Recent advances in nanoalloys for selective electrochemical CO₂ reduction

Abstract

The electrochemical carbon dioxide reduction reaction (CO₂RR) represents a compelling strategy to close the anthropogenic carbon cycle by converting CO₂ into fuels and chemicals using renewable electricity. Despite significant advances, the performance of CO₂RR catalysts remains fundamentally constrained by linear scaling relationships in monometallic systems, which couple the adsorption energetics of key intermediates. This intrinsic limitation restricts independent optimization of reaction steps, leading to poor selectivity, high overpotentials, and persistent competition from the hydrogen evolution reaction (HER). Nanoalloy catalysts have emerged as a powerful platform to overcome these constraints by introducing atomic-scale heterogeneity that is inaccessible to single-metal surfaces. Through controlled integration of multiple elements, nanoalloys enable decoupling of adsorption energetics, generation of distinct active sites, and spatial separation of reaction functions, thereby allowing selective stabilization of intermediates and access to reaction pathways that are thermodynamically or kinetically disfavored on monometallic catalysts. In this review, we move beyond conventional composition-based classifications and present a product-centered framework that directly connects nanoalloy design principles with CO₂RR reaction pathways and target products. We highlight three central synergistic mechanisms that govern nanoalloy behavior: electronic effects arising from d-band center modulation, strain effects induced by lattice mismatch, and ensemble effects enabled by tandem catalysis. We discuss how rational control of these mechanisms, together with microenvironment engineering, provides leverage over local reaction kinetics and selectivity, spanning syngas production, C₁ oxygenate formation, and C–C coupling toward C₂₊ products. Finally, we outline remaining challenges and emerging opportunities in nanoalloy catalyst design, emphasizing strategies to bridge atomic-level mechanistic insight with durable, scalable systems for industrial CO₂ electrolysis.

---

Khanh Quang Nguyen

Khanh Quang Nguyen is currently a researcher in the Advanced Materials and Applications Research Group (AMA Laboratory) at HUTECH University, Vietnam. He received his Bachelor's degree in Chemistry from the VNUHCM-University of Science in 2023. He is currently pursuing a Ph.D. degree in Materials Science at the same institution. His research primarily focuses on the design of advanced nanomaterials for green energy applications, specifically exploring electrochemical CO2 reduction (CO2RR) into sustainable fuels using bimetallic and nanoalloy catalysts. Additionally, his interests encompass photoelectrochemical (PEC) water splitting for hydrogen evolution and photocatalysts for environmental remediation.

Hong-Huy Tran

Dr Hong-Huy Tran is currently a Postdoctoral Researcher at the University of Pennsylvania with a joint affiliation in the School of Engineering and Applied Science and the School of Dental Medicine. He obtained his BSc degree in Materials Science from the University of Science, Vietnam National University, in 2017. He then worked as a visiting scholar at Chung Yuan Christian University and a researcher at the Ho Chi Minh University of Technology in 2018. He obtained his PhD degree in Materials Engineering from the Université Grenoble Alpes, French National Center for Scientific Research (CNRS), France in 2022. He was awarded an International Fellowship from IDEX, Freiburg Rising Stars Academy 2025–2026, and Rising Stars in Soft and Biological Matter co-sponsored by the University of Chicago and the University of California San Diego. His research integrates soft matter physics, catalytic nanomaterials, and robotic control to create adaptive, multifunctional materials and dynamic systems to address challenges in health, sustainability, and advanced manufacturing.

Huan Tan Ngo

Huan T. Ngo is a PhD candidate in Electrical and Computer Engineering at Northeastern University. He received his BSE in Materials Science and Engineering from Case Western Reserve University, with a minor in Physics. His research spans thin-film materials processing, optical sensors, spectroscopic characterization for optoelectronic materials, and computational design tools via machine learning. Prior work includes electrochemical processing of titanium nanotube arrays, photovoltaics, and solid oxide fuel cell degradation. His current interests lie at the intersection of nanoscale materials engineering and device physics, with applications in photonics, biomedical sensing, wearable electronics, and electrochemical systems.

Viet Van Pham

Dr Viet Van Pham is an Associate Professor, Head of the Department of Science and Technology, and Director of the CIRTech Institute at HUTECH University, Vietnam. He also serves as the Group Leader of the Advanced Materials and Applications Research Group (AMA Laboratory). He received his PhD in Materials Science from the VNUHCM-University of Science in 2018. His current research focuses on the synthesis and characterization of novel nanomaterials for photocatalysis and electrocatalysis, with specific applications in NOx and CO2 reduction, wastewater treatment, and photo/electrochemical water splitting for hydrogen evolution. Dr Pham has published over 100 peer-reviewed articles and 6 books/book chapters. He actively serves in editorial roles for several prestigious journals under Springer and World Scientific. Furthermore, he has led multiple special and thematic issues as Lead Guest Editor for high-impact journals, focusing on nanomaterials, renewable energy, and sustainable development. His scientific contributions have been recognized with the Golden Globe Youth Science and Technology Award (2018) and the Promising Young Physicist Award (2021).

---

1. Introduction

Climate change driven by greenhouse gas emissions, particularly CO₂, has become one of today's most urgent global challenges. The escalating frequency of extreme weather events, rising sea level, and widespread ecological disruptions highlight the need for decisive, coordinated international actions to reduce CO₂ emissions. The 2015 Paris Agreement set the goal of limiting global temperature rise to well below 2.0 °C, with efforts to cap warming at 1.5 °C, and called on nations to achieve net-zero emissions by the mid-21st century.^1,2 Most recently, COP30 in Brazil (2025) reinforced this mandate by launching the “Belém Mission to 1.5 °C″ and committing to mobilize $1.3 trillion annually by 2035 to accelerate the implementation of low-carbon technologies.^3,4 To realize these ambitious commitments, major economies have solidified “Net Zero 2050” targets, highlighting the urgent demand for clean energy solutions and effective CO₂ mitigation technologies.^5,6

Within this context, the approaches based on CO₂ reduction reactions have emerged as a highly promising strategy. However, the conversion of CO₂ is fundamentally challenging due to the inherent thermodynamic stability of the molecule, which possesses a linear geometry and a high C=O bond dissociation energy.⁷ Traditional CO₂ conversion technologies, such as thermocatalytic hydrogenation, high-temperature reforming, and photochemical reduction, typically require harsh operating conditions (elevated temperatures and pressures or intensive light sources), resulting in high energy consumption and poor overall carbon efficiency.⁸ In particular, photochemical CO₂ conversion often suffers from intrinsically low conversion efficiency, which severely limits its practicality even under strong illumination.⁹ Moreover, the economic viability of these thermocatalytic process is constrained by two fundamental barriers,¹⁰ the high cost of green hydrogen feedstock-currently $6–$8 per kg compared to the U.S. DOE's ‘Hydrogen Shot’ target of $1 per kg (ref. 11) and the substantial thermodynamic energy input required to drive the reaction, which often challenges the NETL energy efficiency benchmark of 100 kJ per mol CO₂.¹²

These constraints hinder the potential large-scale deployment, making the electrochemical CO₂ reduction reaction (CO₂RR) an appealing alternative due to its ability to operate under ambient conditions and couple directly with renewable energy sources to convert CO₂ into value-added fuels and chemicals.^13–15 The product spectrum spans from C₁ compounds such as carbon monoxide (CO),^16–18 formic acid (HCOOH),^19–21 methane (CH₄),^22–24 to C₂ products such as ethylene (C₂H₄),^25–27 ethanol (C₂H₅OH),^27–29 and even C₃₊ products like n-propanol (C₃H₇OH).^30–32 However, selectivity remains the central challenge in CO₂RR. The reaction involves multiple competing pathways that share similar intermediates, often resulting in complex product mixtures and significant competition from the hydrogen evolution reaction (HER).^33,34 While activity and stability are also important, achieving high selectivity is the most critical bottleneck that limits practical CO₂ electrolysis.

To address these challenges, catalyst design plays a pivotal role. Over the past decades, monometallic catalysts have been extensively investigated; however, their performance is fundamentally constrained by intrinsic barriers, specifically: (1) linear scaling relationships (LSRs) that link the adsorption energies of key intermediates and (2) the consequent inability to selectively stabilize desired intermediates while suppressing competing reactions.^35,36 To overcome these specific limitations, nanoalloys have been proposed as an enabling approach. By strategically combining different elements at atomic and nanoscale levels, alloy catalysts can simultaneously exploit the synergy of the electronic, strain, and ensemble effects to modulate intermediate adsorption energies and create new active sites.^37–40 Furthermore, advanced nanostructural designs, such as core–shell, Janus, and heterostructures, enable precise spatial control over active site distribution, thereby strongly directing product selectivity.^41–45 Owing to these advantages, nanoalloys provide a versatile platform for tuning intermediate adsorption and directing CO₂RR pathways. However, despite rapid progress over the past decade, several critical challenges remain. Existing studies often investigate individual alloy systems or isolated synergistic effects; thus, a holistic perspective on how electronic, strain, and ensemble effects collectively govern product selectivity remains fragmented.

Moreover, recent perspectives have further crystallized the critical requirements for next-generation catalysts. Moon et al. recently highlighted that for industrial-scale deployment, decoupling catalyst synthesis from electrode fabrication, specifically via the deposition of pre-synthesized nanoparticles, is superior to direct growth methods.⁴⁶ This insight places a premium on the precise engineering of robust nanoalloy building blocks that can be scaled up as catalyst inks. Concurrently, Kim et al. emphasized that accessing high-value C₃₊ products requires overcoming significant kinetic barriers involving multi-electron transfers, necessitating not just system-level cascade designs but intrinsic surface sites capable of stabilizing complex intermediates.⁴⁷

In response to these evolving demands for precise, product-specific active site engineering, this review establishes a comprehensive, product-centered classification of nanoalloy catalysts for CO₂ electroreduction. Rather than solely surveying individual materials or synthesis methods, we organize nanoalloys according to their dominant target products, providing direct comparison of structure–performance relationships across different reaction pathways (Fig. 1). We begin by discussing the fundamental limitations of monometallic catalysts under LSRs, followed by an in-depth examination of the key synergistic effects in nanoalloys, including electronic effects, strain engineering, ensemble effects, tandem catalysis, and strategies for suppressing HER, all of which play crucial roles in governing product selectivity. Building on these principles, the main sections analyze representative alloy systems for each major CO₂RR product, highlighting mechanistic insights, structure–property trends, and design rules. Finally, we outline remaining challenges, knowledge gaps, and future opportunities for rational nanoalloy design to advance next-generation CO₂ electrolysis technologies. Ultimately, by classifying nanoalloys based on target products rather than elemental composition, this review provides a practical roadmap for decoding selectivity. This approach enables researchers to directly compare diverse structural strategies for specific pathways, facilitating the rational design of catalysts tailored for desired CO₂ reduction outcomes.

