Urea Electrosynthesis Accelerated by Theoretical Simulations

Urea is not only a primary fertilizer in modern agriculture but also a crucial raw material for the chemical industry. In the past hundred years, the prevailing industrial synthesis of urea heavily relies on the Bosch–Meiser process to couple NH3 and CO2 under harsh conditions, resulting in high carbon emissions and energy consumption. The conversion of carbon‐ and nitrogen‐containing species into urea through electrochemical reactions under ambient conditions represents a sustainable strategy. Despite the increasing reports on urea electrosynthesis, a comprehensive review that delves into a profound, atomic‐level comprehension of the fundamental reaction mechanisms is currently absent. In this Perspective, recent advancements in electrochemical urea synthesis from CO2/CO and various nitrogenous species (i.e., N2, NOx−, and NO) under ambient conditions are presented, with special emphasis on theoretical understanding of the C─N coupling reaction mechanisms. Several key strategies to facilitate the C─N coupling are then pinpointed, which not only enhance their applicability in practical experiments but also highlight the significant progress achieved in this field. At the end, the major obstacles and potential opportunities in advancing urea electrosynthesis accelerated by theoretical simulations and in situ techniques are discussed. This review is hoped to act as a roadmap to ignite fresh insights and inspiration for the development of electrocatalytic urea synthesis.


Introduction
Urea (CO(NH 2 ) 2 ), with the highest nitrogen content of 46% in all solid nitrogenous fertilizers, is indispensable to the modern DOI: 10.1002/adfm.202313420agriculture, helping to alleviate the challenges posed by the world's growing population. [1]It also holds immense versatility as a key raw material for producing important chemical products such as urea−melamine−formaldehyde resin, urea formaldehyde, and barbiturates, as well as a significant additive in reducing NO x emissions in the exhaust gases from diesel and lean-burn natural gas engines. [2]Ensuring the sustainability and efficiency of the urea industry is of paramount significance for the progress and well-being of human society.The current predominant industrial method for the urea production was patented in 1922 using ammonia (NH 3 ) and carbon dioxide (CO 2 ) as the reactants, known as the Bosch−Meiser process. [3]3c,4] Though CO 2 can be readily obtained by various industrial processes, the industrial NH 3 synthesis is still dominated by the century-old Haber−Bosch process (N 2 + 3H 2 → 2NH 3 ), which involves harsh conditions (400−500 °C, 150−300 bar) due to the strong binding energy of the N≡N bonds (941 kJ mol −1 ). [5]This process consumes more than 2% of the global energy supply and generates ca.1% of greenhouse gas. [6]Therefore, this runs counter to the strategy of sustainable development, and exploring a green alternative for urea synthesis under ambient conditions is an urgent need.
Electrosynthesis emerges as a highly promising green technology, primarily owing to its emphasis on sustainable energy conversion and storage. [7]5c,10] In fact, Furuya and colleagues achieved the first instance of electrochemical urea synthesis in 1995, producing urea through the simultaneous electrochemical reduction of CO 2 and NO 3 − /NO 2 − using Cu-loaded gas-diffusion electrodes. [11]Since then, the progress regarding electrosynthesis of urea remained relatively limited until 2020, when Chen et al. reignited research enthusiasm in this field. [12]Their pioneering work showcased the potential of mild-condition urea production from N 2 and CO 2 , demonstrating that PdCu particles on TiO 2 sheets could produce urea at a rate of 3.36 mmol g h −1 with a Faradic efficiency of 8.92%. [12]Subsequently, an increasing number of studies have been dedicated to improving the catalytic efficiency for urea production through the electrochemical reduction of CO 2 and NO 3 − /NO 2 − at ambient conditions. [13]The utilization of wastes and earth-abundant molecules as carbon and nitrogen sources for urea electrosynthesis is indeed a promising avenue for achieving carbon/nitrogen neutrality, artificial nitrogen fixation, and mitigating environmental pollution.However, existing electrocatalysts are still in their infancy, and their catalytic activity and selectivity are often hindered by intrinsic challenges, including the difficulties in chemisorption of inert molecules (e.g., CO 2 and N 2 ), suppression of the side reduction reactions of CO 2 , and selective promotion of the C−N coupling reaction. [14]This raises a critical issue of understanding the internal mechanism during the emerging electrosynthesis process.
Along with the experimental research, theoretical simulation has become an essential tool in both scientific and engineering practices, emerging as a powerful technique for comprehending the reaction mechanisms and behaviors of urea electrosynthesis at an atomic level.The valuable insights derived from the simulations, such as the activation of reactants, reaction pathways, reaction barriers, bond breaking, and coupling, as well as reaction dynamics, can be utilized and analyzed to uncover the underlying principles and descriptors that guide the optimization of catalytic performance and experimental synthesis.Currently, density functional theory (DFT) calculations, based on the time-independent Schrödinger equation, represent powerful methods to deal with many-body problems. [15]They provide a computationally cost-effective and sufficiently accurate approach, enabling the calculation of the ground state energy of a given system and effectively quantifying specific subordinate properties of the material.7b,16] Therefore, a relatively straightforward strategy involves combining experimental characterizations with DFT-based atomistic understanding to gain insights into catalyst structures and reac-tion mechanisms, as well as to identify key reaction descriptors.This approach effectively changes the traditional trial-anderror method, contributing to the efficient optimization of urea electrosynthesis.
In this perspective, we offer a comprehensive summary of recent progress in electrochemical urea synthesis, and mechanisms of the C−N coupling reaction involving CO 2 /CO and various nitrogenous species under ambient conditions, mainly from theoretical perspectives.We then discuss the several key strategies aimed at modulating the active sites and optimizing reaction pathways to expedite the C─N coupling process.At the end of the perspective, the major challenges and prospects in urea electrosynthesis accelerated by theoretical simulations are discussed.We expect the insights from theoretical perspectives can provide useful guidance to steer the research of urea electrosynthesis and move it towards the practical applications.

