C─N Coupling Enabled by N─N Bond Breaking for Electrochemical Urea Production

Urea electrosynthesis under mild conditions has emerged as a promising alternative strategy to replace the harsh industrial HaberBosch process, which is however limited by sluggish CN coupling and low selectivity. Here, a novel mechanism based on the synergistic effect of NN bond cleavage and CN coupling for highly efficient urea production is proposed. It is found that dual vanadium atoms anchoring onto defective graphene (V2N6) can activate the adsorbed *N2, in which the stable N≡N bond can be gradually weakened until being broken after two protonation steps, with superior thermodynamic and kinetic feasibility. CO molecules can be easily adsorbed on the dissociated *NH, followed by an exothermic CN coupling to form the urea precursor *NHCONH with a low kinetic energy barrier of 0.20 eV. The dual‐atom V2N6 not only exhibits superior intrinsic activity for urea formation, with a limiting potential of −0.26 V, but also can significantly suppress the competitive N2 reduction and hydrogen evolution reactions. This study presents a new avenue for developing novel mechanisms and efficient catalysts for urea electrochemical synthesis.


Introduction
Urea (NH 2 CONH 2 ), with a high nitrogen content (46%), is regarded as one of the most vital nitrogen fertilizers in agricultural production to support about 27% of the world's population, it is also a basic raw material for manufacturing industries including pharmaceuticals, cosmetics, and plastics. [1]Currently, the industrial synthesis of urea is predominately from DOI: 10.1002/adfm.202305894ammonia (NH 3 ) and carbon dioxide (CO 2 ) under harsh conditions (150−200 °C, 150−250 bar), which requires large energy inputs. [2]1b,3] Obviously, this industrial urea synthesis is far away from the demands for sustainable development, and a mild route using clean energy is desiderated.
A burgeoning alternative strategy to reduce energy consumption for urea production is the electrocatalytic reduction of N 2 and CO 2 under ambient conditions, involving a multi-step proton-electron coupling process (N 2 + CO 2 + 6H + + 6e − → NH 2 CONH 2 + H 2 O). [4]Nevertheless, this ideal technology for urea production is greatly hampered by the intrinsic low activity and selectivity, which results from the inertness of N 2 and CO 2 , [5] the sluggishness of C─N coupling reaction, [6] and the competing reactions at reduction potential. [7]hus, experimental and theoretical efforts have been devoted to the exploration of efficient electrocatalysts.7a] In 2021, Yuan et al. designed Mott-Schottky hetero-structural Bi/BiVO 4 hybrids [6a] and perovskite BiFeO 3 /BiVO 4 hybrids [6b] with excellent electrocatalytic performances toward urea production from the adsorption of both N 2 and CO 2 molecules.Recently, Zhu et al.  have theoretically proposed 2D Mo 2 B 2 and Cr 2 B 2 materials as excellent electrocatalysts for urea formation with limiting potentials of −0.49V to −0.52 V, respectively. [8]4a] Although significant progress has been made in the exploration of electrocatalysts for urea production, two main challenges for the simultaneous reduction of N 2 and CO 2 still remain.The first one is the intrinsic selectivity of urea formation, in which the multiple side reactions and parallel N 2 /CO 2 reduction reactions strongly compete with the desired C─N coupling reaction, resulting in complex product distribution and low urea yield.To tackle this issue, an alternative strategy has been developed by utilizing carbon monoxide (CO) to replace CO 2 as the source of the carbonyl moiety for urea synthesis. [11]oreover, CO is also one of the main contributors to air pollution issue, which results from the incomplete combustion process of hydrocarbons.It is of great importance to the converting of CO into other nontoxic products. [12]Another challenge during the electrocatalytic urea synthesis is the sluggish C─N coupling for forming *NCON urea precursor, leading to the low activity of urea formation. [13]This is because *CO and *N 2 are usually adsorbed on the different sites, and inserting *CO into *N 2 needs to overcome a high bond energy of N 2 to break the N≡N bond. [8,14]ith this in mind, we propose a potential mechanism based on the N─N bond breakage facilitating the followed C─N coupling, as illustrated in Figure 1a.To achieve this mechanism, the most fundamental step is to efficiently adsorb and activate the inert N 2 molecule.Since the N 2 chemisorption with sideon configuration is the ideal initiation to activate and dissociate the N≡N bond, we expect that dual-metal sites (Figure 1b) can be the desired catalyst to activate and polarize the N≡N bond.According to the "acceptance-donation" mechanism, dualmetal sites with empty d-orbitals can accept lone-pair electrons of N 2 , and the partially filled d orbitals donate electrons back to the antibonding orbitals of the adsorbed N 2 molecule, thus, the N≡N triple bonds are weakened and activated (Figure 1c). [15]n the basis of N─N bond cleavage, the C─N coupling through the dissociated *NH can occur rapidly, as shown in Figure 1d.In this work, by virtue of density functional theory (DFT) calculations, we show that the electrochemical synthesis of urea on the homonuclear vanadium atoms anchoring onto graphene support is thermodynamically and kinetically favorable through the proposed mechanism.Remarkably, the competitive nitrogen reduction reaction (NRR) and hydrogen evolution reaction (HER) can be significantly overwhelmed, endowing good selectivity.

