of Ni–WO₃ is discovered to regulate the adsorption of *COO intermediates for urea splitting.¹⁰² Benefiting from the electron transfer from Ni to WO₃, the adsorption of urea exhibits the dual binding of the –NH₂ group and C=O group over Ni and WO₃ simultaneously, where such bridge adsorption mode of Ni–NCO–W lowers the C–N cleavage barrier, promoting the consecutive transformation of *COO over the WO₃ site and the formation of N₂ at the Ni site, enabling the coupled desorption of CO₂ and N₂ effectively. With the construction of the Ni–O–Mn unit site, Sun et al. reported the highly effective self-supported NiMn metal organic framework (NiMn–MOF) loaded with ultrafine Pt nanocrystals as Pt_NC/NiMn–MOF for the electrochemical UOR.¹⁰³ Within the Ni–O–Mn unit site, the electronic repulsion in high-spin Ni–O enhances the π-symmetry between Mn³⁺ and O²⁻, which favors the electron transfer throughout the catalyst. The introduction of Pt_NC further regulates the charge redistribution, thus optimizing the chemisorption behavior to reduce the energy barrier of RDS.

Fig. 11 (a) Comparison of the polarization curves of Ni–WO_x with and without 0.33 M urea. (b) Schematic of the coupling between the UOR and the CO₂R for cell electrolyzers. (c) Polarizations curves for the UOR|CO₂R electrolyzer compared with OER|CO₂R. (d) Adsorption mode of urea on Ni–WO_x with calculated charge density difference.¹⁰⁰ Copyright 2021, Wiley-VCH GmbH. (e) Proposed mechanism of the UOR over the W–NT system. (f) Free energy profile of W–NT in terms of the UOR and OER.¹⁰¹ Copyright 2023, Wiley-VCH GmbH.

By noting about the unique role of W, Cat et al. also reported the construction of single-atom W-doped nanoporous P-Ni(OH)₂ catalysts as np/W–P-Ni(OH)₂ to lead the chemical-electrochemical coupled pathway for the UOR.¹⁰⁴ As shown in Fig. 12a, the conventional pathway for Ni-based catalysts follows the self-oxidation reaction to create the NiOOH phase, allowing for the formation of key intermediates including *CONHN and *CON₂, where the RDS is assigned to the formation of CO₂ with a higher overpotential. Differently, with the proceeding of the chemical-electrochemical coupled pathway, np/W–P-Ni(OH)₂ enables the almost thermoneutral lattice hydroxyl dehydrogenation of Ni(OH)₂ to form the Ni(OH)O phase, thus leading the highly effective pathway with a low energy barrier. As revealed by the high-angle annular dark-field scanning transmission electron microscopic (HAADF-STEM) image, the np/W–P-Ni(OH)₂ presents the core–shell-like morphology with the atomic distribution of W in the shell (Fig. 12b). With the electrochemical measurement in 1 M KOH with 0.33 M urea, the np/W–P-Ni(OH)₂ sample shows the lowest Tafel slope of 52.6 mV dec⁻¹, demonstrating the most favorable kinetic process of the UOR (Fig. 12c). The practical display in hydrogen production with the coupled UOR system further states the decreased cell potential of 316 mV, compared with the OER performance, thus denoting the promising application (Fig. 12d). To unveil the dynamic kinetics of the UOR, potential-dependent Bode plots in Fig. 12e exhibit that the doping of W into np/P-Ni(OH)₂ shifts the phase peak of urea oxidation towards higher frequencies, which is indicative of the enhanced electron transfer property. With the notation about W-induced favorable kinetics, the calculated PDS of W–P-Ni(OH)₂ is the adsorption of urea with the free energy change of 1.176 eV, lower than the PDS of the dehydrogenation of urea into *CONHNH₂ (1.94 eV) for pure Ni(OH)₂, as shown in Fig. 12f. After the formation of *NH₂ (Fig. 12g), the subsequent PDS for W–P-Ni(OH)₂ is the deprotonation of *NHOH into *NHO with a free energy gap of 1.045 eV, lower than the PDS of pure Ni(OH)₂ as the deprotonation of *NHO into *NO. Besides the incorporation of W, some commonly used transition metals have exerted analogous modulation effects. Xie et al. developed electron-delocalized Ni–Co active pairs to achieve efficient UOR performance by doping Co into Ni(OH)₂ and integrating with the CoNi alloy.¹⁰⁵ Electron-delocalized Ni–Co sites within the catalyst lead to the upshifted d-band center, orbital hybridization and electron transfer between the electron delocalized Ni–Co sites and the N atoms in urea are significantly enhanced, which decreases the electron transfer from H to N in the N–H bond and weakens the N–H bond. As demonstrated by in situ characterizations and theoretical investigation, the RDS of the UOR is *CONH₂NH₂ → *CONH₂NH, and Ni^2+δ–O–Co^2+δ delivers a rate-determining step with an energy barrier reduced to 1.58 eV, far superior to that of pure Ni(OH)₂.

Fig. 12 (a) Schematic of various UOR pathways. (b) HAADF-STEM image of the np/W–P-Ni(OH)₂ sample. (c) Calculated Tafel plots for the electrochemical UOR. (d) Overall performance of hydrogen production via the UOR‖HER system. (e) Potential-dependent Bode plots of np/W–P-Ni(OH)₂ for the UOR. (f and g) Free energy profiles of pure Ni(OH)O and W–P-Ni(OH)₂ along the UOR progress.¹⁰⁴ Copyright 2025, the Royal Society of Chemistry.