Mechanism of Electrochemical Urea Production
The prevailing industrial approach for urea production from ammonia and carbon dioxide is limited, primarily by the low activity and selectivity of the CO 2 reduction reaction, as well as the harsh conditions required for NH 3 synthesis.Consequently, it is urgent and imperative to seek more efficient strategies in the current era.In this regard, the direct conversion of nitrogenous molecules (e.g., N 2 , NO x − , and NO), seamlessly integrated into the CO 2 reduction process, becomes a more enticing option.In the following, we summarize emerging alternatives for urea production through electrocatalytic technologies in Table 1, and elaborate on the discussion according to the different feedstocks.

N 2 and CO 2 Coupling
N 2 , being the most abundant atmospheric composition, presents itself as an exceptionally attractive source of nitrogen for urea synthesis. [17]However, due to the remarkable bonding strength of the N≡N triple bond (940.95kJ mol −1 ), direct cleavage of N 2 for the C−N coupling has been recognized as a significant challenge.Inspired by the successful electrocatalysis of N 2 reduction, [9b] the urea electrosynthesis has been achieved through the simultaneous reduction of N 2 and CO 2 at ambient conditions (N 2 + CO 2 a U L is the limiting potential for urea formation from the theoretical study. Generally, this electrochemical reaction involves three key steps: 1) co-activation of N 2 and CO 2 molecules; 2) C−N coupling reaction; 3) hydrogenation of reaction intermediates. [18]n 2020, Wang and co-workers showcased the potential of ambient-condition urea production by electrochemically coupling N 2 and CO 2 on the PdCu/TiO 2 catalyst in an aqueous environment, with a Faradic efficiency (FE) of 8.92%. [12]The DFT calculations revealed that the reaction mechanism begins with N 2 + CO 2 that involves six consecutive protonation and reduction steps, as illustrated in Figure 2a.Specifically, the initial two protonation steps are the formation of *COOH intermediate and the selective reduction to *CO, respectively, followed by the C─N coupling reaction of *CO and *N 2 to form a tower-like key intermediate *NCON.Subsequently, the formed *NCON species is further reduced by four proton-coupled electron transfer (PCET) steps, continuously hydrogenated to *NCONH, *NHCONH (alternative pathway)/*NH 2 CON (distal pathway), *NH 2 CONH, and *NH 2 CONH 2 , respectively.The potential-limiting step, which requires the highest energy input among all the PCET steps, is highly dependent on the employed catalyst.Since the adsorbed *CO 2 is preferentially reduced to a *CO prior to the C−N coupling, this reaction route is referred to as the CO 2 RR pathway in our perspective (marked in blue in Figure 2a and Figure 3a).This classic mechanism has been applied to investigate the catalytic activity of a variety of materials, including 2D TM 2 B 2 (TM = Mo, Ti, and Cr), [19] Mott−Schottky BiFeO 3 /BiVO 4 heterostructure, [14b] Ni 3 (BO 3 ) 2 , [14c] InOOH, [20] Si 2 -g-C 6 N 6 , [21] and so on, as listed in Table 1.
A recent theoretical study, conducted by Su et al., pointed out that the direct C−N coupling to the formation of *NCON species is not the only reaction pathway for urea production. [22]he C−N bond construction can also be realized by coupling *CO and *N 2 H x (x = 1, 2, and 3) intermediates, as shown in Figure 2b-d. Based on the varying sequences of the CO 2 RR and NRR during the urea formation process, these proposed alternatives are named the NRR pathway marked in orange and the NRR & CO 2 RR mixed pathway marked in green, respectively (see Figure 2b-d).To be specific, in the NRR pathway, the PCET steps convert the adsorbed *N 2 to an *N 2 H x intermediate, followed by two PCET steps to reduce *CO 2 to a *CO species (NRR & CO 2 RR mixed pathway).Figure 2b 2c).This reaction mechanism of the *NHCONH precursor has been theoretically validated on the MoP(101) surface, resulting in a limiting potential for urea production of −0.27 V (see Table 1 and Figure 3b). [18]In addition, the coupling reaction between *CO and *NH 2 NH or *NH 2 NH 2 can also facilitate the formation of C-N bonds through various reaction pathways, as shown in Figure 2d.For example, Shen et al. systematically explored the catalytic activity and electronic properties of transition metal single atom anchored on -borophene nanosheets (M@-B) during CO 2 and N 2 electroreduction. [24]Their DFT calculations demonstrated that the most feasible path for the urea production on the Nb@-B catalyst is where the *CO and *NH 2 NH 2 species are crucial for the C−N bond construction.In general, the preceding discussion outlines the potential reaction pathways for coupling N 2 and CO 2 , implying that the reaction mechanisms behind C−N coupling still lack a definitive consensus and exhibit considerable flexibility.