Fig. 1 Schematic representation of how nanoalloy synergistic effects govern product selectivity across CO₂RR pathways.

2. Fundamentals of CO₂RR

CO₂RR-based processes enable the conversion of CO₂ into value-added chemicals through the use of electrical energy, particularly when powered by renewable energy sources.^14,15 Beyond directly reducing CO₂ emissions, this technology provides a pathway for energy storage in the form of chemical bonds within high-value fuels, thereby expanding the potential for sustainable fuel and chemical production and contributing to the global goal of carbon neutrality.⁴⁸ The key technological advantage of CO₂RR lies in its feasibility under ambient temperature and atmospheric pressure, in sharp contrast to conventional CO₂ conversion technologies that typically require elevated temperatures and substantial energy input.^49,50

2.1. Thermodynamic characteristics

CO₂ is an exceptionally stable molecule, with carbon existing at its highest oxidation state. The C=O double bond in CO₂ has a very high bonding energy, approximately 750 kJ mol⁻¹.⁷ This implies that the reduction of CO₂ is an endothermic reaction, requiring a significant amount of energy input to break these strong bonds and form new products. In electrochemistry, this energy is supplied in the form of an electrode potential. As shown in Table 1, the standard reduction potentials for various CO₂RR pathways in aqueous solution at pH 7 are highly negative; for example, the reduction of CO₂ to CO is −0.53 V versus the standard hydrogen electrode (SHE), while reduction to CH₄ is −0.24 V vs. SHE.⁵¹ However, for the reaction to proceed at an appreciable rate, a much more negative applied potential than the standard reduction potential is required. The difference between the operational potential and the standard reduction potential is referred to as the overpotential. High overpotentials represent one of the major challenges of CO₂RR, as they lower the overall energy efficiency of the process.

Table 1 Electrochemical CO₂ reduction reactions and their corresponding electron/proton transfer numbers

Reaction	Electron/proton transfers	E° (V vs. SHE, pH 7)
CO2 + 2H^+ + 2e^− → HCOOH	2/2	E^0 = −0.610 V
CO2 + 2H^+ + 2e^− → CO + H2O	2/2	E^0 = −0.530 V
2CO2 + 2H^+ + 2e^− → H2C2O4	2/2	E^0 = −0.913 V
CO2 + 4H^+ + 4e^− → HCHO + H2O	4/4	E^0 = −0.480 V
CO2 + 6H^+ + 6e^− → CH3OH + H2O	6/6	E^0 = −0.380 V
CO2 + 8H^+ + 8e^− → CH4 + 2H2O	8/8	E^0 = −0.240 V
2CO2 + 12H^+ + 12e^− → C2H4 + 4H2O	12/12	E^0 = −0.349 V
2CO2 + 12H^+ + 12e^− → C2H5OH + 3H2O	12/12	E^0 = −0.329 V
2CO2 + 14H^+ + 14e^− → C2H6 + 4H2O	14/14	E^0 = −0.270 V
3CO2 + 18H^+ + 18e^− → C3H7OH + 5H2O	18/18	E^0 = −0.310 V
2H^+ + 2e^− → H2	2	E^0 = −0.420 V

---

2.2. Reaction mechanism and kinetic challenges

CO₂RR is a complex electrochemical process occurring at the triple-phase interface between the catalytic electrode (solid), the electrolyte (liquid), and the reactant (gas).^52–54 The reaction proceeds through a series of elementary steps: initially, dissolved CO₂ diffuses to the electrode surface, where it is adsorbed and undergoes electron/proton transfers to form intermediates, before finally desorbing as products. A critical kinetic bottleneck in this sequence is the initial activation of the linear CO₂ molecule to form the bent radical anion *CO₂˙⁻, which requires a large reorganization energy and a high overpotential.^49,55 Furthermore, the major kinetic challenge in aqueous media is the competition from the HER.^33,34,56 HER generally exhibits more favorable kinetics on many metallic surfaces; therefore, it directly competes with CO₂RR for electrons and active sites on the catalyst surface. Mechanistically, HER is a simple two-electron transfer process, whereas deep CO₂ reduction to high-value products involves complex pathways requiring up to 12 or even 18 electron/proton transfers (Table 1). Moreover, since protons are much smaller and more agile than CO₂ molecules, this competitive advantage is further amplified. Consequently, an effective catalyst must not only exhibit high activity but also high selectivity-defined by the faradaic efficiency (FE), representing the fraction of current utilized for a specific product relative to the total input current.⁵⁷ Strategies to achieve this typically involve destabilizing the *H intermediate to suppress HER while facilitating the multi-step CO₂ reduction pathway.⁵⁶

2.3. Performance metrics

To enable systematic comparison and reporting, studies commonly employ key performance indicators such as overpotential, current density, FE, stability, and selectivity.^58,59 Comprehensive reporting of these parameters is essential for standardized evaluation and to accelerate progress in the field. A summary of these metrics is provided in Table 2.

Table 2 Key performance metrics in CO₂RR

Metric	Definition	Formula	Unit	Significance
FE	Ratio of charge used to produce a specific product relative to total charge	FE = (nNF/Qtotal) × 100	%	Measures product selectivity
Current density (j)	Rate of electrochemical reaction per unit electrode area	jproduct = jtotal × FEproduct	mA cm^−2	Reflects production rate; critical for industrial scaling
Overpotential (η)	Difference between applied potential and equilibrium potential	η = Eapplied − E°′	mV	Indicator of energy efficiency; lower η means lower energy consumption
Selectivity	Ability to preferentially form a desired product	Commonly quantified by FE	—	Reduces product separation cost and enhances economic value
Stability	Duration of stable catalyst operation	Operating time (h) with stable FE and j	h	Critical for practical applications and operational cost

---

3. Intrinsic limitations of monometallic catalysts and the strategy of using nanoalloys in CO₂RR

3.1. Limitations of monometallic catalysts under LSRs

In CO₂RR, metals play a central role, not only serving as the surface for adsorption and activation of CO₂ molecules, but also determining the reaction mechanism and the selectivity of the final products.⁶⁰ Each metal exhibits distinct adsorption properties toward reaction intermediates, leading to a preference for the formation of specific products. Based on the main products, monometallic catalysts are generally categorized into four main groups, as summarized in the Table 3.

Table 3 Overview of monometallic catalysts for CO₂RR

Metal group	Main products	Key intermediates	Advantages
Pb, Hg, In, Sn, Cd, Tl, Bi	HCOOH	*OCHO	High selectivity for formate; weak competition from HER
Au, Ag, Zn, Pd, Ga	CO	*COOH, *CO	High CO selectivity
Ni, Fe, Pt, Ti	H2	*H	Strong HER activity, unfavorable for CO2RR
Cu	CO, HCOOH, CH4, C2H4, C2H5OH, …	*CO, *CHO, *COH, *OCCO, …	Unique ability to reduce CO2 to multi-carbon products

---

One of the simplest and most commercially promising pathways is the production of CO, a key C₁ product. The earliest seminal work on monometallic catalysts was reported by Hori et al.,⁶¹ who demonstrated that Au and Ag exhibit exceptionally high selectivity toward CO due to their favorable binding of COOH and weak adsorption of CO. Following this foundation, subsequent studies have systematically improved the performance of monometallic Au and Ag through nanostructuring, support engineering, and crystal-phase control. Representative advances include nanostructured Au catalysts achieving FE_CO values of ∼70–77% in neutral media, conductive-core architectures that enhance electron transport,^62,63 and the emergence of 4H-phase Au nanomaterials with FE_CO exceeding 90% at −0.7 V vs. reversible hydrogen electrode (RHE).⁶⁴ Similarly, Ag-based systems-especially triangular nanoplates, 3D foams, and ultrathin nanowires-have achieved FE_CO values approaching 95–99% with reduced overpotentials and industrially relevant current densities.^65–67

These findings indicate that even with a single metal, the design of unique crystal structures can significantly enhance performance; however, both Au and Ag ultimately remain limited to producing CO and are unable to effectively drive the formation of multi-carbon products. Such observations directly reflect the Sabatier principle: an ideal catalyst must bind reaction intermediates with a moderate strength, not too weak and not too strong.⁷¹ If the binding is too weak, reactants are poorly activated and rapidly desorb; conversely, if the binding is too strong, intermediates are strongly retained on the surface, leading to the poisoning of active sites and hindering subsequent reaction steps.⁷² This principle is often illustrated visually by a volcano plot, in which, on the weak-binding branch, activity is limited by the adsorption or activation step, while on the strong-binding branch, it is constrained by product desorption (Fig. 2a and b).⁶⁸ The most effective catalysts typically reside near the summit of the plot, where an optimal balance is achieved. The challenge for complex reactions such as CO₂RR is that a single metal may lie close to the peak for one product, yet far from the peak for another.⁷¹ In the case of Au and Ag, both possess nearly optimal binding energies for *COOH, placing them close to the volcano peak for CO formation (Fig. 2a). However, they fall on the weak-binding branch for *CO, causing this intermediate to desorb readily as CO gas. As a result, these metal surfaces cannot retain *CO long enough to drive C–C coupling toward multi-carbon products. Similarly, for post-transition metals such as Sn (Fig. 2b), the stronger affinity for *OCHO intermediates positions Sn near the volcano peak for HCOOH formation, accounting for the high formate selectivity of Sn-based catalysts in CO₂RR. However, from an economic perspective, C₂₊ products possess far greater energy density and market value, making the restriction to CO or formate production a major limitation for practical CO₂ electrolysis applications.⁷³

Fig. 2 (a and b) Volcano-plot correlations between partial current density and binding energy of *COOH and *OCHO intermediates, reproduced from ref. 68 with permission from American Chemical Society, copyright 2017. (c and d) Three-dimensional scaling relations among key pairs of intermediates (*COOH, *CO, *H) and (*OCHO, *OH, *O) on various metal surfaces, showing inherent linear coupling that limits selectivity tuning, reproduced from ref. 69 with permission from WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim, copyright 2018. (e) Schematic illustration of material modifications that selectively stabilize intermediates and break the linear scaling limitation of monometallic catalysts in CO₂RR, reproduced from ref. 70 with permission from Elsevier Inc, copyright 2023.