To realize the transformation of urea into innocuous N₂ for sustainability rather than the formation of NH₃ as the acidic rain resource, Zhan et al. proposed the atomic anchoring of Ni over a Ti foam system with the synthesis of an asymmetric Ni–O–Ti site.¹⁰⁶ As shown in Fig. 13a, it is observed that the atomic Ni site is distributed onto the surface of the TiO_x substrate, where the atomic dispersion is further intensified via the elemental mapping. With the determination of the coordination environment of the asymmetric Ni–O–Ti site (Fig. 13b), the Fourier-transformed extended X-ray absorption fine structure spectroscopy (FT-EXAFS) exhibits the characteristic peak at 1.59 Å of the Ni–O shell, where the absence of Ni–Ni coordination further confirms the atomic dispersion of the Ni site into TiO_x. During the electrochemical UOR process, the asymmetric Ni–O–Ti sites disclose the low potentials of 1.30 V and 1.33 V for the current densities of 10 mA cm⁻² and 100 mA cm⁻², respectively. Meanwhile, Ni–O–Ti sites exhibit a Tafel slope of only 14.2 mV dec⁻¹, notably smaller than 20.0 mV dec⁻¹ for the Ni–O–Ni sites and 20.8 mV dec⁻¹ for the Ni foam. All the electrochemical performance outperforms both of the symmetric Ni–O–Ni sites and the Ni foam. Within a potential range of 1.40 to 1.70 V, the Ni–O–Ti sites still maintain a selectivity of 99% for N₂ evolution, which is consistent with the above-mentioned electrochemical performance result. By feeding the equivalent ratio of CO(¹⁴NH₂)₂ and CO(¹⁵NH₂)₂, the observed mass spectral distribution in online mass spectrometry witnesses the ¹⁴N₂ : ¹⁵N₂ : ¹⁴N¹⁵N ratio of 1 : 1 : 0, indicating the dominant intramolecular N–N coupling over the asymmetric Ni–O–Ti sites (Fig. 13c). To elucidate the electrochemical UOR pathway over asymmetric Ni–O–Ti sites, the in situ FTIR spectroscopy in Fig. 13d shows that the adsorption of urea onto asymmetric Ni–O–Ti sites induces a blueshift of the ν_a() mode compared with that onto Ni–O–Ti sites, indicating strengthening of the bond. Concerning the oxophilicity of the Ti site, such chemisorption of urea can be accounted by the unique configuration of the C=O → Ti format. The lack of peaks attributed to NCO⁻ or NO_x⁻ over the asymmetric Ni–O–Ti site throughout the bias potentials determines the high N₂ selectivity as the clean products. As shown in Fig. 13e, the urea adsorption over Ni–O–Ti manifests the obvious electron accumulation for two bonds, which accounts for the facilitated interactions of the C=O group in comparison with the Ni–O–Ni site. Upon the adsorption of urea onto the asymmetric Ni–O–Ti site, the intramolecular N–N is favored during the initial deprotonation process with the energy change of 0.41 eV, while the breaking of the bond involves a higher energy change of 1.03 eV (Fig. 13f). The production of N₂ with the minor energy change (0.13 eV) via the subsequent deprotonation accounts for the high selectivity of the N₂ product.

Fig. 13 (a) Aberration-corrected HAADF-STEM image of the Ni–O–Ti site with the marked Ni site and elemental mappings. (b) FT-EXAFS spectra for the Ni–O–Ti site. (c) N₂ contents determined by online mass spectrometry over the asymmetric Ni–O–Ti site. (d) In situ FTIR spectra for the electrochemical UOR process with the changing bias potential. (e) Charge density differences for the adsorption of urea over the Ni–O–Ti site and Ni–O–Ni site. (f) Free energy profile for the UOR process on the asymmetric Ni–O–Ti and symmetric Ni–O–Ni sites.¹⁰⁶ Copyright 2024, Springer Nature.

3.4. Localized modulation from rare-earth dopants

The rare-earth species are comprised of Sc, Y, and the whole lanthanide series (ranging from La to Lu), revealing similar physical and chemical properties. The beauty of rare-earth species lies in that the subtle quantity alters the material property, which are capable of serving as the vitamin dopants in the modern industry. In terms of lanthanide series, the valence configuration shares the typical state of 4f^x5d⁰⁻¹6s² (x = 1–14). Despite the lower energy of 5s/5p orbitals compared with the 4f orbitals, the occupancy of dispersive 5s/5p orbitals allows for the outside location, which contributes to the weak shielding of 4f electrons.^107–109 Benefiting from the intra-atomic 4f–4f electron transitions and the weak shielding effect from 5s/5p electrons, the insensitive involvement with surrounding molecules for rare-earth species behaves mostly as the electron modulators.^110–112 Additionally, with the property of 4f-shell-induced radial shrinkage, rare-earth species exhibit the strong spin–orbit coupling and weak ligand field splitting, thus disclosing the unique magnetic properties.¹¹³

Based on the unique role of rare-earth species, Rao et al. initially discovered the direct mechanism of electrochemical UOR by using NdNiO₃ as the catalyst.¹¹⁴ In the direct mechanism, the typical NiO sample delivers the interconversion of Ni²⁺/Ni³⁺ as the crucial step for the oxidation of the urea molecule alongside the charge transfer step. However, with the incorporation of Nd³⁺, the retention of the high-valence NiOOH phase without the reduction into Ni(OH)₂ is observed, providing the sufficient oxidative capacity at the Ni³⁺ site. Benefiting from the exposure of Ni³⁺ sites in NdNiO₃, the enhanced adsorption of OH⁻ is also favored, thus accounting for the well-retained formation of NiOOH. Afterwards, the formation of a NdNiO₃–NiO heterointerface was also unveiled as interface-driven promotion of the electrochemical UOR.¹¹⁵ The construction of such heterointerfaces enhances the activation of the NiOOH phase for better UOR performance, where the modulated interface of charge distribution optimizes the balance between the adsorption of urea and the desorption of *CO₂ with the assistance of strong OH⁻ adsorption. Intrinsically, the scheme of d–p–f gradient orbital coupling framework has been constructed to supply the comprehension over the rare-earth-induced system for heterogeneous catalysis.^116–118 Taking octahedral and tetrahedral structures as the prototype, the metal–oxygen interaction allows for the formation of non-bonding t_1g/t_2u states in the octahedral site and t₁ states in the tetrahedral site.¹¹⁹ With the incorporation of rare-earth into the octahedral site, the low covalency is able to further induce the weak t_1u/t_2u bonding states beyond, which is promising to activate the lattice oxygen as the nucleophilic site for modulating the UOR performance.^120,121

