The precursor reduction method is quite simple and straightforward, and it is probably the most widely employed method for synthesizing metal nanoclusters including AuAg clusters. In this method, the metal-ligand precursor or complex is first formed through the reaction between the metal salts and the ligand. Followed by the addition of a reducing agent, AuAg nanoclusters were produced. To improve the monodispersity of the obtained nanoclusters, post-reaction treatments or purification steps are necessary in most cases.

One typical example using this method is the synthesis of [Au₃₄Ag₂₇(C≡CR)₃₄](PPh₄) (PPh₄: tetraphenylphosphonium) cluster with a concentric four-shell structure^[61]. The Au₃₄Ag₂₇ nanocluster was synthesized by reducing a mixed solution of AuC≡CR and AgC≡CR with NaBH₄ in an ice bath. Note that the order of reactant addition can be critical for obtaining specific AuAg nanoclusters. For instance, during the synthesis of Au₈Ag₃(PPh₃)₇Cl₃ molecule, HAuCl₄ was first treated with PPh₃ (PPh₃: triphenylphosphine) ligand to form some Au(I) complex, and then AgSbF₆ was added to form AuAg alloy complex^[62]. After filtration to remove the AgCl precipitate, NaBH₄ was added to reduce the filtrate. Once the mixture turned red, it indicated that the Au₈Ag₃ cluster had formed, which is isostructural to the Au₁₁(PPh₃)₇Cl₃ nanocluster. Notably, if HAuCl₄ and AgSbF₆ were co-mixed with PPh₃, they react immediately to form the AgCl precipitates; consequently, no AuAg cluster can be obtained. Similarly, the [Au₁₃Ag₁₂(PPh₃)₁₀Cl₈]SbF₆ nanocluster was fabricated using a nearly identical protocol^[54]. Au(I)PPh₃Cl was first prepared and then reacted with AgSbF₆ in an ice bath for 30 min. The mixture was subsequently stirred in the dark for 24 h to yield the Au₁₃Ag₁₂ nanocluster, which comprises two Au₇Ag₆ icosahedra sharing a central gold atom. Interestingly, this procedure can produce two conformational Au₁₃Ag₁₂ cluster isomers with complete reversibility by controlling the reaction temperature^[54].

It is worth noting that the precursor reduction method often fails to achieve precise control over the composition and structure of target alloy clusters. This is primarily due to insufficient regulation of metal nucleation and growth kinetics, which in turn leads to poor selectivity and generally low yield^[63].

The metal exchange reaction usually employs pre-synthesized nanoclusters as the substrate to react with more reactive (or less noble) metal salts. Typically, the as-formed Ag nanocluster can react with Au(I) complex to yield AuAg nanoclusters^[24].

Bootharaju et al. fabricated the Ag₂₄Au(SPhMe₂)₁₈ (SPhMe₂: 2,4-dimethylbenzenethiolate) nanoclusters using this method, where the Ag₂₅(SPhMe₂)₁₈ nanocluster reacts with AuClPPh₃ thanks to the electrochemical redox potential difference between Au(I) and Ag atoms in the Ag₂₅ framework^[64]. Interestingly, Au substitution in the Ag₂₅ core induces some structural distortion in its icosahedral framework^[64]. In another study conducted by Bootharaju et al., adding AuPPh₃Cl to the [PdAg₂₄(SR)₁₈]^2- (SR: thiolate ligand) nanocluster system controllably generated the mono-gold-substituted [AuAg₂₄(SR)₁₈]^- nanocluster, where Au replaces Pd at the central position^[65]. The Au-to-Pd substitution can be visually observed by an orange-to-dark-green color transition^[65].