NO x − and CO 2 Coupling
Utilizing earth-abundant N 2 and CO 2 as reactants is undeniably an environmentally friendly approach to C−N bond formation for urea synthesis.Nevertheless, this technology is impeded by the intensive energy input required for activating the reactants, particularly N 2 , due to their chemical stability.In this regard, NO 3 − and NO 2 − with lower bonding energy of N═O bond (204 kJ/mol) have garnered the attention of researchers as potential alternatives to N 2 . [28]10e] As such, the integration of NO 3 − /NO 2 − and CO 2 into urea (2NO 3 ) would contribute to both energy conservation and environmental protection.
In 1995, Furuya et al. first verified the urea production from CO 2 and NO 3 − /NO 2 − using a Cu-loaded gas-diffusion electrode, delivering an FE of approximately 10% at −0.75 V versus standard hydrogen electrode (SHE). [11]Although the same group made further optimizations to this reaction, the formation  [12] ; b) *NHCONH on the MoP(101) surface [18] ; c) *CO 2 NO 2 on In(OH) 3 {100}facets and V O -InOOH {010} facets [13c] ; d) *NHCO on bonded Fe─Ni pairs; [25] e) *NH 2 CO on Zn nanobelts; [26] reaction energies involved in f) *CO couples with selected N−N dimer and (g) co-adsorbed *NH 2 −*NH 2 and *CO further reaction on Ni 2 Zn/C 9 N 4 . [27]chanism is not fully understood.To date, there are two distinct mechanisms proposed for this process, including 1) the Eley−Rideal mechanism, in which the surface-bound N 1 species couples with incoming gaseous CO 2 to form C−N bonds (N 1 + CO 2 ); 2) the Langmuir−Hinshelwood mechanism, where the adsorbed C 1 and N 1 species integrated into C−N bonds (N 1 + *COOH/*CO). [29]Yu et.13b,c] Once the *CO 2 NO 2 is generated, the proton/electron pairs can continue the NO 2 group through six PCET steps, leading to a generation of *CO 2 NH 2 intermediates.Afterward, the subsequent pro-tonation of *CO 2 NH 2 intermediate into *COOHNH 2 was identified as the potential determining step (PDS) during the urea synthesis processes, which aligns with the operando synchrotron radiation-Fourier transform infrared spectroscopy (operando SR-FTIR) measurements.After the formation of *CONH 2 , the second C−N coupling process was enabled and the resulting *CONO 2 NH 2 species can be further reduced into the final product.
For the Langmuir−Hinshelwood mechanism, Zhang et al. proposed that urea electrosynthesis was achieved through the coupling of *NH 2 and *COOH intermediates on oxygen vacancyrich ZnO (ZnO-V) (Figure 4b), as evidenced by in situ attenuated total internal reflectance Fourier Transform infrared spectroscopy (ATR-FTIR) measurements. [30]13e] By comparing the energy barriers for the C−N coupling of *CO with four N-containing intermediates (i.e., *NO, *N, *NH, and *NH 2 ), it was found that the coupling reaction between *NO and *CO exhibits the lowest energy barrier (0.27 eV), suggesting that *OCNO is the urea precursor.After that, the second C−N bond formation ensued between the *OCNO and *NO species, with a relatively low energy barrier of 0.24 eV, followed by the subsequent hydrogenation reactions to produce urea.Nevertheless, depending on the catalysts used, the coupling between *CO and *NH/*NH 2 can also take place.For example, Wang and co-workers reported that the coupling of *NH with *CO on bonded Fe─Ni pairs plays a pivotal role in forming the first C─N bond of *NHCO, thereby promoting urea production, as shown in Figures 3d and 4d. [25]fter the generation of *NHCO intermediate, the second NO molecule is then attached and converted into the key intermediate of *NHCONO, followed by the consecutive PCET processes to realize the urea formation finally.Moreover, this first C−N coupling mechanism base on the *NH and *CO as the key precur-sors was verified on the Cu(100) surface by ab initio Molecular Dynamics simulations at the neutral solution. [31]It is worth noting that the C−N coupling on Cu(100) surface typically occurs within a narrow potential window, generally in the low overpotential region, while the C−N coupling takes place between adsorbed *NH and solvated CO at high overpotential.13a] Therefore, these studies suggest that the reaction pathways originating from NO 3 − /NO 2 − and CO 2 exhibit a degree of complexity and variations dependent on the specific catalysts employed.