Two-Step Screening of High-Efficiency Dual-Atom Catalyst
To design the potential catalysts, 3d transition metals were anchored onto the defective graphene substrate to construct the dual-atom electrocatalysts for urea production, including 12 homonuclear and 8 heteronuclear DACs, as shown in Figure 2a(i,ii).
We utilized a two-step screening to search for a highly efficient catalyst for electrochemical urea production, which corresponds to two key reactions.The criteria include: 1) We first screened out catalysts with the moderate adsorption energy of N 2 (ΔE * N 2 , * represents the catalyst), which means that not only N 2 can be activated by dual active sites, but also the formed urea can be easily released from DACs; 2) since the tower-like *NCON species is the most important intermediate for urea formation as revealed by Chen et al., [7a] we evaluated the reaction free energies of *N 2 + *CO → *NCON (ΔG 1 ) based on the results of the first step.7a,8,9] Therefore, the dual-atom catalyst with the ΔG 1 < 0 will be selected out.
According to the first criterion, the adsorption energy of N 2 molecules on M 2 N 6 and M 2 N 6 C 4 is calculated and shown in Figure 2a.Except for Co 2 N 6 , N 2 adsorption on the investigated homonuclear M 2 N 6 and M 2 N 6 C 4 is spontaneous with a negative ΔE * N 2 , demonstrating that dual-atom sites have a high affinity for adsorbing N 2 molecules with side-on configuration.In particular, Sc and Ti anchored onto defective graphene substrates exhibit extremely negative values of ΔE * N 2 due to the large proportion of empty d orbitals.In contrast, the N 2 adsorption on V, Cr, Mn, Fe, Co, Ni, Cu, and Zn anchored onto M 2 N 6 is relatively weak, which is attributed to their low percentage of empty d orbitals.
To balance the N 2 adsorption on homonuclear DACs, heteronuclear DACs were further constructed using a strong-adsorption Figure 2b displays the Gibbs free energies of the *N 2 + *CO → *NCON step for V 2 N 6 , V 2 N 6 C 4 , VCrN 6 , VFeN 6 , VNiN 6 , VCoN 6 , CrFeN 6 , VCrN 6 C 4 , TiVN 6 C 4 , and TiCrN 6 C 4 DAC candidates following the first criterion.Among the 10 selected DACs, it is noted that only V 2 N 6 exhibits an exothermic ΔG 1 of −0.52 eV, while the other catalyst candidates undergo an endothermic process.This indicates that these DACs are not suitable as catalysts for urea production due to their poor ability to directly couple C and N atoms.Based on the above two-step screening, V 2 N 6 is identified as the most promising DAC for urea production due to its moderate affinity to N 2 molecules and superior ability to couple C and N atoms.Thus, in subsequent sections, we investigate the complete urea formation process on the V 2 N 6 surface and systematically explore the activity origin and reaction mechanism to evaluate the catalytic activity and selectivity of this candidate.