Remarkably, the metal exchange of the Au(I)-to-Ag(I) process can be precisely controlled and directly visualized by mass spectrometry. In 2017, Yao et al. documented a surface motif exchange (SME) reaction to selectively replace the surface motifs of [Ag₄₄(SR)₃₀]^4- nanoclusters, leading to the formation of core-shell structured [Ag₃₂@Au₁₂(SR)₃₀]^4- nanoclusters^[66]. Such a process was analyzed in detail using mass spectrometry and tandem mass spectrometry. Theoretical calculations also disclosed that the Au@Ag core shell stabilizes the Ag₂₀ shell, which can prevent the Au atoms from diffusing inwardly. Such a dedicated SME reaction opens a new door to the precise synthesis of AuAg nanoclusters with pre-designed alloying sites^[66].

Although the metal exchange reaction can achieve atom-level cluster-to-cluster structural transformation, most current examples are limited to single-atom exchanges or involve only a very small number of metal atoms^[64,67]. Customized nanocluster synthesis for regulating the number of doped metal atoms via the metal exchange reaction remains quite challenging. This is predominantly attributed to the inherently uncontrollable galvanic reduction process during the exchange reaction.

Anti-galvanic reduction (AGR) refers to the reaction in which metal ions or complexes are reduced by less reactive (more noble) metal nanoclusters. AGR may seem to contradict the Galvanic theory, but such a reaction occurs on ultrasmall metal nanoclusters due to their quantum confinement effect.

In 2012, Wu first discovered this novel redox phenomenon of [Au₂₅(SC₂H₄Ph)₁₈] nanoclusters, which can react with AgNO₃ at a 5.2:1 weight ratio, yielding [Au₂₂Ag₃S₁₂]^- and [Au₂₃Ag₂S₁₂]^- as the dominant products^[68]. Following that, Yao et al. systematically studied this reaction, where two Ag atoms can be added onto [Au₂₅(SC₂H₄Ph)₁₈] to form Au₂₅Ag₂(SC₂H₄Ph)₁₈ under well-controlled reaction conditions and precise reactant-substrate stoichiometry^[69]. The successful preparation of monodisperse, high-purity Au₂₅Ag₂(SC₂H₄Ph)₁₈ validates the effectiveness of AGR for targeted metal doping.

Beyond simple metal replacement or surface metal addition, the introduction of foreign metal atoms may induce size and structural changes in parent metal nanoclusters, leading to the formation of new nanoclusters. One classic case is the synthesis of the [Au₁₃Ag₁₂(PPh₃)₈(BMIm)₂I₈]SbF₆ (BMIm: 1,3-bimethylimidazole) nanocluster with a rod-like carbene structure reported recently by Zhang et al.^[70]. By reacting the pre-synthesized Au₉ nanoclusters with freshly prepared Ag-BMIm under vigorous stirring in the dark, the [Au₁₃Ag₁₂(PPh₃)₈(BMIm)₂I₈]SbF₆ nanocluster can be obtained within a few minutes.

It is worth noting that the unique physicochemical properties arising from the size and surface effects of metal nanoclusters enable AGR to proceed spontaneously^[23,24], but this reactivity does not apply to all types of noble metal nanoclusters. It is restricted to specific reaction systems and depends on the reaction conditions^[71,72]. Therefore, expanding the universality and applicability of AGR is promising for producing novel atomically precise AuAg and other alloy nanoclusters.

Ligand exchange is another effective strategy for fabricating atomically precise AuAg nanoclusters^[73]. It involves partial or complete replacement of the present ligands with incoming ligands on pre-synthesized AuAg nanoclusters. In addition, the extra ligand can initiate etching in the reaction system, leading to size transformation of the parent AuAg nanoclusters.