NO and CO 2 Coupling
Given that there are up to 16-electron reduction process for nitrate to urea and the intrinsic instability of nitrite, using the NO as the nitrogen source is emerging as the other process for the urea production, as NO is one of the major pollutants in the atmosphere and a crucial intermediate in the nitrate/nitrite electroreduction.To achieve this, Zhang et al. explored the catalytic activity of Zn nanobelts (Zn NBs) under ambient conditions, which reached a yield rate of 15.13 mmol g h −1 and FE of 11.26% for urea at a current density of 40 mA cm −2 . [26]Their theoretical simulations indicated that the generated C−N bond originates from the step-by-step coupling of *CO and *NH 2 intermediates, as shown in Figure 5a.In the proposed mechanism, the existence of absorbed *NO molecules can further stabilize *COOH and *CO intermediates, facilitating the subsequent C−N coupling process.When comparing the combinations of *COOH/*CO and N-related intermediates, it is found that the formation of *NH 2 CO is both thermodynamically and kinetically favored (Figure 3e).Then, the second C−N bond can be generated via the route of *CONH 2 and *NH 2 coupling, ultimately leading to the urea release.Recently, Jiao et al. proposed a unique mechanism for synthesizing urea from CO 2 and NO based on the *NO-dimerization, and demonstrated that triple-atom catalyst (TAC)Ni 2 Zn/C 9 N 4 can effectively and selectively catalyze the co-reduction of NO and CO 2 to valuable urea. [27]This TACsprompted concurrent N─C─N coupling mechanism correlates with independent CO 2 RR and NORR processes, as presented in Figure 5b.Compared to conventional ones, this coupling approach displayed two distinct features: 1) early *NO dimerization that generates the unique coupling precursor H 2 N*-*NH 2 ; and 2) concurrent N−C−N coupling derived from an insertion reaction involving H 2 N*-*NH 2 and *CO (Figure 3f-g).This is attributed to the unique configuration of the catalysts which enables the H 2 N*-*NH 2 species to have moderate N−N bond to interact with *CO.

CO as Carbon Source
A commonality we can find from the above reaction mechanisms discussion is that CO-like (*CO) and ammonia-like (*NH x ) pre-cursors generally serve as the crucial intermediate for the C−N coupling.However, the complexity of byproducts generated from CO 2 reduction is head-scratching and inevitable during the urea production, leading to the low urea yield.In this context, utilizing carbon monoxide (CO) to replace CO 2 as the source of the carbonyl moiety for urea synthesis is an alternative strategy.Moreover, CO is one of the major contributors to air pollution, stemming from the incomplete combustion process of hydrocarbons.Therefore, many researchers are dedicated to exploring the direct utilization of CO as a carbon source.For instance, Su et al. studied theoretically dispersed dual-metal anchored on N-doped graphene as electrocatalysts to synthesize urea from CO and N 2 . [32]The DFT results specified three main endothermic steps during the urea production, including the C−N coupling reaction (*N 2 + CO → *NCON), the final hydrogenation step (*NH 2 CONH + H + + e− → *NH 2 CONH 2 ), and the desorption of urea (*NH 2 CONH 2 → * + NH 2 CONH 2 ).Meanwhile, the adsorption energy of *NCONH (ΔE(*NCONH)) was identified as the principal descriptor to screen other potential electrocatalysts for urea production, with an effective range of −1.0 eV < ΔE(*NCONH) <0.5 eV.Recently, Liu et al. proposed a novel mechanism based on the synergistic effect of N−N bond cleavage and C−N coupling for highly efficient urea production from CO and N 2 . [33]Their theoretical results suggested that dual vanadium atoms anchoring onto defective graphene (V 2 N 6 ) can effectively activate the adsorbed *N 2 , where the stable N≡N triple bonds are gradually weakened until being broken after two protonation steps.As displayed in Figure 6a,b, the CO molecule can readily bond with two dissociated *NH species, resulting in an exothermic C−N coupling to form the urea precursor *NHCONH with a low kinetic energy barrier of 0.20 eV.In another study, Wei and co-workers investigated the electrocatalytic performance of dual atom catalysts (DACs), namely TM 2 /g-CN, for urea synthesis employing N 2 O and CO as the nitrogen and carbon sources. [34]s shown in Figure 6c, the N 2 O + CO reaction mechanism is  [33] c) urea production from N 2 O and CO; [34] activation barriers for C−N coupling on (111) facets d) between *CO and *N and e) *CONH and *N intermediates. [35]vided into three steps: where Cu is predicted to be the most selective catalyst to produce urea. [35]It is suggested that *CO + *N and *CONH + *N are two key reaction routes for possible C−N couplings, leading to the formation of *NCO and *NCONH intermediates (Figure 6d,e).

Key Strategies for C─N Bond Formation
After examining C−N coupling mechanisms and summarizing recent advances for urea electrosynthesis, the key catalytic strategies used in reported studies are outlined in this section.