Copper (Cu) is the only metal capable of overcoming this limitation, as it resides at the very summit of the volcano plot for the *CO intermediate.⁷⁴ The binding energy between Cu and *CO is sufficiently strong to maintain a substantial surface coverage of *CO, yet not so strong as to hinder subsequent dimerization or hydrogenation steps. This delicate balance endows Cu with a unique ability to catalyze the formation of multi-carbon products.⁷⁵ However, Cu faces another critical challenge: poor selectivity. Because of its capacity to stabilize *CO and catalyze multiple reaction pathways simultaneously, the Cu surface typically generates a complex mixture of products. This makes it extremely difficult to selectively obtain a single high-purity product.⁷⁶

Ultimately, all monometallic catalysts are constrained by the intrinsic limitation of the LSRs (Fig. 2c and d).^69,77–80 This represents a deeper thermodynamic barrier, explaining why it is fundamentally challenging to simultaneously optimize every step in a complex reaction sequence on a single-metal surface. Fundamentally, LSRs are universal correlations between the adsorption energies of chemically similar reaction intermediates (e.g., *COOH, *CO, and *CHO, all of which bind through the carbon atom) across diverse catalyst surfaces, originating from bond-order conservation principles and the d-band model.^81–83 These relationships are predominantly established through first-principles computational methods, such as density functional theory (DFT),^81,82 and can be further validated experimentally by measuring the heats of adsorption using techniques like single crystal adsorption calorimetry (SCAC).⁸⁴ Because of these strong correlations, if a metal binds *CO strongly, it will also bind *COOH strongly. Consequently, the binding energy of one intermediate cannot be independently tuned without altering the binding energies of other intermediates in the reaction pathway. The implication of LSRs is a significant reduction in the degrees of freedom available in catalyst design, creating an extended Sabatier barrier that monometallic catalysts cannot overcome. For example, to optimize CH₄ formation, it would ideally be necessary to stabilize the *CHO intermediate relative to *CO. However, due to LSRs, any attempt to strengthen *CHO binding will simultaneously strengthen *CO binding, leaving the energy barrier between them nearly unchanged.^85–87

3.2. From limitations to the strategy of nanoalloys

The existence of scaling relationships compels researchers to search for materials capable of breaking or deviating from these thermodynamic constraints (Fig. 2e). A classic demonstration of this strategy was reported by Lu et al. on Mn–Ni heteronuclear dual-atom catalysts (DACs).⁷⁹ In conventional single-metal catalysts, simultaneously achieving strong *COOH binding to lower the activation barrier and weak *CO binding to promote desorption is impossible. In contrast, Mn–Ni DACs exhibit a synergistic “task division”: the Ni center, with a reduced valence state due to electron transfer from Mn, preferentially binds to the carbon atom in *COOH and *CO, facilitating intermediate stabilization and CO release. Meanwhile, the Mn center, in a higher valence state, interacts strongly with the oxygen atom in COOH, thereby promoting CO₂ protonation. This cooperative mechanism enables intermediates to shuttle between the two metal centers, breaking the LSRs and achieving a remarkable FE_CO of 98.7% at −0.7 V vs. RHE.

Such studies highlight the urgent need to develop new catalytic materials with the ability to tune electronic and geometric structures for more precise control over intermediate interactions. In this context, nanoalloys have emerged as a pivotal strategy. By integrating multiple elements and enabling structural design at the nanoscale, they offer superior flexibility in tailoring electronic properties, providing abundant active sites, and opening reaction pathways that monometallic catalysts remain challenging to achieve.

To fully realize these sophisticated nanoscale designs, selecting an appropriate synthesis strategy is a crucial prerequisite. Currently, various methodologies have been employed to fabricate nanoalloys, each offering distinct advantages, limitations, and applicable scenarios in precisely controlling the composition, morphology, and atomic arrangement of the catalysts, as systematically summarized in Table 4.

Table 4 Summary of mainstream synthesis methods for nanoalloys

Synthesis method	Operating principle	Advantages	Disadvantages	Applications
Chemical reduction	Employs reducing agents (e.g., NaBH4, ascorbic acid, hydrazine) to reduce metal ions into their metallic states	- Facile operation, cost-effective, and easily scalable for mass production	- Often involves toxic chemicals	- Colloidal synthesis of well-dispersed alloy nanoparticles^88–90
		- Allows precise control over particle size and elemental composition	- Difficult to co-reduce metals with large differences in reduction potentials, which may lead to phase segregation rather than alloying without proper ligands	- Fabrication of core–shell structures via successive reduction^91
Magnetron sputtering	Utilizes plasma (typically Ar) to bombard metal targets, depositing an alloy film onto a substrate	- Very high product purity	- Requires expensive vacuum equipment and high energy consumption	- Direct deposition of nanoalloy films onto gas diffusion layers (Cu–Sn alloy nanofilm;^92 CuAg electrodes;^93 Cu/Ag and Cu/Al alloys^94)
		- Precise control over elemental composition	- Less morphological diversity and challenging 3D shape control compared to wet-chemical methods
		- High uniformity and reproducibility for large-area, self-supporting electrodes
Electrochemical deposition	Simultaneous reduction of metal ions from an electrolyte directly onto the electrode surface driven by an applied potential/current	- Avoids contamination from chemical reducing agents	- Certain electrolytes may generate byproducts or toxic gases during plating	- Fabrication of self-supported, porous alloy architectures directly on working electrodes (mesoporous AuCuNi ternary alloy films;^95 porous Cu–Ni alloy;^96 porous Cu–Pd alloy layers;^97 PtCu alloy^98)
		- Composition and size can be tuned via current density, applied potential, or pH	- Difficult to achieve atomic-level uniformity for complex multicomponent alloys
		- Enhances adhesion between the catalyst layer and the electrode
Galvanic replacement	Uses a metal with a lower reduction potential as a sacrificial template to provide electrons for reducing ions of a more noble metal	- Extremely powerful technique for fabricating hollow structures to maximize 3D spatial accessibility and surface area	- Often requires subsequent thermal treatment steps to ensure complete alloying and phase uniformity	- Fabricating nanoarchitectures (hollow Au–Cu nanorods;^99 Cu–in nanocrystals;^100 AuAg alloy nanoparticles;^101 highly alloyed AuAg aerogel^102)
Laser Ablation in liquid (LALT)	A high-energy laser vaporizes a solid metal target and decomposes metal salts in solution, followed by rapid quenching	- “Green” synthesis with fast reaction times	- Requires expensive, specialized laser equipment	- Fabrication of highly metastable, defect-rich nanoalloys (RuAu SAA;^103 HEAs nanoparticles;^104 AuFe alloy nanoparticles^105)
		- Ultra-fast quenching overcomes thermodynamic limits, enabling the alloying of classically immiscible metals	- Challenging to scale up for massive industrial production
Carbothermal Shock (CTS)	An ultrafast process involving instantaneous melting, decomposition, and solidification of precursors (within milliseconds) at extreme temperatures (∼2000 K)	- Extremely fast processing time	- Requires extreme peak temperatures and specialized equipment capable of generating massive heating/cooling rates	- Developing multi-element high-entropy alloys (HEAs) uniformly dispersed on carbon supports (PtPdRhRuCe HEAs nanoparticles;^106 FeCoNiPtRu HEAs;^107 RhRuFeCoNiMo HEAs nanoparticles/WO3 nanofibers^108)
		- Capable of uniformly mixing multiple immiscible elements to form perfect solid solutions
Support-assisted synthesis	Carbon networks, MOFs, or polymer frameworks act as spatial confinement and stabilizing agents during metal ion reduction	- Provides spatial confinement that effectively controls particle dispersion and limits agglomeration	- Complex preparation process that is highly dependent on the quality and intrinsic properties of the support materials	- Ideal for fabricating rare-earth nanoalloys, where rare-earth metals possess extremely negative reduction potentials and are highly prone to oxidation^109
		- Successfully inhibits the oxidation of highly reactive metals
Thermal decomposition	Simultaneous decomposition of organic/inorganic metal precursors at appropriate temperatures without additional reducing agents	- Produces nanoalloys with high crystallinity at relatively low temperatures	- Requires precursors with matching decomposition temperatures to ensure homogeneous alloying	- Fabrication of fine-grained, highly crystalline alloy nanoparticles, particularly for noble metal alloys (PdPt alloy nanodendrites^110)
			- Certain precursors are highly toxic
Sonochemical synthesis	Utilizes high-intensity ultrasound to induce acoustic cavitation, creating extreme local temperatures and pressures upon bubble collapse	- Reactions can proceed under normal ambient external conditions	- Difficult to control cavitation uniformity across large volumes, limiting large-scale mass production	- Surface modification of catalyst particles to boost activity and the synthesis of bimetallic nanoalloys with narrow size distributions.^111
		- Facilitates the formation of metastable or amorphous structures

---

Ultimately, an optimal synthesis strategy endows nanoalloys with superior catalytic properties, and the foundation of this superiority lies in unique synergistic effects, cooperative interactions between different components or active centers that enhance overall performance beyond what individual components can achieve. These synergistic effects may manifest through: (i) cooperative action, where active centers coordinate via intermediate transfer; (ii) synergistic interaction, in which the coexistence of multiple catalytic sites enables new reaction pathways; and (iii) mutual promotion, where different active centers accelerate charge, proton, or intermediate transfer.¹¹² These interactions are externally expressed through three principal synergistic effects: the electronic effect, the strain effect, and the ensemble effect. Nanoalloys can overcome the challenges of monometallic catalysts by employing a variety of design strategies.