Au₁₆Ag₁(S-Adm)₁₃ (Adm: Adamantane) was synthesized using this method by Kang et al.^[74]. Firstly, a trace amount of Ag(I)(S-C₆H₁₁) was introduced into the Au₁₈(S-C₆H₁₁)₁₄ solution, yielding a mixture of Au₁₇Ag₁(S-C₆H₁₁)₁₄ and residual Au₁₈(S-C₆H₁₁)₁₄. Subsequently, the HS-Adm ligand was added to the mixture. After the reaction was completed, Au₁₆Ag₁(S-Adm)₁₃ was isolated by preparative thin-layer chromatography (PTLC), thanks to the distinct size and surface property differences between the newly generated nanoclusters and the parent nanoclusters^[74]. In another study, Hu et al. fabricated the [Au₈Ag₅₅(Dppp)₄(C₆H₁₁S)₃₄](BPh₄)₂ [dppp: 1,3-Bis(diphenylphosphino)propane, BPh4: tetraphenylborate] nanocluster using this approach; however, a size-expansion process was observed^[75]. Specifically, Au₇Ag₈ nanoclusters were first prepared and dispersed in CH₂Cl₂, after which Dppp and C₆H₁₂SH were introduced into the reaction system. The mixture was stirred for ~12 h, and monodisperse Au₈Ag₅₅ nanoclusters were obtained after completion of the reaction^[75].

It should be noted that the ligand exchange reaction can facilitate atom-precise surface engineering of pre-formed alloy nanoclusters, allowing precise customization toward catalytic^[74,75], optical^[76], or biomedical applications^[77] through functionalized ligands. Nevertheless, during the reaction process, the cluster structure is prone to unintended alterations^[78], imposing fundamental constraints on those applications that require specific atomic structure or configurations.

Overall, the above four most widely employed synthetic methods possess different features. The comparison of their advantages, limitations, structural precision, and scalability is summarized in Table 1.

There are other synthetic approaches that have shown great promise for the synthesis of atomically precise AuAg nanoclusters. For example, photochemical synthesis provides an accessible platform to tailor nanoclusters with specific sizes. In 2023, Wang et al. reported the synthesis of atomically precise [Ag₂₂Au₃(4-^tBuPhC≡C)₂₀(Dpppe)₃](SbF₆)₃ [dpppe: 1, 5-bis(diphenylphosphino)pentane] nanoclusters via this method^[79]. In a typical trial, AgSbF₆ was dissolved in ethanol, followed by the addition of a CH₂Cl₂ solution containing 4-tert-butylphenylacetylene, 1, 5-bis(diphenylphosphino)pentane, and HAuCl₄. Et₃N was then introduced under vigorous stirring, and the mixture was irradiated with a 5 W white light-emitting diode (LED) at room temperature for 24 h to yield the target nanocluster of Ag₂₂Au₃^[79].

Inter-cluster reactions can form bimetallic nanoclusters through chemical reactions between different cluster molecules, an extremely complex and mysterious process involving atomic exchange and bond reorganization^[80]. For instance, Krishnadas et al. documented the fabrication of [Au₁₂Ag₃₂(FTP)₃₀]^4- superatomic nanoclusters via this method, where the dichloromethane solutions of [Au₂₅(FTP)₁₈]^- and [Ag₄₄(FTP)₃₀]^4- with precise stoichiometric ratios were mixed and reacted at 30 °C^[81]. Remarkably, the resulting product of [Au₁₂Ag₃₂(FTP)₃₀]^4- exhibits a geometric and electronic closed-shell configuration, conferring exceptional stability^[81]. Nevertheless, such successful cases of inter-cluster reactions to acquire atomically precise AuAg nanoclusters remain rare, and understanding the complex structural transformation process remains challenging and is still an active area of research in the nanocluster research community.

Hydrogen energy, as an ideal energy carrier characterized by its high energy density and combustion product being solely water, has been widely regarded as one of the key renewable energy sources to achieve the carbon neutrality goal^[85,89]. Recently, hydrogen production through electrochemical water splitting has been attracting increasing research attention, not only because it can provide high-purity H₂ gas for direct use, but also because it can fully utilize renewable electricity generated from solar, wind, tidal, and other intermittent sources^[90-93]. Electrochemical water splitting contains two half reactions: the hydrogen evolution reaction (HER) at the cathode and the oxygen evolution reaction (OER) at the anode. However, for both HER and OER, additional energy input is required to overcome the energy barrier^[94,95]. To facilitate the reaction process, it is essential to lower the energy barriers; therefore, developing efficient and durable catalysts for both the HER and OER is imperative^[96-99].