Catalytic Activity of V 2 N 6 DAC toward Urea Production
Following the above results obtained from two-step screening, we comprehensively explored the full reaction paths of V 2 N 6 .As discussed above, the C─N bond formation can occur through the simultaneous adsorption of *N 2 and *CO intermediates on V 2 N 6 , which is thermodynamically feasible.Therefore, the activation energy barrier for this reaction is also investigated to further evaluate the kinetic feasibility of C─N coupling, as shown in Figure 2c.The calculated kinetic energy barrier for *NCON is 1.18 eV, which is significantly larger than those reported previously (0.79 eV for the PdCu surface, [7a] 0.58 eV for Mo 2 B 2 , [8] and 0.71 eV for Cr 2 B 2 [8] ).This suggests that the direct C─N bond coupling on the V 2 N 6 via simultaneous adsorption of *N 2 and *CO is sluggish.However, the C─N coupling via *N 2 and *CO is not the only reaction pathway for urea formation.13a,16] Inspired by these studies, we propose a potential reaction pathway enabled by breaking the N─N bond, in which the C─N coupling is formed through the intermediate species of *NH and *NHCO, as illustrated in Figure S1, Supporting Information.Based on this proposed mechanism, the catalytic activity and optimal reaction process on V 2 N 6 are further evaluated, as displayed in Figure 3a.
For the adsorption of N 2 molecule, our computations show that N 2 can be adsorbed on the bridge sites of two vanadium atoms through side-on adsorption.This is because the dual-metal site can trap two lone pair electrons at both ends of N 2 and facilitate the back-donation of electrons from the d orbitals of V atoms to the * orbitals of N 2 (Figures 1c and 3b).Due to the exothermic interaction between N 2 and dual vanadium sites, the N─N bond length is elongated by 6% compared to that of the free N 2 molecule, increasing from 1.11 Å to 1.18 Å. Notably, our calculated results indicate that the CO molecule with its lonepair electrons exhibits a stronger binding affinity to the catalytic sites on V 2 N 6 and experimentally confirmed catalysts compared to N 2 , [7a,8,14,17] as listed in Tables S1 and S2, Supporting Information (related discussions can be found in the Supporting Information).These potential issues of N 2 adsorption and activation on the catalyst surface can be addressed by controlling the reactant ratio and gas flow rate during the electrolysis as suggested by the experiments. [18]Once N 2 is adsorbed, the first protonation on *N 2 is also feasible, with a free energy of −0.39 eV.This downhill process significantly enlarges the N─N bond length of *N 2 from 1.18 Å to 1.34 Å, which implies the possibility of N─N bond rupture upon further protonation.Our results indeed demonstrate that two dissociated *NH intermediates (denoted by 2*NH) on V 2 N 6 are energetically favorable, with an ultralow reaction free energy (−1.12 eV) for *NNH + H + + e − → 2*NH step, suggesting that the second protonation thermodynamically promotes the N─N bond rupture.To confirm its kinetic feasibility, we further explored the energy barrier from *NHNH to 2*NH using CI-NEB method, as shown in Figure 4a.Remarkably, the kinetic barrier for breaking the N─N bond on V 2 N 6 is 0.69 eV, indicating that the process of N─N bond rupture is kinetically feasible.To estimate the robustness of N─N bond rupture facilitated by two sequential protonation steps, we also explored the possibility of breaking N─N bond by the first protonation.As shown in Figure S2, Supporting Information, the N─N bond rupture through *NNH species is exothermic by −0.16 eV, implying that the formation of the *N + *NH species on V 2 N 6 DAC is energetically favorable.Nevertheless, the kinetic barrier for the *NNH → *N + *NH pathway is as high as 1.12 eV, which indicates that breaking the N─N bond by the first protonation is truly difficult.As such, our results demonstrate that the stable N≡N bond can be gradually weakened and eventually broken through two successive protonation steps, providing an opportunity for future C─N coupling.
Once the N─N bond is broken, the adsorption of CO becomes another point of concern in our proposed mechanism for urea production from N 2 and CO molecules.We thus compared the priority of CO adsorption and further proton-coupled electron transfer (PCET) step on 2*NH in V 2 N 6 .Figure 3a shows that the adsorption of CO on 2*NH (*NH + *NHCO) is energetically more favorable, which is −1.14 eV lower than that of protonation step to form the *NH + *NH 2 species.Then, thermodynamic evaluation demonstrates that the formed *NHCO can be pulled closer to the *NH with a downhill energy of −0.74 eV, leading to the formation of *NHCONH.7a,8-10,11c,14] Therefore, the C─N coupling of *NH + *NHCO on the V 2 N 6 DAC is thermodynamically and kinetically feasible, which plays a critical role for efficient urea production.
After the generation of the key *NHCONH intermediate through the C─N coupling, its further protonation steps on V 2 N 6 DAC were then explored.It is found that the formed *NHCONH is facile to be protonated to *NH 2 CONH due to its low ΔG value of −0.56 eV.Subsequently, the *NH 2 CONH intermediate will be reduced to the *NH 2 CONH 2 via an endothermic PCET step with a ΔG value of 0.26 eV, which is identified as the potential-limiting step for urea production and the corresponding limiting potential U urea is −0.26V.It is noted that the theoretical U urea of V 2 N 6 is comparable to that of the PdCu surface (−0.64 V), [7a] which is indicative of superior electrocatalytic reactivity toward urea formation.Furthermore, the adsorption energies of *NH 2 CONH 2 on V 2 N 6 is −1.10 eV, which is lower than those of reported catalyst (−1.68 eV for PdCu surface, [7a] −1.28 eV for 2D Mo 2 B 2 , [8] and −1.21 eV for 2D Cr 2 B 2 [8] ), indicating that the formed urea product can be easily released from V 2 N 6 DAC, especially when electrochemical reactions are carried out in flow cells.Additionally, to simulate a real aqueous solution, we correct solvent effects on urea production by employing both an implicit solvent model and an explicit solvent model, respectively (Figure S3, Supporting Information).Here, our focus was primarily on the *NNH, 2*NH, *NH + *NHCO, *NHCONH, and *NH + *NH 2 species, which are associated with the N─N bond rupture and C─N coupling.It is found that the formation of urea precursor, that is, *NHCONH, predicted by the explicit solvent model closely matches that of the implicit model.Notably, although the hydrogen bond from water layer can stabilize the intermediates, our results demonstrate only a slight change in these crucial steps compared to the vacuum simulation, consistent with the previous studies. [8,10,19]