3.2.1. Electronic effect and d-band center engineering. The electronic effect arises when the presence of hetero-metal atoms near the active center alters the local electron density. In other words, chemical interactions between dissimilar atoms redistribute the surface electronic structure. This effect is typically described in terms of shifts in the d-band center or modifications of the density of states (DOS) near the Fermi level of the host metal. The origin of this phenomenon lies in unique metal–metal interactions, which create coordination microenvironments absent on monometallic surfaces. The resulting electronic perturbations and orbital hybridization between heteroatoms alter the energy of the d-band center while simultaneously restructuring the local geometric and electronic states of the active sites.¹¹³

In CO₂RR, the d-band center acts as a central descriptor governing the adsorption strength of intermediates, thereby dictating both catalytic activity and product selectivity. Fig. 3a illustrates the basic principle of this theory: orbitals from an adsorbate interact with the metal's d-band to form bonding and antibonding states. The position of the d-band center (E_d) relative to the Fermi level (E_F) determines the filling of these antibonding states, which in turn dictates the bond strength.¹¹⁴ Specifically, when the d-band shifts upward toward the Fermi level, the adsorption of intermediates becomes stronger, favoring deeper hydrogenation steps and the formation of C₂₊ products. Conversely, when the d-band shifts downward, the interaction with *CO weakens, enabling facile CO desorption and enhancing CO selectivity.¹¹⁸ Both DFT studies and experimental evidence have demonstrated that d-band tuning can be achieved through various strategies, including metal alloying, lattice strain, heterostructure construction, or the incorporation of single-atom catalysts. Among these, nanoalloys offer a versatile platform to finely tune the electronic structure of metals. By systematically varying the atomic composition and coordination environment, the d-band center can be shifted continuously, allowing for the precise optimization of binding energies of key intermediates.⁸⁹ Such tuning not only directs desired reaction pathways but also suppresses competing reactions such as HER.

Fig. 3 Illustration of d-band center engineering to control product selectivity in CO₂RR. (a) Schematic principle of the d-band center theory, reproduced from ref. 114 with permission from Springer Nature Limited, copyright 1995. (b) Product selectivity map as a function of E_d and applied potential, and (c) the corresponding mechanisms on Ag–Cu surface alloys, reproduced from ref. 115 with permission from Wiley-VCH GmbH, copyright 2023. (d) Experimental valence band XPS showing the d-band center shift in Au–Cu alloys, reproduced from ref. 116 under CC-BY-NC-ND. (e) Calculated PDOS showing the upward d-band center shift in B-doped CuIn alloy, reproduced from ref. 117 with permission from Dalian Institute of Chemical Physics, the Chinese Academy of Sciences, copyright 2023.

For instance, doping Cu with a more electronegative metal such as Sn withdraws electron density from Cu, shifting its d-band downward and weakening the adsorption of *COOH and *OCHO. Conversely, alloying with a less electronegative element such as In donates electrons to Cu, shifting its d-band upward and enhancing *COOH adsorption.¹¹⁸ This upward shift effect is also confirmed in Boron-doped CuIn systems, as shown in the calculated projected density of states (PDOS) in Fig. 3e, where both Cu 3d and In 3d bands shift closer to the Fermi level.¹¹⁷ In Au–Cu alloys, increasing Au content shifts the d-band center further from the Fermi level, as clearly demonstrated by the experimental data in Fig. 3d, where the d-band center shifts linearly from −2.82 eV (pure Cu) down to −4.46 eV (pure Au).¹¹⁶ This downward shift weakens adsorption interactions and favors CO production, while the remaining Cu sites still provide sufficient binding to stabilize *COOH.^69,116 Similarly, in Pd–Au systems, the presence of Au lowers the d-band of Pd, thereby reducing *CO binding and promoting efficient CO desorption.⁸⁹ In Cu–Ni and Cu–Ag alloys, interactions with oxygenated intermediates are weakened, which enhances selectivity toward liquid carbonyl products while suppressing HER.¹¹⁹ Furthermore, in Co_xNi_1−x/N-CNFs, modulation of the d-band center significantly alters intermediate binding energies, with Co_0.75Ni_0.25/N-CNFs achieving FE as high as 85.0% for CO at moderate overpotentials.¹²⁰

A clear example of how electronic effects govern product selectivity in nanoalloys is demonstrated in the study by Wei et al.¹¹⁵ In Ag–Cu surface alloys, adjusting the Ag/Cu ratio tunes the d-band center, directly controlling the *CO binding strength. This relationship allows for systematic control over the final product, as illustrated in the 3D selectivity map in Fig. 3b, which maps product FE based on E_d and applied potential. A Cu-rich surface (e.g., Ag₁₆Cu₈₄), corresponding to a high E_d and strong *CO binding (Fig. 3c₁), facilitates C–C coupling to form *COCO and yields C₂ products with an FE up to ≈60%. At intermediate compositions (e.g., Ag₄₃Cu₅₇) with a moderate E_d and moderate *CO binding (Fig. 3c₂), C–C coupling is suppressed, but hydrogenation to *CHO is favored, thus shifting selectivity to C₁ products (CH₄ and CH₃OH) with an FE of ≈41%. Finally, at Ag-rich compositions (e.g., Ag₈₃Cu₁₇) with a low E_d and weak *CO binding (Fig. 3c₃), leading to facile desorption before further reaction, resulting in high selectivity for CO (≈74% FE).

Advanced alloy designs have also demonstrated combined electronic and geometric effects. For example, B-doped CuIn catalysts, as seen in Fig. 3e where the Cu 3d E_d shifts from −2.807 eV to −2.746 eV, exhibit optimized electronic structures that provide abundant active sites and improved charge transfer kinetics, enhancing CO₂-to-CO selectivity over a wide potential range.¹¹⁷ Similarly, Zeng et al. demonstrated that Pd–Ag alloy electrocatalysts could surpass the constraints of LSRs for efficient CO production.¹²¹ By fine-tuning the alloy composition, optimizing at Pd_0.75Ag_0.25, they modulated the surface electronic structure to weaken the overly strong binding of *CO on Pd sites. This balanced interaction facilitated the desorption of products while maintaining high activity for CO₂ activation, resulting in a superior FE_CO of 95.3%. In another approach, a thermally induced “atomic replacement” strategy was utilized to transform PtNi encapsulated in Zn-ZIF-8 into PtZn intermetallics coupled with isolated Ni single atoms.¹²² DFT calculations revealed a pronounced synergistic effect between adjacent PtZn and Ni₁ sites. PtZn stabilized the COOH intermediate via Pt–O interactions, thereby significantly lowering the protonation barrier on nearby Ni single-atom sites, while the cooperative PtZn–Ni₁ configuration simultaneously facilitated CO desorption.

3.2.2. Strain effect and surface strain engineering. The strain effect, more specifically lattice strain, arises from atomic size mismatch or lattice incompatibility between different components of a catalytic material. Strain can be categorized into two main types: tensile strain, when the interatomic distance increases, and compressive strain, when this distance decreases.¹²³ Beyond inherent atomic size differences, lattice strain can also be deliberately introduced through advanced nanoscale structural design, which is now recognized as one of the key principles governing the overall performance of catalytic systems. One of the most effective strategies for intentionally inducing strain is the core–shell architecture. A representative study by Sun et al. designed a charge-asymmetry “armor” catalyst (Pd₁Fe SAA@PC), in which a Pd₁Fe single-atom alloy (SAA) core is encapsulated by a phosphorus-doped carbon (P-doped C) shell.¹²⁴ Both experimental characterizations and theoretical calculations confirmed that this P-doped carbon layer imposed a tensile lattice strain of approximately +1.2% on the metallic core. DFT results further demonstrated that this optimal strain substantially increased the energy barrier for over-hydrogenation, thereby underscoring the critical role of lattice strain in governing product selectivity. Another approach is to introduce surface vacancies. Guo et al. convincingly demonstrated this principle in nanoporous PtCu alloys.¹²⁵ Using electrochemical etching, they selectively removed Cu atoms from the surface, generating a high density of Cu vacancies. Structural analyses revealed that these vacancies induced compressive lattice strain in the neighboring Pt lattice. DFT further revealed that Cu vacancies modified the adsorption properties of Pt sites, strengthening binding to desired intermediates (HO*) while weakening binding to undesired intermediates (O*), thereby enhancing catalytic activity.

These structural distortions directly influence the electronic structure of active sites, primarily through modulation of the d-band center. For late transition metals (d-band more than half filled, such as Cu, Ag, Au), tensile strain reduces orbital overlap, narrows the d-band, and shifts its center upward toward the Fermi level, strengthening interactions with intermediates. In contrast, compressive strain shifts the d-band downward, weakening adsorption.¹³¹ For early transition metals (d-band less than half filled), the trend is reversed.¹²³ This fundamental mechanism is illustrated in Fig. 4a, which shows that tensile strain causes an upward d-band shift for late transition metals (strengthening adsorption) but a downward shift for early transition metals (weakening adsorption).¹²⁶ By deliberately tuning the type and magnitude of strain in nanoalloys, the adsorption energies of key intermediates can be modulated, enabling simultaneous control over both activity and selectivity in CO₂RR.^132–134 For example, on bismuth-based catalysts, tensile strain on pristine Bi sites strengthened *OHCO adsorption and lowered the reaction barrier, achieving FE above 90% for HCOOH.¹³² Similarly, Bi@Sn core–shell structures with compressive strain were shown to promote CO₂-to-formic acid conversion.¹³⁵ In copper-based systems, strain plays a crucial role in regulating CO₂RR,¹³³ while rare-earth dopants can induce tensile strain in CuO_x catalysts, enhancing efficiency toward C₂₊ products.¹³⁶

Fig. 4 Strain-induced modulation of electronic structure and catalytic selectivity across representative CO₂RR nanoalloys. (a) Schematic illustration of how tensile strain shifts the d-band center of early and late transition metals, reproduced from ref. 126 with permission from Macmillan Publishers Limited, copyright 2017. (b and c) Theoretical evidence in Cu–Bi system showing that tensile strain facilitates CO₂ activation and stabilizes OCHO intermediates, reproduced from ref. 127 with permission from Wiley-VCH GmbH, copyright 2024. (d–g) Potential-dependent CO selectivity, valence band XPS spectra, and Gibbs free energy profiles for the CuAgSn alloy revealing that lattice distortion and d-band downshift increase *H adsorption energy, suppressing the competing HER and enhancing CO formation, reproduced from ref. 128 with permission from Royal Society of Chemistry, copyright 2022. (h–j) AuPd alloy: controlled phosphorization expands the Pd core to generate tunable tensile strain, enhancing FE_CO through d-band upshift and lowered COOH formation barrier, reproduced from ref. 129 with permission from Elsevier B.V., copyright 2025. (k) Illustration of strain-relaxation-driven optimization, where moderate lattice relaxation balances COOH formation and CO desorption, reproduced from ref. 130 with permission from American Chemical Society, copyright 2022.