In the past few years, bimetallic nanoclusters have emerged as model and effective catalysts for both HER and OER, as they have good catalytic performance, and more importantly, their well-defined structure can offer a great platform to unravel the structure-activity relationship at the atomic level^[30,32,100]. Specifically, for atomically precise AuAg nanoclusters, their uniform size and composition, unique electronic structure, precise Au/Ag stoichiometric ratio, and crystallographically resolved structures endow them with significant advantages as model bimetallic catalysts for gaining deep mechanistic understanding and guiding improved catalyst design^[101]. In 2021, Li et al. reported an Au₃₆Ag₂(SR)₁₈ nanocluster with low thiolate ligand surface coverage, which exhibited ultrahigh HER activity with a very low overpotential to achieve a given current density^[102]. Specifically, Au₃₆Ag₂(SR)₁₈ consists of three icosahedral (I_h) units, reminiscent of the monomeric I_h Au₂₅(SR)₁₈^- and dimeric I_h Au₃₈(SR)₂₄, and these tri-units endowed it with a unique geometric and electronic structure. The outstanding HER performance is attributed to the low ligand-to-metal ratio, low-coordinated Au atoms, and the unfilled superatomic orbitals^[102].

Recently, our group investigated the electrocatalytic performance of atomically precise Au₁₅Ag₂₃(^tC≡CBu)₁₈Br₆ nanoclusters in HER, OER, and OWS^[103]. Au₁₅Ag₂₃ nanocluster adopts a triple-layered core-shell-shell structure (Au₆@Au₆Ag₂₃@Au₃) with a tetragonal bipyramidal Au₆ core, a middle Au-Ag alloy layer, and an outer Au shell co-protected by 18 bulky alkynyl ligands and 6 Br ligands [Figure 1A and B]. As a 14-electron superatom, its unique electronic configuration and synergistic ligand coordination enable distinct catalytic properties. For HER in 0.5 M H₂SO₄, the Au₁₅Ag₂₃ nanoclusters exhibited a low overpotential of 125 mV at 10 mA·cm^-2 and negligible current decay over 16 h, outperforming the AuAg NPs (overpotential of 235 mV) [Figure 1C]. In the OER test, the Au₁₅Ag₂₃/NiFe-layered double hydroxide (LDH) composite demonstrated an ultralow overpotential of 250 mV (10 mA·cm^-2) in 1.0 M KOH, with only ~5% current loss over 30 h [Figure 1D]. X-ray photoelectron spectroscopy (XPS) analysis revealed electron transfer from NiFe-LDH to Au₁₅Ag₂₃, generating high-valent Fe species as the active sites. For OWS, a two-electrode system employing Au₁₅Ag₂₃ as the cathode catalyst and Au₁₅Ag₂₃/NiFe-LDH as the anode catalyst required an applied voltage of 1.51 V at 10 mA·cm^-2 in 1.0 M KOH, surpassing the control system of AuAg NPs paired with NiFe-LDH (1.65 V @ 10 mA·cm^-2, Figure 1E). Moreover, the AuAg NC||Au₁₅Ag₂₃/NiFe-LDH combination exhibited outperformed long-term stability in the 50 h continuous test, manifested by the only ~4% decay of the initial current density, while the AuAg NP||NiFe-LDH combination lost ~30% of the initial current after 50 h [Figure 1F]. The superior catalytic performance was attributed to synergistic co-protection by alkynyl and bromine ligands, which significantly enhanced the electron transfer and active site accessibility. This study highlights the AuAg nanoclusters as multifunctional catalysts to promote HER, OER, and OWS toward hydrogen production.

Electrochemical CO₂RR represents a promising approach that can not only mitigate the growing global concern over the harmful effects of excessive CO₂ emissions but also generate value-added carbon-based fuels and chemical feedstocks^[104-106]. However, the development and implementation of CO₂RR is hindered by several serious challenges such as the thermodynamic stability of the CO₂ molecule and the competitive HER reaction. Moreover, CO₂RR involves a very complex reaction pathway with multiple proton/electron transfer steps. Therefore, the CO₂RR catalyst plays a central role in activating the CO₂ molecule, determining the reaction pathway, and governing the catalytic activity and final product selectivity^[107-109].