Activity Origin of V 2 N 6 DAC toward Urea Production
Due to the high activity of the proposed N─N bond rupture mechanism, V 2 N 6 DAC has been identified as a promising electrocatalyst for urea production.However, the activity origin of dual vanadium atoms is not yet clear.It is well-known that the activity of an electrocatalyst is essentially governed by its electronic structure.Therefore, to gain deeper insights into the specialty of V 2 N 6 for urea synthesis, its electronic properties including partial density of states (PDOS), pCOHP analyses, charge density difference (CDD), and Bader charge were investigated.As can be seen in Figure S4, Supporting Information, when the spin orientation is downward, there are numerous d-orbital antibonding states of V atoms, which suggests a strong capability for trapping the lone pair electrons of free N 2 molecules.After N 2 adsorbed on V 2 N 6 , the overlapped PDOS between V 3d and *N 2 2p orbitals are observed just below the Fermi energy, as shown in the upper panel of Figure 5a.These overlapped PDOS mainly consist of V (3d xz )─*N 2 (2p z ) coupling and V (3d z 2 )─*N 2 (2p z ) coupling (Figure S5, Supporting Information), which demonstrates that the empty d orbitals of vanadium atoms can accept the  electrons of N 2 to strengthen the N 2 adsorption.Meanwhile, the components of V (3d yz ) and *N 2 (2p y ) correspond to the * back-bonding between V and *N 2 , resulting in the activated N≡N bond.This N 2 activation based on the acceptancedonation mechanism (Figure 1c) on the dual vanadium atoms can be vividly described using electronic configurations of *N 2 , as shown in Figure S6, Supporting Information.
Since breaking the N─N bond is prerequisite for efficient urea production, the PDOS of 2*NH on V 2 N 6 DAC is then analyzed.The upper panel of Figure 5b shows that the main overlap between V 3d orbitals and N 2p orbitals is located in a deeper region below the Fermi level, as compared to that of *N 2 .This indicates that the V─N bonding state is significantly enhanced after two PCET steps, giving rise to the feasibility of the N─N bond breaking.These observations can be further justified by the pCOHP analyses for bonding mechanism between the adsorbed N and V atoms, as depicted in the bottom of Figure 5a,b.The covalent bonding and antibonding states can be characterized using negative and positive overlap population, respectively.The -IpCOHP values up to the Fermi energy level were also computed to get the quantitative description of the covalent bonding strength between V and N atoms.As shown in the bottom of Figure 5a,b, the -IpCOHP values of spin-up and spin-down states for 2*NH are much lower than that of *N 2 in V 2 N 6 , leading to the extremely stronger V─N covalent bonding strength.
In addition, the CDD analyses and Bader charge calculations intuitively reveal the electron transfer between dual-atom sites and N atoms before and after two protonation steps on *N 2 intermediate.As evident from Figure 5c, the charge transfer process between the anchored V atoms and the adsorbed N 2 is bidirectional, and the charge accumulation and depletion can be found on both sides.These characteristics are in good agreement with the donation/back-donation picture in Figure 1c, namely, dual V atoms can "pull" the N 2 lone-pair electrons and "push" electrons into the N 2 antibonding orbitals, leading to the significant increase in elongation of N─N bond (from 1.11 Å in the free gas phase to 1.18 Å).Additionally, as shown in Figure 5c, it is evident that there is a net electron transfer of 0.71 e from the V atoms to N atoms.In contrast, the analyses of Bader charge and CDD for *N 2 adsorption on undesirable candidates such as Cr 2 N 6 , Mn 2 N 6 , and VFeN 6 reveal fewer net electron transfer from TM atoms to N atoms, which are 0.21 e, 0.52 e, and 0.64 e, respectively, as illustrated in Figure S7, Supporting Information.Therefore, the interaction between the N 2 molecule and V 2 N 6 catalytic sites is intrinsically stronger, facilitating both the effective adsorption and activation of the inert N 2 molecule as well as subsequent reactions.After two PCET steps, there is a significant increase in charge accumulation around the N atoms and charge depletion around the V atoms.As shown in Figure 5d, the number of electrons transferred from V atoms to N atoms in the 2*NH intermediate is more than twice that of the *N 2 intermediate.This further verifies that the strength of V─N bond can be significantly enhanced through the two protonation steps, where the covalent interaction between V and N atoms is strong enough to break the stable N─N bond.
Furthermore, we also investigated the PDOS of N1 and N2 for 2*NH intermediate in V 2 N 6 , as shown in Figure S8, Supporting Information.It is found that N2 atom has more antibonding orbitals than that of N1, which suggests that N2 will be more active in adsorbing the CO molecule.This can be ascribed to the different configurations of N1 and N2, where N1 is adsorbed on the bridge site of two V atoms, while N2 is stabilized on the top site of a V2 atom (Figures 3b and 5d).Thus, our DFT results demonstrate that breaking the N─N bond can intrinsically activate N atoms, which provides more activity for CO adsorption and further C─N coupling.
As the key intermediate for urea production, similar electronic analyses were performed on *NHCONH in V 2 N 6 DAC.atoms to stabilize the tower-like *NHCONH species, as shown in Figure 6c.