With the notation about the gradient orbital coupling, Qiang et al. reported the construction of a La-incorporated 3D ordered macroporous NiO heterostructure (3DOM La₂O₃–NiO) catalyst for the electrochemical UOR.¹²² As shown in Fig. 14a, the La element shares the unique configuration of 4f⁰ after the electron loss of valence 6s² states. The gradient orbital coupling within the La–O–Ni site can be constructed based on the covalency for Ni–O with bonding and antibonding (Ni–O)* bands, where the σ and π conjunction is contributed by the O-sp orbitals. Upon the construction of a La–O–Ni gradient orbital coupling framework, the charge transfer between La₂O₃ and NiO can be facilitated with the thermo-balance through the Fermi level leverage, which can be derived from the work function of La₂O₃ (2.70 eV) versus NiO (3.84 eV). The morphology of La₂O₃–NiO in Fig. 14b reveals the even uniformity of good size and abundant porosity with pore wall, providing the favorable mass transport condition during heterogeneous catalysis. To further understand the electronic interaction between La₂O₃ and NiO, the charge density difference in Fig. 14c presents the observable electron transfer from La to Ni through the La–O–Ni bridge. For 3DOM La₂O₃–NiO, the well-behaved electrochemical performance reveals voltages of 1.24 V, 1.30 V, and 1.38 V to approach current densities of 10 mA cm⁻², 50 mA cm⁻², and 100 mA cm⁻², respectively. Benefiting from the advantageous performance of 3DOM La₂O₃–NiO, the practical measurement was performed using a membrane electrode assembly (MEA)-based anion exchange membrane water splitting (AEMWE) device with the electrolyte of 1 M KOH + 0.33 M urea solution (Fig. 14d). The proposed reaction mechanism on La–O–Ni bridge is termed the high-valence nickel mechanism (HNM) (Fig. 14e). To be specific, the electrochemical UOR process proceeds through the adsorption of urea onto the Ni site, which is followed by the gradual cleavage of N–H bonds, thus favoring the formation of the *OCNN intermediate. It is determined that the PDS of UOR is assigned to be the cleavage of the N–H bond of *OCNH₂NH into *OCNH₂N with a free energy gap of 1.22 eV, which is larger than the 1.04 of La₂O₃–NiO for the PDS of *OCNH₂N → *OCNHN (Fig. 14f). Additionally, Zheng et al. demonstrated a Ce-doped Ni₃S₂ catalyst loaded onto a Ni foam for the highly effective UOR.¹²³ With the construction of gradient orbital coupling of Ce 4f-S 3p-Ni 3d, the d-band center of Ni site is downshifted from −2.46 eV to −3.03 eV, thus weakening the adsorption of *CO during the UOR. With the Ce doping, the corresponding energy barrier for the dehydrogenation of urea is assigned to be 0.22 eV, which is significantly lower than that for the oxidation of *CO as the RDS (1.68 eV) for the counterpart. Similarly, Zhang et al. also reported a Ce-induced NiS catalyst (Ce–NiS) for the highly-effective UOR.¹²⁴ By tuning the Ce loading content, the performance follows the trend of 3% Ce–NiS > 5% Ce–NiS > 1% Ce–NiS, in which the best-behaved Ce–NiS presents a high faradaic efficiency of 91.39% compared with NiS of 67.52% for hydrogen production. Furthermore, the electrochemical activation energy of the UOR delivers the decreased result of 8.72 kJ mol⁻¹ to 5.68 kJ mol⁻¹ with Ce doping, also confirming the promotive role of Ce.

Fig. 14 (a) Qualitative molecular orbital diagram of gradient orbital coupling within the La₂O₃–NiO system. (b) SEM image of La₂O₃–NiO at a calcination temperature of 450 °C. (c) Charge density difference of La₂O₃–NiO, where the yellow and cyan areas represent the electron accumulation and depletion. (d) Electrochemical stability of the UOR coupled with the HER at a current density of 1 A cm⁻², where the inset shows the cell model for the urea-assisted hydrogen production. (e) Proposed UOR mechanism over La³⁺–O–Ni^δ+(δ ≥ 3). (f) Gibbs free energy profile under the standard condition of the UOR process for La₂O₃–NiO and NiO surfaces, where the inset shows the atomic model for the catalyst.¹²² Copyright 2025, Wiley-VCH GmbH.

By noting about the unique role of asymmetric Ce–O–Ni sites, Liu et al. reported the self-supported construction of Ce-doped Ni–MOF catalyst (Ce–Ni–BDC) with the assistance of benzene-1,4-dicarboxylic acid as the ligand (Fig. 15a).¹²⁵ Under the solvothermal growth of Ni–BDC over the Ni foam, the prepared nanosheets provide the fluffy and biscuit-like structure with abundant interlayers of ∼25 nm that favor the ion exchange of Ce³⁺ into Ni–BDC (Fig. 15b). The atomic coordination of Ce–Ni–BDC states the extended Ni–O bond length compared with Ni–BDC, which can be attributed to the formation of the Ce–O–Ni site, as shown in Fig. 15c. Accordingly, the Ce–O bond length shows subtle extension compared with CeO₂ at ∼1.87 Å, owing to the weaker bond strength with the formation of the Ce–O–Ni site (Fig. 15d). The electrochemical performance (Fig. 15e) undergoes a remarkable negative shift of potential of 268 mV in 1 M KOH + 0.5 M urea compared to 1 M KOH electrolyte, Ce–Ni–BDC exhibits only 1.261 V at 10 mA cm⁻² and a Tafel slope of 44.33 mV dec⁻¹, while Ni–BDC presents a higher potential of 1.345 V at 10 mA cm⁻² and a larger Tafel slope of 73.32 mV dec⁻¹, illustrating the enhancement of UOR activity through the formation of the asymmetric Ce–O–Ni site. By examining the in situ Raman spectroscopy, the distinctive peak of the symmetric C–N stretching vibration mode at 1009 cm⁻¹ confirms the stable adsorption and transport of urea molecules from 1.2 to 1.7 V (Fig. 15f). Concerning the oxidative environment, the newly formed peaks at 460 and 578 cm⁻¹ can be attributed to Ce-γ-NiOOH at potentials over 1.3 V, accompanied by the disappearance of Ce–O–Ni signals. For the Ni–BDC sample, the emergence of peaks at 460 and 563 cm⁻¹ stems from γ-NiOOH at a potential of 1.45 V, denoting the slower reconstruction process and higher required potential. With the formation of active Ce-γ-NiOOH phase, the Ce site promotes the obvious charge transfer towards Ni sites, which can enhance the UOR performance (Fig. 15g). To be specific, the electron transfer from Ce favors the formation of asymmetric Ce–O–Ni within Ce-γ-NiOOH, which is consistent with the spectra result, as indicated by Fig. 15h. With the further incorporation of the Ce–Ni–BDC catalyst into a direct Zn–urea battery similar to the Zn–air battery, the anode reaction proceeds through the electrochemical urea splitting into CO₂ and N₂ in replace of the OER (Fig. 15i). Under the periodic shifts of current density, the galvanostatic discharge–charge profile shows a flexible response, indicating promising practical application (Fig. 15j). Similarly, Wang et al. unveiled the industrial-level UOR performance by supplying Ce-doped Ni₅P₄ (Ce–Ni₅P₄) as the catalyst.¹²⁶ The introduction of Ce into Ni₅P₄ facilitates the electron transfer from P and Ni towards the Ce site, in which the increased chemical state of Ni site boosts the reconstruction into NiOOH with a higher UOR kinetics, thus decreasing the energy gap of RDS and the deprotonation of *CONHN into *CONN. Besides the use of Ce, the incorporation of La into β-Ni(OH)₂ as La:β-Ni(OH)₂ exhibits the fast reaction kinetics, where the electronic interaction between La and Ni ions is delivered to accelerate the electron transport for electrical conductivity, thus adjusting the electrochemical UOR performance.¹²⁷