Atomically precise AuAg nanoclusters hold great merit as model and effective catalysts for CO₂RR, not only due to their ultrasmall size and high activity and the exclusive selectivity for CO formation, but also because their atom-precise structure provides atomic-level insight into the reaction mechanism^[110,111]. Specifically, with the aid of in-situ spectroscopic study and theoretical calculations, such an atomically precise structure is beneficial for identifying the real catalytic activity, elucidating the synergistic effect between Au and Ag, disclosing the surface ligand role, and even understanding the catalyst structure evolution/reconstruction mechanism^[75]. For instance, Xu et al. reported the evolution from Au₂₄Ag₂₀ monomer clusters into Au₄₃Ag₃₈ dimeric nanoclusters, where the Au₂₄Ag₂₀ kernel of Au₁₂@Ag₂₀ is assembled into Au₁₂@Ag₁₉-Au-Au₁₂@Ag₁₉, forming the kernel of Au₄₃Ag₃₈^[112]. Interestingly, the monomeric Au₂₄Ag₂₀ catalysts outperformed dimeric Au₄₃Ag₃₈ catalysts in CO₂RR^[112].

The kernel and shell structure in the metal core can significantly affect the CO₂RR catalytic properties of AuAg nanoclusters, including the catalytic activity and stability. In 2022, Seong et al. reported a core-shell AuAg nanocluster of AuAg₁₂@Au₁₂(SEtPh)₁₈ (SEtPh: 2-phenylethanethiolate)^[110]. By engineering the active sites through replacing the Ag₁₂(SR)₁₈ shell of Ag₂₅ with an Au₁₂(SR)₁₈ shell, the resulting AuAg₁₂@Au₁₂ NC exhibited dramatically enhanced performance compared to Au₂₅. AuAg₁₂@Au₁₂(SEtPh)₁₈ consists of an AuAg₁₂ core and an Au₁₂ shell protected by 18 phenylethanethiolate (SEtPh) ligands through dimeric staple units. The overall framework of AuAg₁₂@Au₁₂(SEtPh)₁₈ resembles the icosahedral structure of Au₂₅(SEtPh)₁₈, where the thiolate ligands bond to the metal core-shell via S atoms, and the core-shell interface is stabilized by metal-metal bonds [Figure 2A]. Figure 2B and C demonstrates that the AuAg₁₂@Au₁₂/GDE catalyst exhibits a CO onset potential of 1.62 V and achieves a CO current density (j_CO) of 339 mA·cm^-2 at a full-cell potential of 2.66 V in a zero-gap CO₂ electrolyzer, outperforming the Ag NP catalyst. During 24-hour galvanostatic electrolysis at a current density of 200 mA·cm^-2, the full-cell potential remains stable at 2.13 ± 0.03 V with an initial CO selectivity of 90% and an energy efficiency of 57%, highlighting its excellent stability and practical applicability^[110]. To better comprehend the reaction mechanism, the authors also performed operando attenuated total reflection surface-enhanced infrared absorption spectroscopy (ATR-SEIRAS) to confirm the active site. During CO₂RR, the ^*CO formation is the key step, and the C=O stretching band of ^*CO for Au₂₅ NC is centered at 2,096 cm^-1[110]. Interestingly, the ^*CO signal observed for AuAg₁₂@Au₁₂(SEtPh)₁₈ is located at 2,072 cm^-1, which is consistent with that of Au₂₅ NC. This indicates that the Au site in AuAg₁₂@Au₁₂(SEtPh)₁₈ serves as the active site for CO₂RR, and the transplanted Au atoms function as the active site for CO₂-to-CO electroreduction. These findings demonstrate that atomic-level engineering of metal shells effectively enhances CO₂-to-CO electroreduction, offering a novel strategy for designing high-performance electrocatalysts.