Electrocatalytic Selectivity of NRR and HER on V 2 N 6 DAC
During the process of electrocatalytic urea production, NRR and HER are two main competitive reactions, resulting in the decrease of faradaic efficiency.As reported by previous studies, [15a,20] dual-atom sites usually exhibit basal plane activity for N 2 electroreduction to NH 3 .Would the formed 2*NH species also be further reduced to NH 3 on our selected V 2 N 6 DAC?To address this issue, the thermodynamics of the electrochemical NRR on V 2 N 6 was investigated to get some deep insights.As shown in Figure 3a, although the formation of *NH + *NH 2 species on V 2 N 6 is exothermic, the *CO adsorption on 2*NH is more energetically preferred, suggesting that the further reduction on 2*NH is not prioritized.Moreover, Figure S11, Supporting Information illustrates the complete electrochemical NRR process on V 2 N 6 via an enzymatic pathway, with a limiting potential (U NRR ) of −0.53 V. Encouragingly, it is noted that the U NRR value is two times higher than that of U urea , indicating that NH 3 formation can be greatly suppressed on the V 2 N 6 catalyst.
As for the competing HER on V 2 N 6 , we first considered the interaction between H and the dual active sites.The adsorbed H (*H) is found to be stable on the bridge site of dual vanadium atoms, as shown in Figure S12a, Supporting Information, with a negative ΔE *H value of −1.19 eV (Figure 7a).Despite H atoms exhibiting stronger adsorption compared to N 2 molecules, the continuous H adsorption shows a lower priority than that of N 2 adsorption, as depicted in Figure 7a.To be specific, the second PCET step on pre-adsorbed *H to form either *H 2 (Figure S12b, Supporting Information) or *H + *H (Figure S12c, Supporting Information) is found to be endothermic with ΔE values of 0.84 eV and 0.22 eV, respectively, while the formation of *NNH on V 2 V 6 is a slightly endothermic process with a ΔE value of only 0.03 eV (Figure 7a).This indicates that either the Heyrovsky step (*H + H + + e − → H 2 ) or the Tafel step (*H + H + + e − → *H + *H) on V 2 N 6 DAC is blocked and its selectivity toward urea production is promising.
According to the aforementioned results, we deduce that the formation of NH 3 and H 2 can be greatly suppressed on V 2 N 6 DAC, demonstrating its high selectivity for urea synthesis.Considering the significant H adsorption capability, a plausible reaction mechanism for urea production on V 2 N 6 DAC in an acidic solution is proposed, as illustrated in Figure 7b.Due to the elevated concentration of H + in acidic environments, the N 2 adsorption can initiate with the pre-adsorbed *H intermediate, leading to the preferential formation of *NNH via the hydrogenation of *H + *N 2 on V 2 V 6 .Then, the subsequent electrochemical steps remain unaffected, where the sequential protonation step can effectively break the N─N bond to generate the 2*NH, followed by the facile CO adsorption and C─N coupling.Unlike the conventional *NCON pathway, the effortless C─N coupling on V 2 N 6 DAC is attributed to the dissociated *NH through the reaction *NH + *NHCO → *NHCONH, rather than the process *N 2 + *CO → *NCON.Once the key intermediate *NHCONH is formed, the formation of urea becomes straightforward.