Fig. 15 (a) Schematic of the Ce–Ni–BDC loaded on a Ni foam. (b) SEM image of Ni–BDC. (c) Ni K-edge spectra of Ce–Ni–BDC and Ni–BDC. (d) Ce L-edge spectra of Ce–Ni–BDC with reference to standard CeO₂. (e) Polarization curves of Ce–Ni–BDC for the UOR and OER. (f) Potential-dependent in situ Raman spectra of Ce–Ni–BDC during the UOR. (g) Charge density difference for Ce-γ-NiOOH, where cyan and purple regions represent the electron accumulation and depletion. (h) Proposed electron transfer over Ce–Ni–BDC. (i) Schematic of the Zn–urea battery. (j) Galvanostatic discharge/charge curves at various current densities.¹²⁵ Copyright 2025, Wiley-VCH GmbH.

4. Summary and outlook

In this review, the local hetero-ion modulation over Ni-based catalysts has been systematically focused for the electrochemical UOR, where the hetero-ion species range from the main group, transition-metal block to the rare-earth block. Beginning with the discussion of various pathways over Ni-based catalysts, the progress of urea splitting into CO₂ and N₂ can be triggered with the assistance of molecular N–N coupling, nucleophilic attack by O species, and adsorbate evolution. From these distinctive pathways, it can be witnessed that the dynamic phase transformation of Ni-based catalysts shares the vital role from the chemical redox behavior among oxidative Ni^δ+ species. To lead the intrinsic comprehension over the complex UOR system, the universal scaling relationships have been constructed based on various energy descriptors. The intrinsic insights into electronic structure modulation by hetero-ion species over various Ni-based catalysts are exclusively provided to understand the structure–performance relation when designing highly effective Ni-based catalysts. Undoubtedly, these inspiring progresses pave the way for future development in electrochemical urea splitting in replace of highly energy-consuming OER and in the agricultural pollution treatment. To guide the future research and echo for the advancement of data-driven catalysis, a novel outlook is provided based on a closed-loop framework for the designing and screening of promising catalysts for the UOR (Fig. 16). Notably, although hetero-ion modulation provides an effective regulatory strategy for electrochemical urea oxidation, the UOR still faces stability issues. Especially, during the long-term stability test, hetero-ion modulation exhibits no relatively pronounced improvement since it is prone to agglomeration and reconstruction, which severely affects the stability of Ni-based catalysts. Therefore, it is still necessary to develop modification design strategies to promote the stability of Ni-based catalysts in electrochemical urea oxidation.

Fig. 16 Closed-loop framework for future investigation of UOR: data-driven science, theoretical investigation, catalyst synthesis and application, and advanced physical characterizations.

4.1. Data-driven science and theoretical investigation

With the exploit of the abundant data resources reported, the characteristic samples for the UOR can be obtained for data compact as the reference source. By focusing on these resources, the key features of experimental data can be further reorganized according to the vectorized properties by large language models (LLMs) to construct the frontier database. During the input of LLMs, the natural language orders take the effect combined with the user plate and typical derivation strategies from artificial intelligence, noted as the prompt engineering (e.g. zero-shot, one-shot, few-shot, and chain-of-thought). Afterwards, the theoretical screening for catalysis can be well extended based on these vectorized training datasets with the assistance of machine learning (ML), including the coordination number around Ni (CN_Ni), the electronegativity of dopants (χ_d) with contents, mean-field adsorption energy of Ni surface (E_m), and coverage of adsorbates (θ), constructing the ML-driven relation like E_m ← αCN_Ni+ βχ_d + γΣ_iθ_i. To reach out for the high accuracy, the first-principles framework should be proceeded over the critical catalytic descriptors for the UOR. Thus, with the input of first-principles data resources, the training for machine learning potential (MLP) further provides the radical dismantlement for the dimension and time series limitations of first-principles framework.

4.2. Catalyst synthesis and practical application

With the guidance of data science, the catalyst synthesis can be purposeful by referring to the general properties of target materials for the UOR. For instance, the catalyst synthesis can be realized using tailored preparation strategies, such as solid-state reaction, solvothermal reaction, molecular assembly reaction, and atomic layer reaction under corresponding reaction conditions, which aims at the intrinsic synergism between Ni and other active metal sites. After the catalyst synthesis, the instant electrochemical measurement of the UOR can be accessed through the multi-functional electrochemical lab, in terms of polarized curves, Tafel slopes, electrochemical impedance, and so forth. Based on the screened catalysts, the practical application can be achieved in comparison to the energy consumption related to the OER for validation. Generally, the electrochemical screening for performance offers the experimental validation and the materials resources for the physical characterizations, which is also the necessary feedback for the theoretical guidance. Besides, since the metal site typically acts as the main active site, the electrochemical screening of the metal dissolution should also be performed to monitor the catalyst stability for the high-performance UOR.

4.3. Advanced physical characterizations

For advanced physical characterizations, catalyst samples are collected both before and after the UOR for subsequent physical characterizations to investigate the phase and surface transformation. To be specific, the combination of X-ray-derived spectra and surface information detection provides the pre-catalyst view for the basic properties of the obtained catalyst samples. Concerning the dynamic surface changes during the UOR process, the operando techniques give instant signals of surface coverage of adsorbates and phase composition with the shifting of bias potentials. For urea oxidation, conversion into N₂ and CO₂ induces the complex surface state change to influence the catalytic behavior for the transformation of adsorbed N_xH_y species. With the capture of these subtle signals, the theoretical modelling can be promisingly rationalized to lead the consistence between time and spatial dimensions for electrochemical urea splinting. In terms of the subtle signal from the metal coordination, its potential-dependent shift should also be determined to analyze the metal dissolution beyond the adsorption and desorption of chemical intermediates during the UOR.