Efficient Activation of Feedstocks
During the co-reduction of CO 2 and N-containing feedstocks, the adsorption and activation of reactants are prerequisite for urea production.Taking N 2 as an example, its nearly inert N≡N triple bond poses a challenge for cleavage and reduction under ambient conditions, and even breaking a single bond may require a significant energy input of ca.9b] Moreover, the negative electron affinity (−1.9 eV) and high ionization potential (15.85 eV) as well as the large energy gap (∼10.8223c] To activate N 2 molecules, the electron "'acceptance and back-donation"' mechanism has been widely recognized as the efficient route, [74] as shown in Figure 7a.Within this mechanism, the empty d orbitals in the metal centers could accept the lone-pair electrons of N 2 ( donation), then electrons in the partially occupied d orbitals would be donated into the antibonding orbitals of the adsorbed N 2 molecule (* back-donation), thus, the N≡N triple bonds are weakened and activated.
Like N 2 , the relatively short distance between oxygen and carbon atoms in CO 2 gives rise to a certain degree of triple bond characteristic of C═O bonds.Consequently, CO 2 is chemically inert with an initial C═O bond energy of 806 kJ mol −1 , therefore, CO 2 hydrogenation normally needs to overcome a high energy barrier. [75]The apparent electron "'acceptance and backdonation"' process mentioned above also guarantees the effective activation for CO 2 , where the occupied 1 g orbitals of CO 2 first donate their electrons to the empty d orbitals of metal sites, while the occupied metal d orbitals donate electrons back to the lowest unoccupied 2 u orbitals of the CO 2 molecule (Figure 7b). [76]urrently, extensive efforts have been devoted to optimizing the interaction between C/N feedstocks and the active sites for the efficient urea electrosynthesis.For example, Su et al. systematically explored the adsorption and activation of N 2 on the longdistance dual-metal anchored on N-doped graphene using the DFT calculations. [32]The calculated adsorption-free energies of the side-on *N 2 are less than −1.00 eV, indicating a strong chemisorption of N 2 .This can be attributed to a tricoordinate and long-distance pattern in which dispersed dual metals collaborate to enhance the interaction between the side-on *N 2 and substrates.Furthermore, the integrated-crystal orbital Hamilton population (ICOHP) analysis of N−N bonds revealed that an  [41] ; d)*CO 2 reduction to CO (up) and free energy changes and activation barriers of *CONH 2 formation (down) on pure Pd (111) and Te-doped Pd (111) surface, respectively [13a] ; e) Free energy profiles of electrochemical urea production on Mo 2 B 2 , Ti 2 B 2 , and Cr 2 B 2 . [19]crease of ICOHP (N─N) value (a smaller ICOHP value denotes a stronger binding strength) corresponds to a decrease of ΔG value of *N 2 + CO → *NCON, which suggests that the enhanced activation of the adsorbed *N 2 can facilitate the C─N bond formation.In an alternative study, Huang et al. theoretically evalu-ated the two pairs of B and transition metal (TM) embedded into graphene-like porous C 2 N monolayer (TM 2 −B 2 @C 2 N) for possible active electrocatalysts toward urea production from N 2 and CO 2 . [41]As shown in Figure 7c, N 2 adsorption on the early-TMdoped systems prefers aside-on or dissociative pattern, while N 2 adsorption tends to favor an end-on or physical manner on the late-TM-doped systems.Among them, early metal Cr 2 −B 2 @C 2 N with a moderate co-adsorption free energy of N 2 and CO 2 is selected as the promising electrocatalyst with a low limiting potential of −0.37 V versus RHE in the neutral environment.Furthermore, doping the known catalysts is a common strategy used to increase the interaction.13a] The DFT calculations demonstrated that the formation of *COOH is promoted on the Te−Pd(111) surface compared to Pd(111) surface, indicating that the Te doping can significantly activate CO 2 and boost the CO 2 RR performance (Figure 7d).Additionally, the doped surface exhibits a stronger binding capability for *CO and *NH 2 , thereby facilitating the first C−N bond construction both thermodynamically and kinetically.
Notably, it is crucial to emphasize the significance of maintaining a moderate activation for reactants, ensuring that reactions proceed optimally.According to the Sabatier principle, [77] the catalytic efficiency of a given surface will be maximum when the adsorbate-active site interaction strength reaches the optimal value.In other words, the adsorption strength of reactants on the catalyst should strike a delicate balance, which cannot be too weak (resulting in insufficient activation of reactants for the subsequent reactions) or too strong (so that over-activation impedes the smooth release of the formed urea).For instance, the adsorption energies of N 2 (CO 2 ) on 2D Mo 2 B 2 , Ti 2 B 2 , and Mo 2 B 2 are −0.93 (−1.04), −1.27 (−1.65), and −0.85 (−1.07) eV, respectively, while the adsorption energies of *NH 2 CONH 2 are −1.28,−1.55, and −1.21 eV, respectively (Figure 7e). [19]The results indicated that Ti 2 B 2 shows the most pronounced activation of N 2 and CO 2 , thereby leading to the strongest adsorption strength with *NH 2 CONH 2 and increasing the difficulty of urea desorption.