Stability of V 2 N 6 DAC
In order to establish a connection with the experimental synthesis of V 2 N 6 DAC predicted here, we studied its thermodynamic stability in terms of diffusion and aggregation of vanadium atoms. [21]First, the adsorption energy (E ad ) of V atom at the defective graphene support is evaluated to assess the diffusion of dual V atoms, which can be calculated by the following equation.
where E V 2 N 6 , E V , and E N 6 @G indicate the energies of V 2 N 6 catalyst, single vanadium atom, and the defective graphene with two double vacancies surrounded by six N atoms.It is found that the E ad of V 2 N 6 DAC is −12.18 eV, indicating that the adsorption of two V atoms at defect sites is thermodynamically favorable than being leached.Moreover, the bulk cohesive energy (E coh ) of V 2 N 6 is also checked by the equation below.
where E bulk is the energy of the V metal bulk.Our results show that E coh of V 2 N 6 is −8.95 eV, leading to E ad − E coh < 0, which suggests that dual vanadium atoms embedded into the defective graphene are energetically preferred than the metal aggregation.
To further evaluate the thermal stability of the promising V 2 N 6 catalyst, AIMD calculations at 300 K for 9000 fs were performed, as shown in Figure S13, Supporting Information.The AIMD results showed that V 2 N 6 is kinetically stable at 300 K and no obvious structural disruption is observed during a total simulation time of 9000 fs, suggesting that atomic V atoms will not diffuse on the surface to form a cluster.Therefore, it is demonstrated that V 2 N 6 DAC has a good thermal stability at 300 K. Except for the thermodynamic stability, eligible electrocatalysts should exhibit outstanding surface stability under electrochemical situations.To determine this, the dissolution potential U diss [11c,22] for the V metal dopants was computed to assess the dissolution-resistance of metal atoms from the substrate into water under an applied potential.Here, the U diss is defined as where U o diss is the standard dissolution potential of bulk metal, n is the number of electrons involved in the dissolution, and E f = 1/2 E ad represents the formation energy of V 2 N 6 DAC, respectively.In general, a catalyst with U diss > 0 V is regarded to be electrochemically stable.Our calculated result shows a positive U diss value of 1.87 V, evidencing the good stability of V 2 N 6 under electrochemical environments.