Notably, the establishment of the closed-loop framework can significantly accelerate the discovery of high-performance UOR catalysts and sharpen the microscale comprehension of reaction mechanisms, enabling more efficient exploration of structure–activity relationships and dynamic catalytic processes.

Author contributions

All of the authors contributed to the literature search writing this review.

Conflicts of interest

The authors declare no competing interests.

Data availability

The authors confirm that the data supporting this review are available on request.

Acknowledgements

This work was financially supported by the National Natural Science Foundation of China (22279062), the Basic Research Program of Jiangsu (BK20250033), the 78th General Funding Program of the China Postdoctoral Science Foundation (211090B62508) and JSPS KAKENHI (JP25K01737). The authors are grateful for the support from the National and Local Joint Engineering Research Center of Biomedical Functional Materials and the Priority Academic Program Development of Jiangsu Higher Education Institutions. T. Lu acknowledges the Center for Computational Materials Science, Institute for Materials Research, Tohoku University, for the use of MASAMUNE-IMR (No. 202512-SCKXX-0208) and the Institute for Solid State Physics (ISSP) at the University of Tokyo for the computational resources.

References

1. S. Rekker, G. Chen, R. Heede, M. C. Ives, B. Wade and C. Greig, Nat. Clim. Change, 2023, 13, 927–934.
2. R. Wang, W. He, Y. Xue, Y. Yu and B. Liu, Nat. Food, 2026, 7, 27–37.
3. T. J. Wallington, M. Woody, G. M. Lewis, G. A. Keoleian, E. J. Adler, J. R. R. A. Martins and M. D. Collette, Joule, 2024, 8, 2190–2207.
4. H. Lin, Z. Sui, F. Li and W. Wu, Matter, 2025, 8, 102293.
5. Y. Yang, Y. Cheng, J. P. H. Poon, Y. Zhou, X. Qian, S. Xia, X. Li, J. Xue, L. Zhang and H. Zhang, Nat. Sustain., 2025, 8, 1223–1233.
6. H. Zhang, Y. Zuo and J. Huang, Chem. Synth., 2025, 5, 52.
7. J. Xu, G. Zhang, S. Fang, J. Li, X. Xiao, Y. Ye, Z. Zhang, R. Xu, J. Zhou, Y. Wu, W. Huang and Y. Cui, Nat. Chem. Eng., 2026, 3, 160–168.
8. K. R. Bassett, S. F. Hupperts, S. Jämtgård, L. Östlund, J. Fridman, S. S. Perakis and M. J. Gundale, Nature, 2026, 650, 629–635.
9. Y. Zhang, Nat. Rev. Chem., 2017, 1, 0057.
10. W. Feng, B. Chang, Y. Ren, D. Kong, H. B. Tao, L. Zhi, M. A. Khan, R. Aleisa, M. Rueping and H. Zhang, Adv. Mater., 2025, 37, 2416012.
11. M. Liu, W. Zou, S. Qiu, N. Su, J. Cong and L. Hou, Adv. Funct. Mater., 2024, 34, 2310155.
12. G. Zhan, L. Hu, H. Li, J. Dai, L. Zhao, Q. Zheng, X. Zou, Y. Shi, J. Wang, W. Hou, Y. Yao and L. Zhang, Nat. Commun., 2024, 15, 5918.
13. L. Miao, W. Jia, X. Cao and L. Jiao, Chem. Soc. Rev., 2024, 53, 2771–2807.
14. Z. Li, L. Sun, Y. Zhang, Y. Han, W. Zhuang, L. Tian and W. Tan, Coord. Chem. Rev., 2024, 510, 215837.
15. K. Dashtian, S. Shahsavarifar, M. Usman, Y. Joseph, M. R. Ganjali, Z. Yin and M. Rahimi-Nasrabadi, Coord. Chem. Rev., 2024, 504, 215644.
16. N. Du, C. Roy, R. Peach, M. Turnbull, S. Thiele and C. Bock, Chem. Rev., 2022, 122, 11830–11895.
17. Y. Gao, H. Yu, W. Wang, P. Chi, S. Liu, H. Rao, C. Bian and J. Ge, Joule, 2026, 3, 102337.
18. C. R. Wang, J. M. Stansberry, R. Mukundan, H.-M. J. Chang, D. Kulkarni, A. M. Park, A. B. Plymill, N. M. Firas, C. P. Liu, J. T. Lang, J. K. Lee, N. E. Tolouei, Y. Morimoto, C. H. Wang, G. Zhu, J. Brouwer, P. Atanassov, C. B. Capuano, C. Mittelsteadt, X. Peng and I. V. Zenyuk, Chem. Rev., 2025, 125, 1257–1302.
19. C. Rong, K. Dastafkan, Y. Wang and C. Zhao, Adv. Mater., 2023, 35, 2211884.
20. Z. Chen, Q. Fan, J. Zhou, X. Wang, M. Huang, H. Jiang and H. Cölfen, Angew. Chem., Int. Ed., 2023, 62, e202309293.
21. H. Li, Y. Lin, J. Duan, Q. Wen, Y. Liu and T. Zhai, Chem. Soc. Rev., 2024, 53, 10709–10740.
22. F. Wang and S. S. Stahl, Acc. Chem. Res., 2020, 53, 561–574.
23. C. Liu, F. Chen, B.-H. Zhao, Y. Wu and B. Zhang, Nat. Rev. Chem., 2024, 8, 277–293.