Optimization of Active Sites
The second design strategy is to optimize active sites to guarantee the efficient reduction of CO 2 and nitrogenous molecules for selective C−N coupling.For the urea formation from *N 2 and *CO, both intermediates have occupied  orbitals, giving rise to the electrostatic repulsion to their coupling. [78]To this end, the artificial frustrated Lewis pairs (FLP), composed of a Lewis acid and Lewis base that are sterically prevented from bond formation, is adopted to effectively integrate active sites for achieving chemisorption and reaction with various gas molecules. [79]pecifically, the Lewis acidic site serves as a  acceptor for *NN while the Lewis base donates to the * orbital of *NN and thereby promotes C─N coupling, as depicted in Figure 8a.14c] The calculations of the projected density of states (pDOS) and projected crystal orbital Hamilton population (pCOHP) revealed that the occupied  orbital of N 2 interacts with the d orbital of the Lewis acidic Ni site to form the d- bonding state, as shown in Figure 8b.Simultaneously, the interaction between the * orbital of N 2 and the filled p orbital of the Lewis basic O site will generate a p-* bonding state, facilitating electron donation from the filled O p orbitals into the empty * orbital of N 2 (Figure 8c).14c] During the whole urea synthesis process, Lewis basic hydroxyl groups efficiently capture the C atom of CO 2 , then Lewis acidic Ni cooperates to reduce the adsorbed *CO 2 to CO. Afterwards, the N 2 molecule is captured on the FLP sites in a highly feasible side-on configuration, followed by the direct coupling of released CO with *NN through the  orbital electron transfer to generate the *NCON urea precursor.13e] As shown in Figure 8e, the coupling of *CO with four N-containing intermediates is more selective than that of corresponding hydrogenation reactions on V O -enriched CeO 2 , and the coupling reaction via *NO and *CO is preferred with the lowest energy barrier of 0.27 eV.In comparison, a larger energy barrier of 0.52 eV for the coupling of *NO and *CO on V O -deficient CeO 2 is observed.It is noted that the hydrogenation of N-containing species (*N and *NH) on V O -deficient CeO 2 is more favorable, resulting in a higher NH 3 yield rate as the byproducts (Figure 8e).14b] This endows the BiFeO 3 (CO 2 reduction to *CO) and BiVO 4 surfaces (N 2 adsorption and C−N coupling of *NCON) with local nucleophilic and electrophilic regions, respectively.As shown in Figure 8h-i, the calculated adsorption energies of N 2 and CO 2 on BiFeO 3 /BiVO 4 were −0.17 eV and −0.06 eV, respectively, which were much lower than those observed on the individual surfaces of BiVO 4 and BiFeO 3 .As a result, the BiFeO 3 /BiVO 4 hybrids delivered a urea yield rate and FE of 4.94 mmol h −1 g −1 and 17.18% at −0.4 V versus RHE in 0.1 M KHCO 3 .Recently, Wang et al. synthesized a bonded Fe─Ni diatomic electrocatalyst for significantly improved efficiency of urea synthesis, with a high urea yield rate of 20.2 mmol h −1 g −1 and FE of 17.8%. [25]Their DFT calculations revealed that the bonded Fe─Ni pairs act as highly efficient sites for coordinated adsorption and activation of NO 3 − and CO 2 , in contrast to isolated diatomic and single-atom catalysts, where the generation of *NHCO and *NHCONO intermediates is identified as two key steps of C−N coupling.

Conclusion and Outlook
14b] urea under ambient conditions, which is regarded as a viable alternative to the energy-intensive industrial processes.In this perspective, the recent experimental and theoretical research progress on urea electrochemical synthesis is reviewed (Table 1).The construction of C−N bonds from abundant carbon (CO 2 /CO) and nitrogen sources (N 2 , NO 3 − , NO 2 − , and NO) based on diverse reaction mechanisms is summarized.Furthermore, several successful strategies in electrocatalytic C─N coupling for the sustainable syntheses of urea are highlighted.To date, although some breakthroughs have been made powered by DFT simulations in recent years, the current performance of C−N coupling is still inferior due to some significant obstacles, including poor selectivity, relatively low FE, and divergent C-N coupling mechanisms.To advance catalyst design and innova-tive strategies in the field of electrochemical urea synthesis, we propose some perspectives to further improve the electrocatalytic performance.

In Situ Characterizations
The discovery of efficient and cost-effective electrocatalysts for urea production has a strong dependence on an in-depth understanding of the C−N coupling reaction mechanism.As we have discussed, significant progress has been made in determining the reaction mechanisms and key intermediates in electrocatalytic C−N coupling by virtue of theoretical simulations.DFT calculations point to a wide range of CO 2 -derived intermediates (including *CO 2 , *CO, and *COOH) and Nderived intermediates (such as *N 2 , *N, *NH, *NH 2 , *NO 2 , and *NO) being associated with the coupling process, but they have yet to be directly and conclusively identified.10c,80] However, these operando measurements are often challenging to perform as they entail the integration of a functional electrochemical cell with the characterization instrument.Moreover, these techniques generally provide information on a few key reaction intermediates during the reaction process, such as the catalyst oxidation state or the vibrational spectra of surface-bound intermediates. [78]n the other hand, these reaction intermediates identified by operando measurements are not always the ones participating in the C─N coupling step, especially when multiple reaction processes to different products that occur concurrently.To this end, advanced in situ isotope-labelling characterization techniques (e.g., 13 CO 2 , 15 NO 3 − , and 15 N 2 ) combined with control experiments are suggested for investigating the generation and transformation processes of reactive intermediates.