Furthermore, by doping Au or Pt atoms into Ag nanoclusters, the as-formed AuAg or PtAg nanoclusters exhibited drastically different CO₂RR properties from the parent Ag nanoclusters. In a recent study, Yoo et al. synthesized structurally precise monomeric PtAg₂₄(IPBT)₁₈ (IPBT: 2-Isopropylbenzenethiol) and AuAg₂₄(IPBT)₁₈ nanoclusters, which exhibited different CO₂RR activities^[113]. AuAg₂₄(IPBT)₁₈ exhibits a highly symmetric nanoscale framework, with isopropylphenyl groups of the thiolate ligands exposed on the cluster surface, ensuring good chemical stability and defining its electronic structure [Figure 3A-C]. In a zero-gap membrane electrode assembly (MEA) electrolyzer, the cluster achieves a CO partial current density of -202.2 mA·cm^-2 at -3.6 V with 90.4% Faradaic efficiency of CO (FE_CO), while maintaining > 90% FE_CO over 24 h at -3.2 V, demonstrating exceptional stability under industrial-relevant current densities [Figure 3D and E]. PtAg₂₄(IPBT)₁₈ displays inferior catalytic activity with a smaller current density and lower FE_CO. Operando ATR-SEIRAS measurements were then conducted to investigate the activity differences between the two clusters. The adsorbed CO (^*CO) was observed in a range of 1,850-1,810 cm^-1 for both PtAg₂₄(IPBT)₁₈ and AuAg₂₄(IPBT)₁₈. However, at a given applied potential, PtAg₂₄ clusters exhibit a red shift of the ^*CO wavenumber by ~10 cm^-1 compared to AuAg₂₄. This suggests a stronger interaction between ^*CO and the PtAg₂₄ cluster, which is consistent with the slower ^*CO desorption. AuAg₂₄ exhibits relatively weak ^*CO adsorption and easier desorption, which contributes to a higher CO current density. Finally, density functional theory (DFT) calculations were performed to elucidate the influence of the central atom on CO₂RR performance. Figure 3F presents the Gibbs-free energy profiles of CO₂RR on AuAg₂₄ and PtAg₂₄ nanoclusters, revealing that the ^*COOH → ^*CO step is the rate-determining step. AuAg₂₄ exhibits a lower energy barrier (0.63 eV) compared to PtAg₂₄ (0.72 eV), indicating more favorable CO formation kinetics. Figure 3G further compares the adsorption Gibbs free energies of ^*CO₂ and ^*CO intermediates, showing weaker ^*CO adsorption on AuAg₂₄ (0.32 eV) than on PtAg₂₄ (0.26 eV), which facilitates the CO desorption. These theoretical insights align with the experimental observations, confirming that the single Au atom at the core of the icosahedral Ag₁₂ unit electronically modulates the Ag shell to optimize the intermediate binding, thereby enhancing both the activity and selectivity for CO₂-to-CO conversion^[113]. This work showcases the single-atom doping in tuning the electrocatalytic performance of AuAg nanoclusters, and provides a deep mechanistic understanding of the single-atom-doping effect.

It is worth noting that the CO₂RR performance of the AuAg₁₂@Au₁₂, AuAg₂₄ cluster-based catalysts is at least comparable with, if not superior to, most of the recently reported Au/Ag-based nanostructure catalysts under identical or similar conditions, in terms of FE_CO, CO current density, and long-term stability [Table 2]. To sum up, for AuAg nanoclusters in CO₂RR, the metal core composition and configuration, as well as the employed surface capping ligands, are the main factors affecting the catalytic properties, as they can significantly tune the active sites, optimize the adsorption of reaction intermediates, and modulate the energy barrier of the rate-determining step.