A notable example of doping-induced strain effects is demonstrated by Yu et al., who incorporated Cu into the Bi lattice to introduce tensile strain, enabling the optimized Cu_1/6-Bi catalyst to achieve a formate FE efficiency exceeding 95%, an industrially relevant current density (−317 mA cm⁻²), and excellent operational stability beyond 120 hours.¹²⁷ As shown in Fig. 4b and c, theoretical analyses demonstrate how Cu incorporation modifies the Bi lattice and the reaction energetics. Fig. 4b depicts DFT-calculated free-energy profiles for the *CO₂ → *OCHO → *HCOOH pathway, showing that the Cu_1/6-Bi (2.4% tensile strain) surface significantly lowers the activation barrier compared to pristine Bi. Fig. 4c shows the corresponding PDOS, where hybridization between Bi p and Cu d orbitals stabilizes OCHO intermediates near the Fermi level. Collectively, these results demonstrate that tensile strain not only facilitates CO₂ adsorption but also stabilizes key intermediates, leading to superior selectivity.

Strain engineering can also be applied to suppress competing HER. Du et al. fabricated CuAgSn alloy electrodes via magnetron co-sputtering.¹²⁸ Incorporation of Sn into the CuAg lattice induced significant lattice distortion and compressive strain, which downshifted the d-band center and substantially increased the Gibbs free energy for *H adsorption. As a result, HER was strongly suppressed, and CO selectivity improved markedly, with FE reaching 93% at −0.85 V. Fig. 4d–g provide comprehensive experimental and theoretical evidence supporting this mechanism. Fig. 4d presents FE_CO values across potentials, confirming the superior CO selectivity of CuAgSn over CuAg. The XPS spectra in Fig. 4e reveal a clear downshift of the d-band center after Sn incorporation, consistent with DOS calculations, which further confirm that the d-band center of CuAgSn (−2.42 eV) lies lower than that of CuAg (−2.28 eV). The free-energy diagram in Fig. 4f and g shows that COOH formation remains energetically favorable, while *H adsorption becomes less favorable, explaining the strong HER suppression and high FE_CO.

In another study, Jia et al. developed a continuously tunable tensile strain strategy (1.7–4.4%) on AuPd alloys by synthesizing Pd@Au₄Pd₁ core–shell structures and expanding the Pd core via phosphorization.¹²⁹ Electrochemical measurements showed that at 4.4% tensile strain, FE_CO increased to 94.1% (at −0.55 V vs. RHE), compared to only 75.9% for the strain-free sample, illustrating a near-linear increase in CO selectivity with strain. The theoretical origins of this performance enhancement are elucidated in Fig. 4h–j. Fig. 4h summarizes the DFT-calculated d-band centers, showing a progressive upshift with greater lattice expansion. Fig. 4i presents the corresponding free-energy diagrams for *CO₂ → *COOH → *CO pathway, indicating that tensile strain lowers the *COOH formation barrier while facilitating *CO desorption. Finally, Fig. 4j illustrates the physical mechanism by which phosphorization expands the Pd core, generating controlled tensile strain that optimizes the binding energies of intermediates. Together, these results demonstrate that tensile strain tuning enables simultaneous enhancement of CO₂RR activity and HER suppression.

However, recent studies have shown that the relationship between strain and catalytic activity is not strictly linear. Rather than simply maximizing strain, a more nuanced approach is to identify an optimal strain level. In line with the Sabatier principle, excessive strain may over-stabilize intermediates, hindering CO desorption and reducing overall activity. To address this, Hao et al. proposed a thermodynamically driven strain relaxation strategy.¹³⁰ They synthesized PdNi alloy nanoparticles within electrospun carbon nanofiber nanoreactors, where controlled annealing temperatures induced atomic rearrangements that relaxed lattice strain from 3.2% to an optimal 2.3%. Fig. 4k vividly illustrates this concept. As the synthesis temperature increases, the PdNi lattice relaxes, enabling improved atomic mixing and strain equilibration. The relaxed s-PdNi/CNFs-1000 catalyst exhibits a FE_CO of 96.6% at −0.88 V vs. RHE, compared with only 68.9% for the more strained s-PdNi/CNFs-800 sample. DFT calculations confirm that moderate strain relaxation optimizes *COOH formation and *CO desorption, leading to the best balance between activity and selectivity. This work highlights strain relaxation as a powerful design principle for simultaneously optimizing *COOH formation and *CO desorption, thereby enhancing both activity and selectivity in CO₂ reduction.

3.2.3. Ensemble effect and tandem catalysis. The ensemble effect is a fundamental concept in heterogeneous catalysis, describing the change in activity when the geometric configuration of atomic ensembles varies with chemical composition. DFT calculations have shown that this phenomenon can be described using a linear interpolation model: the adsorption energies of intermediates on alloy surfaces are approximately the weighted average of the constituent atoms. This means that by combining two metals, the adsorption energies of intermediates can be tuned according to an “average scaling,” thereby breaking the inherent limitations of monometallic catalysts.¹³⁷ A classic experimental demonstration of this effect is the in situ STM and electrochemical study of PdAu(111) surface alloys by Maroun et al., where the concept of a “critical ensemble” was introduced-defined as the smallest atomic ensemble required for a specific reaction step.¹³⁸ The study showed that CO adsorption only required a single Pd atom (monomer), whereas *H adsorption required at least a pair of adjacent Pd atoms (dimer). This finding highlighted that even a simple change from *CO to *H adsorption could completely alter the geometric requirements of the active site at the atomic level. This principle later became the foundation for later developments in tandem catalysis concepts.

In tandem catalysis, instead of considering only a single reaction step, different atomic ensembles are “assigned functions”: one active site is optimized for the initial step, while another nearby site is designed for the subsequent step.¹⁴² As illustrated schematically in Fig. 5a, this cooperative architecture enables sequential CO₂-to-CO and CO-to-C₂₊ conversion on spatially adjacent yet electronically distinct sites. The first site (e.g., Ag) efficiently activates and reduces CO₂ to CO, while the second site (typically Cu) binds the CO intermediate more strongly, facilitating C–C coupling to form multicarbon products. Such functional differentiation between weak CO-binding and strong CO-binding centers ensures efficient intermediate transfer, and balances adsorption energies. A prominent example is the PTF(Ni)/Cu catalyst, in which isolated Ni–N sites efficiently reduce CO₂ to CO. The CO then migrates to adjacent Cu nanoparticles, which serve as secondary active sites for C–C coupling, producing ethylene with high selectivity (FE of C₂H₄ = 57.3%).¹⁴³ Following this principle, various bimetallic tandem strategies have proven effective in tuning activity and product selectivity. Kang et al. controlled selectivity between HCOOH and CO by varying the Ni/Co ratio in Ni–Co alloys.¹⁴⁴ In situ infrared spectroscopy and theoretical analysis revealed that Ni sites favored *OCHO formation (leading to HCOOH), while Co sites favored *COOH (leading to CO). A Ni-rich surface (Ni₇Co₃) achieved HCOOH selectivity of 98.5%, while a Co-rich surface (Ni₃Co₇) shifted selectivity to CO with 85.4%.

Fig. 5 Ensemble and tandem catalysis mechanisms in CO₂RR. (a) Conceptual schematic of the general tandem mechanism showing sequential CO₂-to-CO and CO-to-C₂₊ conversion through CO spillover and C–C coupling between complementary active centers; (b) hierarchical Cu-Ag-NC tandem electrode enhancing CO retention and C–C coupling, reproduced from ref. 139 with permission from American Chemical Society, copyright 2025. (c) Multilayer Cu₉₀Al₁₀/Cu₂₉Ag₇₁ architecture achieving stable C₂₊ selectivity in acidic media, reproduced from ref. 94 with permission from Wiley-VCH GmbH, copyright 2025. (d) Cu–M (M = Au, Ag, Pd) interfaces modulating product distribution through differential intermediate binding, reproduced from ref. 140 under the Creative Commons CC BY license. (e) Schematic synthesis of Ag–Pd nanoalloys, (f) XPS spectra showing Pd 3d downshift with Ag incorporation, and (g) composition-dependent FE_CO identifying optimal Ag₃₅Pd₆₅ performance, reproduced from ref. 141 with permission from Wiley-VCH GmbH, copyright 2023.

Similarly, deposition of Au, Ag, and Pd on Cu(200) surfaces, normally inert, created Cu–M interfaces that enriched Cu⁺ centers and modulated intermediate adsorption.¹⁴⁰ As illustrated in Fig. 5d, Cu–Au stabilized *COOH and *CO intermediates, generating high local *CO coverage that promoted C–C coupling, yielding C₂H₄ with FE ≈ 43.2%. In contrast, Cu–Ag interfaces bound much weaker to *CO, favoring facile desorption and resulting in high CO selectivity (FE ≈ 48.0%). Meanwhile, Cu–Pd interfaces lowered the barrier for *OCHO formation, steering the reaction toward HCOOH with FE up to ≈50.7%. These results highlight that designing multifunctional active sites in a single nanoalloy is an effective strategy for directing reaction pathways.