24. H. Sun, X. Xu, L. Fei, W. Zhou and Z. Shao, Adv. Energy Mater., 2024, 14, 2401242.
25. Q. Zhu, J. Xiao, C. Deng, T. Huang, H. Ding, L. Zhang and G. Xu, Chin. J. Struct. Chem., 2025, 44, 100651.
26. N. Zhao, N. Zhang, K. Liu, S. Liu, Y. Wang, R. Jia, Y. Dai, M. Xue and G. Zhao, Nano Res. Energy, 2025, 4, e9120206.
27. Q. Li, J. Wang, Y. Shi, H. Li, H. Yang, K. Xiang, W. You and J. Liang, Mater. Horiz., 2025, 12, 9952–9965.
28. X. Gao, S. Zhang, P. Wang, M. Jaroniec, Y. Zheng and S.-Z. Qiao, Chem. Soc. Rev., 2024, 53, 1552–1591.
29. X. Wang, J.-P. Li, Y. Duan, J. Li, H. Wang, X. Yang and M. Gong, ChemCatChem, 2022, 14, e202101906.
30. P. Qiao, G. Li, X. Xu, D. Wang, F. Wang, L. Xu, L. Lu, H. Cong and M. Sun, Adv. Funct. Mater., 2025, 35, 2421136.
31. S. Xu, X. Ruan, M. Ganesan, J. Wu, S. K. Ravi and X. Cui, Adv. Funct. Mater., 2024, 34, 2313309.
32. H. Jiang, M. Sun, S. Wu, B. Huang, C.-S. Lee and W. Zhang, Adv. Funct. Mater., 2021, 31, 2104951.
33. R. K. Singh, K. Rajavelu, M. Montag and A. Schechter, Energy Technol., 2021, 9, 2100017.
34. J. Gautam, S.-Y. Lee and S.-J. Park, Adv. Energy Mater., 2025, 15, 2406047.
35. J. Li, S. Wang, J. Chang and L. Feng, Adv. Powder Mater., 2022, 1, 100030.
36. W. Liu, X. Niu, J. Tang, Q. Liu, J. Luo, X. Liu and Y. Zhou, Chem. Synth., 2023, 3, 44.
37. G. Gnana kumar, A. Farithkhan and A. Manthiram, Adv. Energy Sustainability Res., 2020, 1, 2000015.
38. C. Zhang, S. Chen, L. Guo, Z. Li, C. Yan and C. Lv, Chin. J. Chem., 2024, 42, 3441–3468.
39. Q.-X. Huang, F. Wang, Y. Liu, B.-Y. Zhang, F.-Y. Guo, Z.-Q. Jia, H. Wang, T.-X. Yang, H.-T. Wu, F.-Z. Ren and T.-F. Yi, Rare Met., 2024, 43, 3607–3633.
40. X. Tang, Y. Xin, R. Chen, X. Ren, L. Gao, H. Xu, P. Yang and A. Liu, J. Mater. Chem. A, 2025, 13, 34122–34148.
41. S. Lu, X. Zheng, L. Fang, F. Yin and H. Liu, Electrochem. Commun., 2023, 157, 107599.
42. A. S. Rasal, H. M. Chen and W.-Y. Yu, Nano Energy, 2024, 121, 109183.
43. Q. Qian, Y. Zhu, N. Ahmad, Y. Feng, H. Zhang, M. Cheng, H. Liu, C. Xiao, G. Zhang and Y. Xie, Adv. Mater., 2024, 36, 2306108.
44. W. Li, X. Lu and Z. Li, Adv. Energy Mater., 2026, 16, e04716.
45. Q. Li, Y. Wang, T. Pan, Y. Zhu and H. Pang, Sci. China Mater., 2025, 68, 317–340.
46. K. Dashtian, S. Shahsavarifar, M. Usman, Y. Joseph, M. R. Ganjali, Z. Yin and M. Rahimi-Nasrabadi, Coord. Chem. Rev., 2024, 504, 215644.
47. J. Wang, M. Sun, X. Zhang, J. Liu, J. He, W. Ge, S. Kong, G. Zhang, M. Gao, Z. Sun and X. Shi, Adv. Mater., 2026, 38, e15043.
48. T. Li, Z. Zheng, Z. Chen, M. Zhang, Z. Liu, H. Chen, X. Xiao, S. Wang, H. Qu, Q. Fu, L. Liu, M. Zhou, B. Wang and G. Zhou, Energy Environ. Sci., 2025, 18, 4996–5008.
49. J. Liang, S. Zhu, D. Chen, Y. Li, D. Zhou, N. Meng, Y. Liao, H. Sun and J. Kong, Adv. Powder Mater., 2025, 4, 100344.
50. Q. Jin, M. X. Garcia-Ortiz and L. Árnadóttir, J. Catal., 2026, 453, 116503.
51. C.-J. Huang, H.-M. Xu, T.-Y. Shuai, Q.-N. Zhan, Z.-J. Zhang and G.-R. Li, Small, 2023, 19, 2301130.
52. B. Zhu, Z. Liang and R. Zou, Small, 2020, 16, 1906133.
53. J. Wang, Y. Sun, Y. Qi and C. Wang, ACS Appl. Mater. Interfaces, 2021, 13, 57392–57402.
54. Z. Ma, X. Liu, H. Wang, L. Xie, Z. Sun, H. Zhan, X. Mi, S. Zhan and Q. Zhou, Adv. Energy Mater., 2026, 16, e02563.
55. H. Zhang, X. Meng, J. Zhang and Y. Huang, ACS Sustainable Chem. Eng., 2021, 9, 12584–12590.
56. M. Song, X. Yang, C. Guo, S. Zhang, J. Ma and H. Gao, EcoEnergy, 2025, 3, 470–481.
57. S. Ding, J. Duan and S. Chen, EcoEnergy, 2024, 2, 45–82.
58. T. D. Kühne, M. Iannuzzi, M. Del Ben, V. V. Rybkin, P. Seewald, F. Stein, T. Laino, R. Z. Khaliullin, O. Schütt, F. Schiffmann, D. Golze, J. Wilhelm, S. Chulkov, M. H. Bani-Hashemian, V. Weber, U. Borštnik, M. Taillefumier, A. S. Jakobovits, A. Lazzaro, H. Pabst, T. Müller, R. Schade, M. Guidon, S. Andermatt, N. Holmberg, G. K. Schenter, A. Hehn, A. Bussy, F. Belleflamme, G. Tabacchi, A. Glöß, M. Lass, I. Bethune, C. J. Mundy, C. Plessl, M. Watkins, J. VandeVondele, M. Krack and J. Hutter, J. Chem. Phys., 2020, 152, 194103.