Urea is a basic yet important chemical raw material, as it is not only an essential nitrogen fertilizer for crop growth but also serves as a critical feedstock for producing a series of valuable chemicals, such as urea-formaldehyde, barbiturates, and urea-melamine-formaldehyde resins^[120,121]. Currently, the industrial synthesis of urea heavily relies on the reaction between CO₂ and NH₃ at high temperatures (150 °C) and high pressures (150-250 bar). These harsh reaction conditions require expensive equipment and high energy input and, more importantly, generate significant CO₂ emissions, which are a major contributor to global warming^[122,123].

Electrochemical urea synthesis provides a new route for green urea production, as it can operate under mild conditions, utilize renewable electricity, and convert common wastewater contaminants (NO₃^-, NO₂^-, etc.) into valuable chemicals^[124-126]. Fundamentally, urea electrosynthesis is an extremely complicated process, and the key lies in the construction of the C–N bond. Atomically precise AuAg nanoclusters offer significant advantages as model catalysts for understanding C–N bond formation, as they provide two neighboring metal sites to stabilize key intermediates for both CO₂RR and nitrogenous compound electroreduction. More significantly, their atomically precise structure enables atomic-level insights into the structure-performance relationship^[127,128].

In 2024, Chen et al. reported two Ag-based bimetallic clusters - Ag₁₄Pd(PTFE)₆(TPP)₈ (Ag₁₄Pd) (PTFE: pentafluorothiophenol, TPP: triph-enylphosphine) and Ag₁₃Au₅(PTFE)₁₀(DPPP)₄ (Ag₁₃Au₅) [dppp: 1,3-Bis(diphenylphosphino)propane] - and their catalytic performance toward urea electrosynthesis^[129]. Figure 4A and B presents the crystal structures of the Ag₁₄Pd and Ag₁₃Au₅ clusters, respectively. Ag₁₄Pd consists of a Pd@Ag₁₂ icosahedral center capped by two AgS₃P units and eight TPP molecules. The electrocatalytic urea synthesis was then carried out in 1 M KOH containing 200 ppm NO₃^- with CO₂ gas flowing over the chamber. The urea Faradaic efficiency (FE_urea) and yield rate in the potential range of -0.076 to -0.576 V are illustrated in Figure 4C and D. When the applied potential becomes more negative, the FE_urea for both Ag₁₄Pd and Ag₁₃Au₅ first increases and then decreases. The urea yield rate of Ag₁₄Pd exhibits a similar trend, whereas the urea yield of Ag₁₃Au₅ continuously increases. Ag₁₄Pd achieves the highest FE_urea of 15.8% and urea yield rate of 143.3 mg·h^-1·g^-1 at -0.276 V, surpassing that of Ag₁₃Au₅. The authors then employed DFT calculations to elucidate the underlying reaction mechanism. Figure 4E displays the Gibbs-free energy profile for urea production on Ag₁₄Pd and Ag₁₃Au₅. It can be noted that Ag₁₄Pd possesses a lower energy barrier of 0.63 eV in the rate-determining step of NO → NOH during nitrate reduction to ^*NH₂, while Ag₁₃Au₅ has a much higher energy barrier of 0.81 eV. Additionally, Ag₁₄Pd stabilizes the key intermediates through lower adsorption energies: ^*NH₂ at -2.49 eV vs. -2.42 eV in Ag₁₃Au₅, and ^*CO at -0.19 eV compared to 0.01 eV for Ag₁₃Au₅. This enhanced adsorption capability facilitates the C–N coupling with a lower energy barrier of 0.48 eV from Ag₁₄Pd against that of Ag₁₃Au₅ (0.58 eV). Notably, Ag₁₃Au₅ exhibits stronger Ag → Au electron transfer of 0.81 e^-, which elevates the Ag oxidation state to +0.09 vs. +0.04 in Ag₁₄Pd. Such electronic polarization weakens the intermediate interactions and promotes the ^*NH accumulation, shifting the selectivity toward CO₂ reduction. By reducing the energy barriers, strengthening the intermediate adsorption, and regulating charge polarization, Ag₁₄Pd emerges as a high-performance catalyst for urea electrosynthesis^[129]. However, Ag₁₃Au₅ has demonstrated great potential for urea electrosynthesis, and through composition optimization and ligand engineering, improved catalytic performance toward urea electrosynthesis might be achieved^[129].