Establishing the Principles for Catalyst Design
Since electrocatalytic urea synthesis processes entail the coreduction reactions of carbon and nitrogenous sources, integrating multiple PCET steps and C−N chemical coupling, designing highly efficient and robust catalysts is substantial.However, the rational design of catalysts faces obstacles due to the absence of effective theoretical principles.Therefore, it is of paramount importance to unveil the structure-activity relationships by constructing a universal descriptor.Unlike extensively studied processes including HER, oxygen evolution/reduction reaction (OER/ORR), CO 2 RR, and NRR, where various descriptors such as p/d band centers, [81] spin moments, [82] charge transfer, [83] active surface densities, [84] or atomic properties [85] are commonly employed for rational catalyst design, the research related to electrocatalytic urea synthesis remains relatively scarce.Recently, Chen et al. predicted a new class of 2D materials, namely, transition-metal phosphide monolayers (TM 2 P, TM = Ti, Fe, Zr, Mo, and W), as the potential electrocatalysts for urea production from coupling CO 2 and N 2 . [38]Based on the DFT calculations, a volcano curve between the catalytic activities (U L ) and the adsorption energies of *NCON species (ΔE *NCON ) is obtained.Either too strong (e.g., Zr 2 P and Ti 2 P) or too weak (e.g., Fe 2 P) binding strength of *NCON on TM 2 P monolayers leads to inadequate catalytic activity for urea production.The optimal ΔE *NCON , approximately −7.20 eV, corresponds to a U L of −0.39 V (Mo 2 P).Coincidentally, a similar volcano plot based on the limiting potential as a function of the adsorption free energies of *NCON species (ΔG *NCON ) is observed on various MBenes. [39]hese two studies suggest that the binding strength of *NCON on active sites can serve as a straightforward and effective descriptor to evaluate the catalytic activity for urea synthesis.This is because the *NCON intermediate plays a pivotal role in the urea formation via CO 2 + N 2 coupling.For the M@-B catalyst, the C−N coupling process takes place through the reaction of *CO + *NH 2 NH 2 → *NH 2 CONH 2 , thus, ΔE *NCON or ΔG *NCON cannot be used to describe the structure-activity relationships in this scenario. [24]It is found that the d-band center ( d ) effectively assesses the catalytic performance of the M@-B electrodes, and the promising M@-B catalysts should satisfy −0.4 < U L < 0 V but also −0.5 < ɛ d < 0.30 eV.Although these descriptors mentioned above have the potential to reveal structureactivity relationships, they are heavily relying on the DFT calculations.This reliance on DFT calculations makes the largescale screening and rational design of catalysts based on these features quite inefficient.More importantly, given the diversity of reactants and the complexity of C−N coupling mechanisms, existing descriptors do not demonstrate universal applicability.As such, the development of a simple and general descriptor constructed by the intrinsic atomic properties of catalysts that are readily available in the laboratory is currently a pressing concern for electrocatalytic urea production, as depicted in Figure 9b.

Data-Driven & Machine-Learning Methods
Selective C−N coupling for urea production is a complex process that encompasses multiple possible reaction pathways, and the precise reduction of carbon and nitrogenous precursors relies on meticulous control of reaction conditions and the utilization of appropriate catalysts.A clear understanding of the reaction mechanism in electrochemical urea synthesis holds significant importance in guiding the design of catalysts and the improvement of performance.Nevertheless, current theoretical calculations typically focus on exploring the intrinsic properties of a material based on the catalyst surface structure in the reaction process, often without considering the complex practical environment which is crucial in understanding complete chemical reactions.This is primarily due to the relatively high computational cost associated with DFT simulations, despite their wellestablished workhorse for atomistic understanding.This motivates the development and adoption of more efficient methods, particularly considering the scale of the electrocatalytic interface (comprising thousands of atoms), its dynamic nature (extending to at least nanoseconds), and the multitude of catalysts to computationally assess the electrochemical urea synthesis. [86]In this regard, machine learning (ML) has been proposed and applied because of its capability to handle complex systems and make testable predictions, as illustrated in Figure 9c. [87]Compared to DFT method solving the time-consuming quantum mechanical equations, ML learns the underlying structure-property relationships from a large amount of material data. [88]Recently, the advent of machine learning interatomic potentials (MLIP) [89] and machine learning force fields (MLFF) [90] has enabled the acceleration of DFT computations and the attainment of enhanced precision in large-scale systems.To date, ML methods have been successfully applied in various catalyst systems, such as OER, [91] HER, [92] and NRR. [93]Whereas there are still few applications of ML method in the field of urea electrocatalysis.Qiao et al. investigated a series of different C−N coupling pathways between *CO and various nitrogen-containing intermediates on extensive MXene materials based on DFT calculations and ML-based approach. [55]By using the sure independence screening and sparsifying operator (SISSO) ML algorithm, data-driven formulas for describing the relationship between E ad-CO and E ad-N with atomic physical chemistry features were identified, and 162 MXene materials were screened without time-consuming DFT calculations.This study introduced an efficient strategy for designing catalysts for intricate C−N coupling reactions through the integration of ML, which simplifies the screening process of efficient urea catalysts.With the discovery of structure-activity relationships and reaction mechanisms, the rational design of ML models is poised to become the new standard for elucidating surface behaviors and creating highly efficient catalysts for electrocatalytic urea production.
displays that the *NHCON urea precursor can be generated by the first hydrogenated *NHN and *CO intermediates through either *N 2 + *CO → *NHN + *CO or *NHN + *COOH → *NHN + *CO pathway.Alternatively, the mixed pathways for the C−N coupling reaction become feasible after the second hydrogenation on the adsorbed *N 2 , namely *NH 2 N (distal pathway) or *NHNH (alternative pathway), leading to the formation of *NH 2 CON and *NHCONH intermediates, respectively (Figure