For a systematic view, tandem catalysis strategies can be classified in two ways. First, based on the physical structure of active sites, they include: (1) atom–atom dual-site tandem catalysis, typical of alloys where two metals interact directly at the atomic level; (2) atom-particle dual-site tandem catalysis, combining isolated single atoms with nanoparticles to exploit interface effects; (3) particle–particle dual-site tandem catalysis, usually physical mixtures of two types of metal nanoparticles; and (4) heterogeneous interface dual-site tandem catalysis, where unique effects arise at the junction between two distinct phases.¹⁴⁵ Second, based on the spatial relationship between catalytic components, they can be categorized as: (i) heterogeneous interfaces, where one component is directly anchored to the host metal surface, facilitating spillover of intermediates due to minimal distance; (ii) mixed catalysts, where two types of particles are blended, with performance depending mainly on geometry and diffusion distance of intermediates rather than electronic effects; and (iii) core–shell structures, where one phase encapsulates another, enabling control of intermediate transport.¹⁴⁶ For tandem catalysis to operate efficiently, spatial design that ensures effective transfer of intermediates from one site to another is crucial. A systematic study by Van Der Veer et al. clarified the importance of nanoscale proximity in Cu–Ag catalysts prepared by sputtering.⁹³ Direct comparison of different configurations showed that layered structures (Cu-on-Ag or Ag-on-Cu), though tandem in principle, exhibited poor C₂₊ production due to restricted CO spillover between layers. In contrast, co-deposited alloys, where Cu and Ag atoms were in direct contact, delivered superior C₂₊ efficiency (up to 75%), demonstrating that nanoscale intimacy between sites is critical for successful tandem catalysis.

One of the most sophisticated applications of tandem catalysis is overcoming stability and selectivity challenges in strongly acidic media. Zhang et al. developed a multilayer tandem electrode to promote C₂₊ formation under such harsh conditions.⁹⁴ They found that individual Cu–Ag and Cu–Al alloys were unstable due to dealloying. However, during operation, Cu₂₉Ag₇₁ spontaneously restructured into a stable Ag(111)/Cu(100) interface, achieving very high FE_CO (93.1%). Leveraging this, they designed a tandem electrode with a Cu₉₀Al₁₀ underlayer (highly active for C–C coupling but unstable) coated with the self-restructured Cu₂₉Ag₇₁ layer, as depicted in Fig. 5c. In this configuration, the Cu₂₉Ag₇₁ layer served as both an efficient CO “factory” and a protective shield against acid corrosion, while Al atoms in Cu₉₀Al₁₀ strongly adsorbed OH⁻, creating a locally alkaline environment that stabilized the alloy and lowered the barrier for C–C coupling. This synergistic mechanism enabled the tandem electrode to reach C₂₊ FE of 81.2% at a current density of 648 mA cm⁻², with a single-pass carbon efficiency (SPCE) of 70.4% and stable operation for 30 hours in acidic conditions.

Although tandem catalysis effectively increases local CO concentration, an inherent challenge remains: CO intermediates can desorb prematurely before participating in C–C coupling, reducing C₂₊ efficiency. To address this, Bian et al. developed hierarchical tandem catalysis to “trap and stabilize” CO.¹³⁹ As shown in Fig. 5b, they fabricated an electrode consisting of a bottom Ag layer (for CO₂-to-CO conversion) beneath a top layer of nitrogen-doped carbon (NC)-modified Cu nanowire arrays. Geometrically, the nanowire array extended the CO diffusion path, increasing residence time. Simultaneously, Cu/NC interfaces enhanced CO trapping and stabilization through linear adsorption, confirmed by in situ spectroscopy. This strategy yielded C₂₊ FE up to 87.5% at high current density, with a C₂₊/C₁ ratio of 10.42 and remarkable stability for 50 hours, highlighting how geometrical and interfacial confinement strengthen tandem synergy.

To further enhance tandem catalysis, recent strategies have focused on controlling the microenvironment surrounding active sites. A representative example is the a-Ni/Cu-NP@CMK catalyst reported by Chen et al.¹⁴⁷ In this system, Cu nanoparticles (∼3.2 nm) were confined within hydrophobic mesoporous carbon (CMK-8), pre-doped with atomically dispersed Ni–N₄ sites. This design provided dual synergy: (i) tandem catalysis, where Ni–N₄ efficiently produced CO from CO₂, which was then supplied to adjacent Cu nanoparticles for C–C coupling into ethylene; and (ii) microenvironment engineering, where the hydrophobic CMK-8 restricted water access, strongly suppressing HER and enhancing CO₂RR selectivity. As a result, a-Ni/Cu-NP@CMK achieved ethylene FE of 72.3% at a high current density of 406.1 mA cm⁻² in neutral electrolyte. Another unique approach is the use of polymer coatings to create tandem effects by controlling the microenvironment. Instead of directly participating in reactions, the polymer regulates the local chemical environment at the catalyst/electrolyte interface. For instance, hydrophobic polymer layers on Cu surfaces enrich CO₂ while excluding H₂O near active sites, effectively suppressing HER and enhancing adsorption of intermediates critical for C–C coupling. Additionally, functional groups in polymers can hydrogen-bond with oxygenated intermediates, stabilizing them and steering selectivity toward products such as ethanol.¹⁴⁸ This strategy opens a new paradigm in which tandem effects arise not only from cooperation between catalytic centers but also from subtle interactions between catalysts and tailored polymer overlayers.

Palladium alloying is also an effective strategy for optimizing CO production. While Pd readily activates CO₂, it suffers from overly strong *CO binding that hinders desorption. To address this, Zeng et al. alloyed Pd with Ag, a weaker CO-binding metal (Fig. 5e–g).¹⁴¹ Fig. 5e shows the synthesis route of ultrasmall Ag–Pd nanoparticles (∼3.58 nm) via oleylamine/oleic acid reduction, forming mixed-ensemble surfaces. The resulting ensemble sites containing both Ag and Pd effectively reduced the energy difference between *COOH and *CO. This mechanism either weakened *CO adsorption or strengthened *COOH adsorption relative to pure Pd, facilitating the catalytic cycle. Electronically, alloying with Ag shifted the Pd d-band downward, stabilizing d-electrons and weakening *CO binding (Fig. 5f). Owing to the combined ensemble, electronic, and size effects of ultrasmall nanoparticles (∼3.58 nm), the optimized Ag₃₅Pd₆₅ alloy catalyst achieved FE_CO of 98.9% at −0.8 V vs. RHE (Fig. 5g).

Nevertheless, a persistent challenge in studying the ensemble effect is that alloying often simultaneously alters both electronic structure and geometry, making it difficult to decouple their roles. Gong et al. proposed an alternative approach: comparing different polymorphs of the same Pd₃Bi alloy with identical composition.¹⁴⁹ This allowed geometric effects to be isolated from electronic ones. The study found that ordered Pd₃Bi intermetallics formed small Pd ensembles (1–5 atoms) and were nearly inactive for methanol oxidation reaction (MOR), whereas disordered solid-solution Pd₃Bi created larger ensembles (average ∼5.25 atoms, maximum 6) and exhibited high activity (∼0.5 mA cm_Pd⁻²). Partially ordered structures gave intermediate performance (∼0.1 mA cm_Pd⁻²). XPS and XANES analyses showed nearly identical electronic structures across samples, confirming that activity differences originated purely from ensemble size. This provides compelling evidence that the geometry of active sites can switch catalytic activity “on/off” independently of electronic structure, opening new pathways for designing catalysts with tunable active site ensembles while preserving electronic character.

3.2.4. Inhibition of HER. One of the main challenges of CO₂RR is suppressing the competing HER. Nanoalloy designs can tailor catalyst surfaces to minimize H₂ formation.¹⁵⁰

A representative demonstration of alloying to suppress HER in acidic media is the “metalloid-metal single-atom alloy” Te₁Bi.¹⁵¹ Introducing the nonmetallic Te atom into a Bi host created antagonism sites with dual effects: (i) geometric hindrance preventing *H coupling and thus suppressing H₂ formation; and (ii) OH⁻ adsorption as a proton source, spatially separating proton and electron supply, thereby favoring *OCHO formation at Bi sites and steering selectivity toward HCOOH instead of HER. As illustrated in Fig. 6a, the Te₁Bi single-atom alloy was synthesized via the reduction and alloying of Bi₂O₃ and Te under a 5% H₂–Ar atmosphere, generating an atomically dispersed Te within the Bi lattice. This atomic arrangement introduces alternating Bi and Te sites that modulate surface charge distribution and proton accessibility (Fig. 6b). The Te sites act as electronically negative centers that repel protons and block *H coupling, while neighboring Bi atoms serve as active sites for *CO₂ activation and *OCHO stabilization. The synergistic Bi–Te interaction therefore reconstructs the catalytic microenvironment-suppressing HER and enhancing CO₂ reduction toward formate. Operando Raman, SR-FTIR, and NAP-XPS measurements, combined with theoretical calculations, revealed significantly higher *H adsorption barriers on both Bi sites and Te sites relative to pure Bi (Fig. 6d), while the energy pathway for *OCHO formation was lowered (Fig. 6c). As a result, Te₁Bi achieved FE_HCOOH of 94.5%, SPCE of 40%, and j_HCOOH ≈ 0.2 A cm⁻² in strong acidic electrolyte, far surpassing pure Bi, which mainly produced H₂. This study highlighted that atomic-level alloying can deliberately control the microenvironment and proton pathways to simultaneously suppress HER and enhance CO₂RR selectivity.

Fig. 6 Alloying strategies for suppressing HER during CO₂ electroreduction. (a) Synthesis of Te₁Bi single-atom alloy via reduction and alloying; (b) reaction mechanism of Te₁Bi showing Bi–Te antagonistic sites that hinder *H coupling and promote *OCHO formation; and (c and d) DFT energy diagrams showing higher *H adsorption barriers and stabilized *OCHO intermediates on Te₁Bi, reproduced from ref. 151 with permission from Elsevier Ltd, copyright 2024. (e) Formation of Cu–Zn nanorods by electroreduction of Cu₂O/ZnO precursors on carbon cloth; and (f) Free-energy profiles for HER and CO₂RR on Zn, Cu, and Cu–Zn, reproduced from ref. 152 with permission from Wiley-VCH GmbH, copyright 2024.