Recently, Zhang et al. reported a carbene-ligand-protected AuAg nanocluster of [Au₁₃Ag₁₂(PPh₃)₈(BMIm)₂I₈]SbF₆ (Au₁₃Ag₁₂), and studied the influence of the carbene ligand on the electrocatalytic urea synthesis^[70]. The overall structure of Au₁₃Ag₁₂ is shown in Figure 5A and B. The Au₁₃Ag₁₂ nanocluster features a metal core of two vertex-sharing icosahedra in an S-configuration. Au atoms occupy the core, and they are bonded to the NHC ligands via shorter Au–C bonds (2.075 Å vs. Au–P: 2.30-2.34 Å), while Ag atoms at axial positions coordinate with the iodine ligands through Ag–I bonds (2.624 Å). Interestingly, when anchored on NiFe-LDH, the Au₁₃Ag₁₂ nanoclusters exhibited outstanding electrocatalytic performance for urea synthesis, achieving a urea production rate of 29.5 mmol·g_cat^-1·h^-1 with a Faradaic efficiency (FE) of 34% at -0.5 V vs. reversible hydrogen electrode (RHE) [Figure 5C]. Such remarkable performance substantially surpasses that of AuAg nanoclusters stabilized by conventional ligands such as DPPP or Ph₃P. Furthermore, the electrospray ionization mass spectrometry (ESI) results confirmed that the CO₂ molecules can be chemically bonded onto the cluster surface [Figure 5D]. The authors also used the DFT calculations to study the electronic structure of the Au₁₃Ag₁₂ nanoclusters. Figure 5E-H presents the DFT-simulated l-Au-BMIm and l-Au-PPh₃ and their bond properties. It can be noted that the carbene ligands can strengthen the Au–C bonds (2.01 Å bonding length vs. Au-PPh₃: 2.30 Å) with higher bond energy (74.1 vs. 60.1 kcal/mol). There are frontier orbital overlaps between the Au orbitals and the carbene orbitals, specifically highest occupied molecular orbital (HOMO)-9 and HOMO-3, where HOMO-3 has an s character and HOMO-9 exhibits a p character. Specifically, a constructive overlap exists between Au’s d_xz orbital and C’s p_z orbital in the HOMO-9 orbital, while the HOMO-2 orbital has an s character similar to that of the HOMO-3 orbital in l-Au-BMIm. The orbital overlap between Au and carbene indicates strong electron delocalization, which can optimize the ^*CO and ^*NH₂ adsorption, reduce the energy barrier of C–N coupling, and eventually enhance the NO₃^-/HCO₃^- conversion synergistically^[70]. This study highlights the crucial role of electronic structure tuning of atomically precise AuAg nanoclusters in boosting both the activity and selectivity in electrocatalytic urea synthesis.

As a note, the electrocatalytic activity of the Au₁₃Ag₁₂, Ag₁₄Pd cluster-based catalysts toward urea electrosynthesis - including the FE_urea and the urea yield rate - is comparable to that of other Au/Ag-based nanostructure electrocatalysts, as compiled in Table 3.

Besides the above electrocatalytic reactions, AuAg nanoclusters have also demonstrated great potential in other catalytic reactions such as multiple-component coupling reactions^[135] and photoelectrochemical reactions^[136]. Yet, from the structural view, the (electro)catalytic performance of the AuAg nanoclusters for various reactions, especially the complicated CO₂RR and electrochemical urea synthesis, is mainly governed by the following factors: size, composition, charge state, metal core configuration, ligand type and spatial structure, as well as the metal-ligand bonding structure. These factors can significantly tune the catalytic active site, induce charge redistribution/polarization, optimize the adsorption capability of the reaction intermediates, and alter the energy barrier of the rate-determining step. These cross-reaction mechanistic principles must be considered in future AuAg nanocluster design for catalytic applications.