Figure 4 .
Figure 4. Schematic depiction of representative mechanisms for urea production from NO x − and CO 2 .The first C−N coupling is enabled by a) *CO 2 NO 2 ; b)*NH 2 COOH; c) *ONCO; and (d) *NHCO.

Figure 5 .
Figure 5. Schematic depiction of representative mechanisms for urea production from a) NO and CO 2 ; and b) 2NO and CO 2 .

Figure 6 .
Figure 6.a) Schematic diagram showing the mechanisms for C−N coupling induced by the N-N bond rupture and b) kinetic energy barrier for C−N bond formation during the urea electrosynthesis process;[33] c) urea production from N 2 O and CO;[34] activation barriers for C−N coupling on (111) facets d) between *CO and *N and e) *CONH and *N intermediates.[35] (1) reducing *N 2 O to *NN by two protonation steps; (2) inserting CO directly into *NN via C─N coupling reaction to form the tower-like intermediate *NCON; and (3) hydrogenating *NCON to urea through four PCET steps.Their DFT calculations demonstrated that Cr 2 /g-CN and Co 2 /g-CN are highly active toward electrocatalytic urea formation with low limiting potentials, −0.19 and −0.15 V, respectively.Baggar et al elucidated the possible urea production on different metal surfaces through co-reduction of NO and CO based on DFT calculations,

Figure 7 .
Figure 7. Simplified schematic diagrams of a) N 2 and b) CO 2 bonding to transition metals; c) Adsorption energy of N 2 on TM 2 −B 2 @C 2 N and adsorption modes[41] ; d)*CO 2 reduction to CO (up) and free energy changes and activation barriers of *CONH 2 formation (down) on pure Pd (111) and Te-doped Pd (111) surface, respectively[13a]  ; e) Free energy profiles of electrochemical urea production on Mo 2 B 2 , Ti 2 B 2 , and Cr 2 B 2 .[19]

Figure 8 .
Figure 8. a) The frustrated Lewis pairs for the N 2 activation; the pDOS of N 2 molecule adsorbed at b) Lewis acidic Ni site and c) Lewis basic O site (in OH − group) in artificial frustrated Ni 3 (BO 3 ) 2 -150, and the pCOHP for (b) Ni−N interaction and (c) O−N interaction (right panel) of the N 2 activation; [14c] d) oxyphilic surface sites for binding the oxygen groups from NO 3 − or NO 2 − ;e) comparison of the coupling energy barrier of *NO, *N, *NH, and *NH 2 with *CO and protonation on Vo-enriched CeO 2 and Vo-deficient CeO 2 , respectively; [13e] f) heterojunctions or dual-active sites for the adsorption of carbon and nitrogen intermediates; g) planar average charge density difference (Δ) along the z-direction and visualization (bottom panel)for the BiFeO 3 /BiVO 4 heterojunction, the yellow and cyan color indicate electron accumulation and depletion, respectively, with an isosurface value of 0.013 eÅ −3 ; free energy diagrams for h) N 2 and i) CO 2 adsorption on BiFeO 3 , BiVO 4 and the BiFeO 3 /BiVO 4 heterojunction.[14b]

Figure 9 .
Figure 9. (a) Combining operando measurements will provide atomic information directly during the C─N coupling process; (b) the intrinsic atomic characteristics of the catalyst are integrated to establish a universal descriptor for predicting the reaction activity of urea electrosynthesis; (c) data-driven and machine-learning methods will expedite the discovery of promising electrocatalysts for efficient urea synthesis.

Table 1 .
Electrocatalytic Urea Synthesis by Co-reduction of small molecules at ambient conditions.