To further elucidate the synergy, Liu et al. combined experiments and DFT to demonstrate that Cu–Zn alloys balance CO₂RR and HER by assigning specialized roles to each metal.¹⁵² As shown in Fig. 6e, the Cu–Zn alloy was synthesized by electrodepositing Cu₂O beneath ZnO nanotube arrays on carbon cloth, followed by electroreduction that converted Cu₂O/ZnO precursors into metallic Cu–Zn alloys. This in situ alloying places Cu and Zn in direct contact, enabling the two metals to assume complementary catalytic roles on the same interface. Cu lowered the energy barrier for the initial CO₂ activation to *COOH, a step difficult on pure Zn. In contrast, Zn weakened CO binding relative to pure Cu, facilitating CO desorption instead of deep reduction or surface poisoning. Meanwhile, Cu–Zn alloys exhibited moderate *H adsorption, preventing excessive HER and maintaining CO₂RR dominance. This behavior is quantified in Fig. 6f. On pure Cu, the adsorption free energy of *H (ΔG_*H ≈ −0.060 eV) is favorable, enabling rapid H₂ evolution, while CO desorption remains energetically demanding. On Zn, *H adsorption is too unfavorable (ΔG_*H ≈ +0.158 eV) and *COOH formation is difficult, limiting both HER and CO₂ activation. In contrast, the Cu–Zn alloy exhibits an intermediate *H adsorption energy (ΔG_*H ≈ −0.130 eV), high enough to slow HER, alongside a moderated barrier for CO₂ to *COOH and a lower barrier for *COOH to *CO. This “task division” at the atomic scale allowed flexible tuning of H₂/CO ratios in syngas (0.8–5.8) simply by adjusting the applied potential.

Beyond alloy composition, other approaches include electrode architecture to control HER. Zhang et al. reported a hollow-fiber Cu–Bi alloy electrode.¹⁵³ In this system, CO₂ was delivered “inside–out” through hollow fibers, creating a gas barrier that restricted electrolyte infiltration into pores. This limited local proton availability, strongly suppressing HER while maintaining high CO₂ concentration at the triple-phase boundary. Surface modification via alkaline oxidation and hydrothermal sulfuration yielded Cu₇S₄–CuBi nanoflowers with large electrochemical surface area and superior wettability. At −0.9 V vs. RHE, the electrode achieved FE_formate of 91.27% with a partial current density of 80.12 mA cm⁻², outperforming other Cu@Bi systems. The mechanism involved a rate-determining electron transfer forming CO₂*–, followed by protonation to HCOO*. The synergy of stable Cu⁺, sulfur incorporation, and hollow-fiber architecture promoted formate selectivity and maintained stability for 12 h of continuous electrolysis. This work demonstrated a self-supported electrode strategy that combined high performance, industrial-level current density, and effective HER suppression.

3.2.5. Enhancing stability. Long-term durability remains one of the major barriers to the commercialization of CO₂RR catalysts. As comprehensively analyzed in a recent review by Lai et al., the degradation of stability is a multi-scale challenge involving thermodynamic, kinetic, and structural factors that span from the atomic level to the entire electrolysis system.¹⁵⁴ Thus, alloying strategies not only enhance activity and selectivity but also play a vital role in extending catalyst lifetimes.

At the catalyst level, nanoalloys are particularly susceptible to atomic-scale reconstruction, phase segregation, and elemental leaching. For instance, as previously discussed in the work by Zhang et al.,⁹⁴ Cu–Ag and Cu–Al systems are prone to dealloying and instability under harsh conditions. Furthermore, Lai et al. highlighted that the dissolution-re-deposition process (Ostwald ripening) under cathodic potentials can cause particle agglomeration, significantly reducing the electrochemical active surface area.¹⁵⁴ These atomic-scale instabilities, particularly severe in Cu-based systems under reducing conditions, lead to rapid performance loss and increased HER.

To counteract these degradation mechanisms, intermetallic alloy design, with ordered atomic arrangements, mitigates phase segregation, preserves electronic interactions, and stabilizes intermediate oxidation states (Cu⁺/Cu⁰) critical for C–C coupling. Moreover, the secondary alloy atoms can act as “anchors” that limit Cu atom mobility, directly addressing the issues of Ostwald ripening and agglomeration mentioned above.¹⁵⁵ Thus, Cu–M intermetallics provide both electronic synergy (tuning adsorption energies of intermediates) and structural robustness against long-term degradation. A prime example is the work by Kuang et al. recently synthesized ordered CuAu intermetallics (o-CuAu) at low temperature (∼250 °C).¹⁵⁶ Results showed atomic ordering was key to stability: o-CuAu maintained FE_CO ∼60% for 160 h in a membrane electrode assembly (MEA) cell at 100 mA cm⁻², while disordered CuAu (d-CuAu) quickly lost selectivity due to restructuring into unstable Cu-rich phases. The superior stability of o-CuAu was attributed to high mixing enthalpy and strong Cu–Au interactions, which suppressed phase segregation. Furthermore, highly valent Au atoms in the intermetallic structure facilitated CO₂ activation and stabilized *CO₂˙⁻ intermediates.

Another effective solution to address structural instability during prolonged electrolysis is to introduce a small fraction of a third element to enhance durability. For example, Jia et al.¹⁵⁷ incorporated ∼3 at% Zn into the shell of Cu@CuAu core–shell catalysts. While Cu@CuAu rapidly lost activity, Cu@CuAuZn maintained high CO selectivity (FE 82%) and stable current for 10 h. Detailed analysis showed Zn atoms acted as “locks” in the alloy lattice. Based on solid solution strengthening, Zn increased the energy barrier for Cu and Au atom migration, preventing their dissolution. Post-reaction electron microscopy confirmed that Zn-containing catalysts experienced less corrosion and retained their core–shell structure.

Another demonstration of structural protection is the CoNi nanoalloy encapsulated in N-doped carbon nanotubes (CoNi@N-CNTs).¹⁵⁸ This design combined electronic synergy between Co and Ni with protective N-CNT shells, preventing corrosion, reconstruction, and deactivation under harsh conditions. CoNi@N-CNTs maintained FE_CO above 90% across acidic, neutral, and alkaline media, with current densities up to 732 mA cm⁻², while remaining stable for more than 36 h in acidic conditions. Mechanistic analysis revealed electron transfer from Co to Ni modified the electronic distribution, optimizing adsorption–desorption of intermediates, thereby enhancing CO₂RR selectivity and extending catalyst lifetime.

However, it is important to note that highly ordered intermetallics or heavily protected structures may also eliminate necessary atomic ensembles for catalysis. As demonstrated in Pd₃Bi, ordered phases were completely inactive compared to disordered solid solutions, despite identical composition and electronic structure.¹⁴⁹ This highlights a critical trade-off in catalyst design: stability must be balanced with the presence of active ensembles.

Beyond the catalyst itself, Lai et al. pointed out that stability is also heavily governed by micro- and macro-scale system failures, notably electrode flooding and salt precipitation.¹⁵⁴ In gas-diffusion electrodes (GDEs), the accumulation of liquid products or electrolytes within the pores obstructs CO₂ transport, while the local pH increase facilitates the formation of carbonate/bicarbonate salts that physically block the triple-phase boundary. To address these systemic challenges, operational regulation strategies must be deployed alongside material design. An excellent example of this synergy is the work by Van Der Veer et al., who applied pulsed electrolysis, alternating steady-state CO₂ reduction with brief oxidative pulses.⁹³ These pulses temporarily re-oxidized the surface, forming a stable Cu₂O layer, which not only slowed surface reconstruction but also actively mitigated electrode flooding. This extended catalyst lifetime from ∼1 h to 6 h, underscoring the potential of operational control to sustain performance. Therefore, achieving industrial-scale durability requires a holistic strategy that bridges robust nanoalloy engineering with optimized system integration and advanced operational protocols.

4. Tuning target products via nanoalloy design

One of the most prominent advantages of nanoalloys is their ability to tune product selectivity through careful design of catalyst composition and structure. Depending on the metal constituents, mixing ratios, nanostructure morphology, and electrochemical environment, nanoalloy systems can be directed to preferentially produce specific products.

4.1. C₁ products

In CO₂RR, C₁ products encompass a diverse spectrum, ranging from simple two-electron transfer species like CO and HCOOH to more deeply reduced and complex compounds such as HCHO, CH₃OH, and CH₄. Selectivity within this group is fundamentally governed by the initial competition between two key intermediates: the stabilization of *OCHO favors HCOOH formation, whereas the stabilization of *COOH directs the pathway toward *CO. Beyond initial CO desorption, driving the reaction toward deeper hydrogenation (HCHO, CH₃OH, or CH₄) significantly elevates the kinetic complexity. The selectivity among these challenging targets depends heavily on the catalyst's ability to precisely tune the binding strength of *CO and subsequent downstream intermediates, which ultimately determines whether the C–O bond is preserved or entirely cleaved. While the simpler C₁ products can often achieve very high FE due to their lower kinetic barriers, selectively synthesizing highly reduced C₁ oxygenates and hydrocarbons remains a formidable yet highly rewarding objective to maximize the energy density and commercial value of the outputs.

4.1.1. Carbon monoxide. CO is one of the simplest products of CO₂ reduction, requiring only a two-electron transfer. The widely accepted pathway for CO formation proceeds via the *COOH intermediate. The process begins with CO₂ adsorption and activation on the catalyst surface, followed by a coupled proton–electron transfer to form *COOH, where CO₂ binds through the carbon atom. The next step involves heterolytic C–OH bond cleavage in *COOH, yielding the *CO intermediate and releasing a water molecule. Finally, because of its relatively weak binding to many metals, *CO desorbs easily to form CO (Fig. 7).^{70,72,159,160} Thus, an effective CO-selective catalyst must stabilize *COOH while binding *CO weakly enough to prevent further hydrogenation.

Fig. 7 Reaction network of CO₂RR showing key intermediates and coupling pathways leading to C₁, C₂, and C₃ products.

Śliwa et al. employed a “host–guest” approach to synthesize AgCu nanoalloys with Cu content precisely controlled from 0% to 15%.⁹⁰ As illustrated in Fig. 8a, the synthesis involved the sequential formation of Ag(Cu^x+)Br host particles, co-reduction of interstitial Ag⁺ and Cu^x+ cations to generate metallic AgCu domains, and subsequent removal of the AgBr phase, yielding phase-pure “Cu-in-Ag” alloy nanoparticles with precisely tunable Cu composition. Experimental results revealed that CO selectivity correlated with alloy composition, peaking at 83.2% FE with 5% nominal Cu at −0.5 V vs. RHE (Fig. 8b). Increasing Cu content to 15% sharply decreased CO efficiency to 50%, while HER became dominant. This decline was attributed to the formation of Cu-rich phases beyond solubility limits, which are highly active for HER.