Conclusions
In conclusion, we performed DFT calculations to systematically explore the feasibility of dual vanadium atoms anchoring on the defective graphene as an electrocatalyst for urea production.The activity, selectivity, and stability of the V 2 N 6 DAC were comprehensively studied.Our calculations demonstrate that the inert N 2 molecule can be adsorbed on dual vanadium sites, and the adsorbed N 2 can be easily reduced to two dissociated *NH.This feasible N─N bond rupture increases the affinity of *NH intermediates for binding to the CO molecules and rapidly couples to form *HNCONH precursor with an ultralow energy barrier of 0.20 eV.The limiting potential of urea formation based on the synergistic effect of N─N bond cleavage and C─N coupling for V 2 N 6 DAC is considerably small (−0.26V), which is comparable to that of the PdCu alloy catalyst.Meanwhile, the NRR and HER are significantly suppressed on V 2 N 6 DAC, leading to the excellent selectivity of urea formation.Our work not only explores a new class of electrocatalysts with superior catalytic performance for urea synthesis under ambient conditions, but also proposes a novel mechanism of the intrinsic synergistic effect for balancing reducing N 2 and fixing CO to produce urea, which could promote more experimental and theoretical efforts on developing dual active sites for other multi-electron electrochemical reactions.

Computational Methods
All calculations of structure relaxation and electronic properties were performed using spin-polarized density-functional theory (DFT) methods in the Vienna Ab initio Simulation Package (VASP) with projector-augmented wave pseudopotential (PAW). [29]The Perdew-Burke-Ernzerhof (PBE) functional at the generalized gradient approximation (GGA) level was used to treat the exchange-correlation. [30] In order to incorporate the effects of nonlocal van der Waals interactions that are not included correctly in conventional DFT calculations, the DFT-D3 method was adopted for dispersion corrections here. [31]A cutoff energy was set to 520 eV, and k-points were sampled using the gammacentered mesh with a reciprocal space resolution of 2 × 0.04 Å −1 and 2 × 0.02 Å −1 for structural optimization and static selfconsistent calculations, respectively.
We modeled two defective graphenes with (6 × 6) supercell as a support for a dual-metal site: two double-vacancy surrounded by six N atoms (M 2 N 6 in Figure 2a(i)) and two single-vacancy surrounded by six N and four C atoms (M 2 N 6 C 4 Figure 2a(ii)).A 20 Å vacuum along z-direction was applied to eliminate artifactual interactions between the periodically repeated images.Gas molecules were calculated in 20 × 20 × 20 Å cell with gamma point sampling.All atoms were allowed to relax until the Hellmann-Feynman forces were smaller than 0.01 eV Å −1 , and the convergence criterion for the electronic self-consistent loop was set to 10 −5 eV.It is generally recognized that the large Coulombic repulsion between the localized 3d electrons of transition metals can be described by the DFT + U method. [32]Here the corresponding U -J values for different transition metals were listed in Table S3, Supporting Information. [33]The projected crystal orbital Hamilton population (pCOHP) method was used via the LOBSTER program to understand the chemical bonding between transition metals and adsorbed nitrogen atoms. [34]Solvation of the surface and adsorbates is taken into account by using the Poisson-Boltzmann implicit solvation model [35] and the explicit solvent model developed by Reda et al. [36] In a unit cell, 17 H 2 O molecules which hydrogen atoms can point either toward or away from the surface constructed the hexagonal honeycomb water layer, including a vacancy of H 2 O molecule to provide enough space for N 2 and CO co-reduction, as shown in Figure S14, Supporting Information.
To describe the electrochemical synthesis of urea, the computational hydrogen electrode (CHE) model was used to establish a free energy profile, as proposed by Nørskov and co-workers. [37]he Gibbs free energy change (ΔG) for each elementary step can be computed by: ΔG = ΔE + ΔE ZPE − TΔS + ΔG U + ΔG pH (4) where ΔE is the electronic energy difference between the freestanding and adsorption states of reaction intermediates, which can be directly obtained from DFT calculations.Generally, it is inevitable that the most stable structures of successive intermediates undergo deformation during the whole chemical reaction process.In this work, the static surface was used for each elementary step to calculate the reaction energy excluding the deformation energies of intermediates.Additionally, to account for the underestimation of DFT calculation with PBE functional, the non-adsorbed gas-phase molecule of CO has to include the correction of 0.13 eV. [38]ΔE ZPE is the change in zero-pint vibrational energy and −TΔS is the entropy contribution at room temperature.For each reaction species, its E ZPE and TS can be calculated by the following equations, [39] respectively.
where k B represents the Boltzmann's constant, h is the Planck constant,  i denotes the frequency of the normal-mode of the adsorbed species, R is the ideal gas constant, and T is the system temperature, which is set to 298.15 K here.Thermal correction terms were only needed for the adsorbates and free molecules, which were listed in Table S4, Supporting Information.ΔG U is the free energy contribution related to applied potential U, and U is the operating electrochemical potential relative to the versible hydrogen electrode (RHE).ΔG pH is the correction of H concentration, that is, ΔG pH = k B T × ln10 × pH.In this work, the value of pH was assumed to be zero in a highly acidic solution.Furthermore, the transition states and kinetic barriers for the N─N bond rupture and C─N coupling reaction were identified by the climbing image nudged elastic band (CI-NEB) method. [40]