59. J. VandeVondele, M. Krack, F. Mohamed, M. Parrinello, T. Chassaing and J. Hutter, Comput. Phys. Commun., 2005, 167, 103–128.
60. J. VandeVondele and J. Hutter, J. Chem. Phys., 2007, 127, 114105.
61. M. Krack, Theor. Chem. Acc., 2005, 114, 145–152.
62. J. Hutter, M. Iannuzzi, F. Schiffmann and J. VandeVondele, Wiley Interdiscip. Rev.:Comput. Mol. Sci., 2014, 4, 15–25.
63. C. Hartwigsen, S. Goedecker and J. Hutter, Phys. Rev. B, 1998, 58, 3641–3662.
64. S. Grimme, S. Ehrlich and L. Goerigk, J. Comput. Chem., 2011, 32, 1456–1465.
65. S. Grimme, J. Antony, S. Ehrlich and H. Krieg, J. Chem. Phys., 2010, 132, 154104.
66. S. Goedecker, M. Teter and J. Hutter, Phys. Rev. B, 1996, 54, 1703–1710.
67. S. Grimme, Chem.–Eur. J., 2012, 18, 9955–9964.
68. J. Zhang, J. Zhu, L. Kang, Q. Zhang, L. Liu, F. Guo, K. Li, J. Feng, L. Xia, L. Lv, W. Zong, P. R. Shearing, D. J. L. Brett, I. P. Parkin, X. Song, L. Mai and G. He, Energy Environ. Sci., 2023, 16, 6015–6025.
69. J. Li, C. Yin, S. Wang, B. Zhang and L. Feng, Chem. Sci., 2024, 15, 13659–13667.
70. Y. Luo, H. Zhou and Y. Tong, ChemSusChem, 2026, 19, e202502504.
71. J. Jian, Y. Qiao, F. Chen, C. Liu, W. Liu, Z. Li, M. Pi, Z. Li and Z. Zou, Appl. Catal., B, 2025, 374, 125390.
72. B. K. Boggs, R. L. King and G. G. Botte, Chem. Commun., 2009, 4859–4861,  10.1039/B905974A.
73. D. A. Daramola, D. Singh and G. G. Botte, J. Mater. Chem. A, 2010, 114, 11513–11521.
74. J. Li, J. Li, T. Liu, L. Chen, Y. Li, H. Wang, X. Chen, M. Gong, Z.-P. Liu and X. Yang, Angew. Chem., Int. Ed., 2021, 60, 26656–26662.
75. W. Chen, L. Xu, X. Zhu, Y.-C. Huang, W. Zhou, D. Wang, Y. Zhou, S. Du, Q. Li, C. Xie, L. Tao, C.-L. Dong, J. Liu, Y. Wang, R. Chen, H. Su, C. Chen, Y. Zou, Y. Li, Q. Liu and S. Wang, Angew. Chem., Int. Ed., 2021, 60, 7297–7307.
76. L. F. Huang, M. J. Hutchison, R. J. Santucci Jr, J. R. Scully and J. M. Rondinelli, J. Phys. Chem. C, 2017, 121, 9782–9789.
77. W. Sun, N. Govindarajan, A. Prajapati, J. Huang, H. Bemana, J. T. Feaster, S. A. Akhade, N. Kornienko and C. Hahn, ACS Appl. Mater. Interfaces, 2025, 17, 2365–2375.
78. Y. Zhou and N. López, ACS Catal., 2020, 10, 6254–6261.
79. B. Hammer and J. K. Norskov, Nature, 1995, 376, 238–240.
80. M. T. Greiner, T. E. Jones, S. Beeg, L. Zwiener, M. Scherzer, F. Girgsdies, S. Piccinin, M. Armbrüster, A. Knop-Gericke and R. Schlögl, Nat. Chem., 2018, 10, 1008–1015.
81. F. Abild-Pedersen, J. Greeley, F. Studt, J. Rossmeisl, T. R. Munter, P. G. Moses, E. Skúlason, T. Bligaard and J. K. Nørskov, Phys. Rev. Lett., 2007, 99, 016105.
82. V. Ivanistsev, R. Cepitis, J. Rossmeisl and N. Kongi, Chem. Soc. Rev., 2025, 54, 10956–10976.
83. K. Chan and J. K. Nørskov, J. Phys. Chem. Lett., 2015, 6, 2663–2668.
84. S. Ringe, C. G. Morales-Guio, L. D. Chen, M. Fields, T. F. Jaramillo, C. Hahn and K. Chan, Nat. Commun., 2020, 11, 33.
85. X. Zheng, Y. Mei, Y. Zeng, Q. Hua and S. Lu, Ind. Chem. Mater., 2026, 4, 392–404.
86. X.-F. Cheng, L.-H. Jiang, X.-P. Zhao, W. Ye, S.-X. Wei, J.-H. He and Z.-B. Zhan, ACS Appl. Mater. Interfaces, 2025, 17, 26766–26774.
87. A. J. Tkalych, K. Yu and E. A. Carter, J. Phys. Chem. C, 2015, 119, 24315–24322.
88. X. Zhu, X. Dou, J. Dai, X. An, Y. Guo, L. Zhang, S. Tao, J. Zhao, W. Chu, X. C. Zeng, C. Wu and Y. Xie, Angew. Chem., Int. Ed., 2016, 55, 12465–12469.
89. Y. Tong, P. Chen, M. Zhang, T. Zhou, L. Zhang, W. Chu, C. Wu and Y. Xie, ACS Catal., 2018, 8, 1–7.
90. Z. Ji, Y. Song, S. Zhao, Y. Li, J. Liu and W. Hu, ACS Catal., 2022, 12, 569–579.
91. Z. Zhao, T. Kang, Y. Dong, Y. Zhou, M. Yuan, H. Xu, J. Wang, G. Qiu, J. Ye and X. Chang, Appl. Catal., B, 2026, 386, 126405.
92. S.-K. Geng, Y. Zheng, S.-Q. Li, H. Su, X. Zhao, J. Hu, H.-B. Shu, M. Jaroniec, P. Chen, Q.-H. Liu and S.-Z. Qiao, Nat. Energy, 2021, 6, 904–912.
93. Z. Zhao, H. Zeng, Y. Zhou, M. Yuan and X. Chang, Appl. Catal., B, 2025, 375, 125414.
94. G. Assat and J.-M. Tarascon, Nat. Energy, 2018, 3, 373–386.