Figure 1 .
Figure 1.a) Schematic diagram showing the proposed mechanism for C─N coupling induced by the N─N bond rupture during the urea electrosynthesis process; b) the desired structure of M 1 M 2 dual metal sites; simplified schematic diagram of c) N 2 bonding to dual-metal sites and (d) C─N coupling for *NHCONH formation based on the N─N bond rupture.

Figure 2 .
Figure 2. a) Calculated N 2 adsorption on M 2 N 6 and M 2 N 6 C 4 DACs; b) calculated Gibbs free energy of *NCON formation on M 2 N 6 and M 2 N 6 C 4 DACs; c) kinetic energy barrier for C─N coupling through the adsorptions of *N 2 and *CO on V 2 N 6 .Insets are the optimized structures in the initial (IS), transition (TS), and final states (FS), respectively.

Figure 3 .
Figure 3. a) Free energy profiles of electrochemical urea production on V 2 N 6 DAC and b) Side view of all optimized possible reaction intermediates for urea production on V 2 N 6 DAC.

Figure 4 .
Figure 4. Kinetic energy barrier for a) N─N bond rupture and b) C─N bond formation, respectively.Insets are optimized structures in the initial (IS), transition (TS), and final states (FS) along the N─N bond rupture and C─N bond formation pathway, respectively.

Figure 5 .
Figure 5. PDOS for V 3d orbitals and adsorbed N 2p orbitals and pCOHP analyses of the chemical bonding between the V and adsorbed N atoms of a) *N 2 and b) 2*NH on V 2 N 6 DAC, respectively.The upward and downward arrows indicate spin-up and spin-down states, respectively; the charge density difference of c) *N 2 and d) 2*NH on V 2 N 6 DAC, respectively.The value of the isosurface is set as 0.01 e Å −3 .

Figure 6a and
Figure 6.a) PDOS for V 3d orbitals, N and C 2p orbitals of *NHCONH on V 2 N 6 DAC; b) schematic illustration of the dual-donor mechanism of C─N coupling on V 2 N 6 DAC; and c) charge density difference of *NHCONH onV 2 N 6 DAC, respectively.The value of the isosurface is set as 0.01 e Å −3 .

Figure 7 .
Figure 7. a) Competing hydrogen evolution (pink and green) and urea production (orange) on V 2 N 6 DAC and b) schematic diagram of the possible reaction mechanism for urea production on V 2 N 6 DAC in acidic solution.