95. M. Ben Yahia, J. Vergnet, M. Saubanère and M.-L. Doublet, Nat. Mater., 2019, 18, 496–502.
96. B. Kim, J.-H. Song, D. Eum, S. Yu, K. Oh, M. H. Lee, H.-Y. Jang and K. Kang, Nat. Sustain., 2022, 5, 708–716.
97. D.-H. Seo, J. Lee, A. Urban, R. Malik, S. Kang and G. Ceder, Nat. Chem., 2016, 8, 692–697.
98. D. A. Kitchaev, J. Vinckeviciute and A. Van der Ven, J. Am. Chem. Soc., 2021, 143, 1908–1916.
99. W.-K. Han, J.-X. Wei, K. Xiao, T. Ouyang, X. Peng, S. Zhao and Z.-Q. Liu, Angew. Chem., Int. Ed., 2022, 61, e202206050.
100. L. Wang, Y. Zhu, Y. Wen, S. Li, C. Cui, F. Ni, Y. Liu, H. Lin, Y. Li, H. Peng and B. Zhang, Angew. Chem., Int. Ed., 2021, 60, 10577–10582.
101. M. Liu, W. Zou, S. Qiu, N. Su, J. Cong and L. Hou, Adv. Funct. Mater., 2024, 34, 2310155.
102. W. Jiang, Z. Zhai, J. Wu and S. Yin, J. Colloid Interface Sci., 2026, 702, 138869.
103. H. Sun, Z. Luo, M. Chen, T. Zhou, B. Wang, B. Xiao, Q. Lu, B. Zi, K. Zhao, X. Zhang, J. Zhao, T. He, J. Zhang, H. Cui, F. Liu, C. Wang, D. Wang and Q. Liu, ACS Nano, 2024, 18, 35654–35670.
104. L. Cai, H. Bai, J. Li, F. Xie, K. Jiang, Y.-R. Lu, H. Pan and Y. Tan, Energy Environ. Sci., 2025, 18, 2415–2425.
105. C. Xie, C. Zhou, Y. Zhang, B. Zhou, Y. Yao, B. Li, J. Li, J. Bai, M. Long, K. Jiang, H. Zhu and L. Zhang, Angew. Chem., Int. Ed., 2026, 65, e25119.
106. G. Zhan, L. Hu, H. Li, J. Dai, L. Zhao, Q. Zheng, X. Zou, Y. Shi, J. Wang, W. Hou, Y. Yao and L. Zhang, Nat. Commun., 2024, 15, 5918.
107. Z. Wang, Y. Wang, J. Wang, Y. Song, M. J. Robson, A. Seong, M. Yang, Z. Zhang, A. Belotti, J. Liu, G. Kim, J. Lim, Z. Shao and F. Ciucci, Nat. Catal., 2022, 5, 777–787.
108. H. Fu, Y. Jiang, M. Zhang, Z. Zhong, Z. Liang, S. Wang, Y. Du and C. Yan, Chem. Soc. Rev., 2024, 53, 2211–2247.
109. Y. Jiang, H. Fu, Z. Liang, Q. Zhang and Y. Du, Chem. Soc. Rev., 2024, 53, 714–763.
110. H. Wang, S. Yang, W. Fan, Y. Cui, G. Gong, L. Jiao, S. Chen and J. Qi, ACS Appl. Mater. Interfaces, 2025, 17, 14749–14772.
111. M. Li, X. Wu, K. Liu, Y. Zhang, X. Jiang, D. Sun, Y. Tang, K. Huang and G. Fu, J. Energy Chem., 2022, 69, 506–515.
112. Z. Liu, S. Lee, T. Zhou, J. Yang and T. Yu, J. Colloid Interface Sci., 2025, 692, 137542.
113. S. Zhang, Y. Jiang, J. Zhao and Y. Du, Chem. Soc. Rev., 2026, 55, 1494–1513.
114. N. N. Rao, C. Alex, M. Mukherjee, S. Roy, A. Tayal, A. Datta and N. S. John, ACS Catal., 2024, 14, 981–993.
115. N. N. Rao, C. Alex, S. Tomar, M. S. Naduvil Kovilakath, S.-C. Lee, S. Bhattacharjee and N. S. John, Appl. Catal., B, 2025, 371, 125177.
116. X. Wang, J. Wang, P. Wang, L. Li, X. Zhang, D. Sun, Y. Li, Y. Tang, Y. Wang and G. Fu, Adv. Mater., 2022, 34, 2206540.
117. M. Li, X. Wang, K. Liu, H. Sun, D. Sun, K. Huang, Y. Tang, W. Xing, H. Li and G. Fu, Adv. Mater., 2023, 35, 2302462.
118. L. Sun, X. Li, J. Li, Y. Zeng and S. Lu, Energy Fuels, 2025, 39, 13969–13996.
119. X. Wang, J. Hu, T. Lu, H. Wang, D. Sun, Y. Tang, H. Li and G. Fu, Angew. Chem., Int. Ed., 2025, 64, e202415306.
120. P. Jiang, B. Ge, J. Liu, C. Huang and X. F. Lu, Appl. Catal., B, 2026, 380, 125781.
121. S. Ajmal, A. Rasheed, W. Sheng, G. Dastgeer, Q. A. T. Nguyen, P. Wang, P. Chen, S. Liu, V. Q. Bui, M. Zhu, P. Li and D. Wang, Adv. Mater., 2025, 37, 2412173.
122. X. Qiang, Y. Yao, J. Yin, P. Da, Z. Mu, K. Shen, Y. Sun, Y. Zhang, P. Li, Z. Li, P. Xi and C.-H. Yan, Angew. Chem., Int. Ed., 2025, 64, e202424014.
123. W. Zheng, N. Duan, Y. Yang, P. Wang, Y. Qu, C. Zong and Q. Chen, Inorg. Chem., 2024, 63, 14602–14608.
124. Y. Zhang, W. Zhang, J. Huang, W. Cai and Y. Lai, EES Catal., 2024, 2, 1306–1313.
125. D. Liu, X. Guo, T. Yang, A. Kong, D. Lv, Y. Yang, X. Cui and R. Liu, Adv. Funct. Mater., 2025, 36, e17562.
126. Y. Wang, L. Liang, Y. Tang, W. Xiao, T. Zhou, P. Shen and P. Tsiakaras, Chem. Eng. J., 2025, 520, 165516.
127. P. Hao, Y. Xin, Q. Wang, L. Li, Z. Zhao, H. Wen, J. Xie, G. Cui and B. Tang, Chem. Commun., 2021, 57, 2029–2032.

---

Footnote

† These authors contributed equally to this work.

---