Fig. 8 (a) Host–guest synthesis method for AgCu nanoalloys, (b) dependence of FE_CO and total current on Cu composition, reproduced from ref. 90 with permission from American Chemical Society, copyright 2024. (c) Cu–Sb intermetallic formation via high temperature solid-state synthesis, (d) EXAFS spectra showing Cu–Sb coordination in Cu₂Sb and Cu₃Sb, (e) CO formation efficiency of Cu–Sb intermetallics, reproduced from ref. 161 with permission from American Chemical Society, copyright 2024. (f) Structure of AgCuAu ternary alloy, (g) PDOS of Cu d-band upshift upon alloying, (h) FE_CO of nanoporous Ag₅Cu₅Au₅ over various potentials, (i) free-energy diagram for *COOH formation, and (j) *H adsorption on AgCuAu, reproduced from ref. 162 with permission from Science Press and Dalian Institute of Chemical Physics, Chinese Academy of Sciences, copyright 2024.

An effective strategy to enhance CO selectivity is to form intermetallic compounds with p-block elements such as Sb. As shown in Fig. 8c, Huang et al. synthesized highly ordered Cu–Sb intermetallics, Cu₂Sb and Cu₃Sb, via high-temperature solid-state synthesis methods, enabling precise tuning of Cu coordination.¹⁶¹ They demonstrated that adjusting the coordination environment of Cu atoms by alloying with Sb effectively suppressed HER and C–C coupling, thereby significantly improving CO selectivity. Extended X-ray absorption fine structure analysis (EXAFS, Fig. 8d) revealed an increased Cu–Sb coordination numbers in Cu₃Sb (CN = 4.9) relative to Cu₂Sb (CN = 1.1), producing a distinct electronic environment around Cu active sites. Consequently, Cu₃Sb exhibited outstanding CO₂RR performance, achieving FE_CO up to 97.9% at −0.6 V vs. RHE, with an impressive CO partial current density of ∼42.8 mA cm⁻² (Fig. 8e). This catalyst also showed high stability in both flow cells and MEAs. In situ Raman spectroscopy further confirmed *COOH as the key intermediate, verifying a selective CO pathway without detectable HCOOH or C₂₊ products.

Moving beyond binary systems, multi-component alloys offer greater flexibility in fine-tuning electronic and geometric effects. Wang et al. combined DFT and experiments to design a ternary AgCuAu alloy (Fig. 8f) with superior CO selectivity.¹⁶² DFT analysis revealed that Cu atoms served as the primary active sites, while adjacent Ag and Au atoms modulated the local electronic structure, shifting the Cu d-band center closer to the Fermi level (Fig. 8g). This shift lowered the barrier for *COOH formation (Fig. 8i) and raised the barrier for *H adsorption (Fig. 8j), thereby enhancing CO selectivity while suppressing HER. To validate these predictions, the team synthesized nanoporous NP-Ag₅Cu₅Au₅ by dealloying. Experiments confirmed the predictions: the ternary alloy delivered FE_CO above 90% across a wide potential window of 0.6 V, peaking at ∼96% at −0.573 V vs. RHE (Fig. 8h). This far outperformed both monometallic and bimetallic counterparts. This work compellingly demonstrated the power of multi-component alloy design, where synergistic electronic effects can simultaneously optimize multiple steps, yielding catalysts with high activity, excellent selectivity, and wide potential stability.

An important industrial application of CO₂RR is controlled production of syngas (CO + H₂), a key feedstock for many chemical processes. The challenge lies in deliberately balancing CO₂RR to CO with HER to H₂. Nanoalloys of non-precious metals have proven to be an effective and economical strategy to address this. For instance, Song et al. investigated Cu–Fe alloys supported on N-doped carbon (CuFe/NC).¹⁶³ They demonstrated tunable H₂/CO ratios from 0 to 1.94 by adjusting the Cu : Fe ratio and applied potential. The optimal Cu₁Fe₂/NC catalyst achieved excellent performance, with FE_CO up to 98.91% at −0.7 V vs. RHE, while maintaining stability for 24 h. XPS analysis revealed the underlying mechanism: because Cu is more electronegative than Fe, electron transfer occurred from Fe to Cu, redistributing electronic density, shifting the Cu d-band, and optimizing intermediate binding energies to balance CO₂RR and HER.

Similarly, Cu–Zn alloys have emerged as promising candidates for syngas production. Guo et al. investigated Cu–Zn alloys with precisely controlled Zn loadings prepared by electrodeposition.¹⁶⁴ Systematic characterization revealed a strong composition-dependent electronic interaction between Cu and Zn. Ultraviolet photoelectron spectroscopy (UPS, Fig. 9a and b) showed that the cut-off and Fermi energies varied non-monotonically with Zn content, reaching optimal values for the Cu–Zn-675 sample. This composition exhibited the lowest work function (2.06 eV), indicating facile electron transfer across the Cu–Zn interface. Experimental results further show that the Cu–Zn-675 catalyst exhibited the highest CO₂ conversion (12.7%) and turnover frequency (TOF) (4.62 s⁻¹) (Fig. 9c), along with nearly 90% FE for syngas and a tunable CO/H₂ ratio up to 2.1 at −0.81 V vs. RHE (Fig. 9d). DFT analysis (Fig. 9e and f) revealed that the introduction of Zn tuned the adsorption energetics of key intermediates: the Cu–Zn surface exhibited moderate *COOH and *H binding, lowering the *CO desorption barrier and weakening *H adsorption relative to pure Cu. This balance between CO₂ reduction and H₂ evolution explains the experimentally observed CO/H₂ tunability. Overall, Cu–Zn alloys integrate the favorable CO₂ activation of Cu with the weak CO binding of Zn, creating a bifunctional interface that enables efficient and controllable syngas production with suppressed HER.

Fig. 9 (a) UPS spectra showing secondary electron cut-off edge, (b) Fermi edge, (c) CO₂ conversion and turnover frequency of Cu–Zn catalysts; (d) FEs and CO/H₂ ratios of Cu–Zn-675 at different potentials; (e) free-energy diagram for CO₂RR, (f) for HER on Cu, Zn, and Cu–Zn alloy surfaces, reproduced from ref. 164 with permission from Elsevier Ltd, copyright 2025. (g) Schematic illustration of Cu–Zn nanosheet synthesis via electrodeposition; (h) PDOS of Cu 3d and Zn 3d states in Cu_0.07Zn; (i) comparison of PDOS for Zn and Cu_0.07Zn; (j) charge density difference map of Cu–Zn alloy surface; and (k) Bader charge analysis of Cu–Zn alloy showing charge distribution between Cu and Zn atoms, reproduced from ref. 165 with permission from Wiley-VCH GmbH, copyright 2025.

Along the same lines, Zou et al. synthesized Cu–Zn nanosheet alloys via electrodeposition (Fig. 9g).¹⁶⁵ Their work demonstrated CO/H₂ ratio tunability from 1.1 to 4.3 by adjusting Cu : Zn ratio and applied potential, achieving nearly 100% FE for syngas with excellent durability. XPS and DFT analyses revealed electron transfer from Zn to Cu, which increased the density of Cu 3d states near the Fermi level. As shown in Fig. 9h, the PDOS analysis reveals that the Cu 3d states are positioned closer to the Fermi level than the Zn 3d states, reflecting an increased electron density surrounding Cu atoms. Moreover, in Fig. 9i, compared with pure Zn, the overall d-band of the Cu_0.07Zn nanosheets shifts toward the Fermi level, confirming that Cu incorporation enhances the electronic density near the active Cu sites and strengthens their capacity for CO₂ activation. The differential charge density map (Fig. 9j) reveals charge redistribution at the Cu–Zn interface, where electrons migrate from Zn to Cu, leading to electron-enriched Cu sites. Bader charge analysis (Fig. 9k) quantitatively supports these findings, showing approximately 1.71|e| transferred from Zn to Cu in the active CuZn₅ phase. This substantial charge migration modulates the surface electronic environment, resulting in an optimal balance between CO₂ activation and H adsorption.

4.1.2. Formic acid/formate. HCOOH or its salts (HCOO⁻) is a valuable two-electron CO₂ reduction product, widely used in industry and considered a promising hydrogen carrier. Mechanistically, the pathway for formate formation diverges clearly from the CO pathway at the very first steps. While CO formation proceeds through the *COOH intermediate, in which CO₂ binds to the catalyst surface via the carbon atom (C-bound), high selectivity toward formate relies on the catalyst's ability to preferentially form and stabilize the *OCHO intermediate, where CO₂ is adsorbed through oxygen–metal coordination (O-bound). Several pathways have been proposed for *OCHO formation. In the first (pathway 1), CO₂ accepts one electron to form an adsorbed radical anion (*CO₂˙⁻), which is then directly protonated to generate *OCHO. Alternatively, in pathway 2, CO₂ can interact with surface-adsorbed hydrogen in the form of a metal–hydride (M–H) bond, where CO₂ inserts directly into the M–H linkage to yield *OCHO. The *OCHO intermediate is subsequently hydrogenated and desorbed to produce HCOOH in acidic media or HCOO⁻ in alkaline conditions (Fig. 7).^{51,70,72,160,166} Therefore, an effective catalyst for the production of formate must not only stabilize *OCHO but also suppress competing pathways, particularly HER and *COOH formation.¹⁶⁷ p-block metals such as Sn, In, and Bi typically exhibit these properties, and alloying them with other metals can further enhance *OCHO stabilization at bimetallic interfaces.

Accordingly, a common strategy to enhance formate production is alloying p-block metals (e.g., Bi, Sn, In) with other metals to optimize electronic structure and reaction kinetics. Wu et al. reported a representative study on Bi–Cu alloys, synthesizing Bi₉Cu₁ nanosheet catalysts supported on carbon cloth (Bi₉Cu₁/CC) via electrodeposition (Fig. 10a).¹⁶⁸ The Bi₉Cu₁/CC catalyst demonstrated outstanding performance, achieving FE for formate above 90% across a wide potential range (−0.7 to −1.2 V vs. RHE) in a flow cell (Fig. 10b). In situ ATR-IR spectroscopy confirmed that the fundamental pathway remained via the *OCHO intermediate. However, DFT calculations revealed the electronic effect of Cu incorporation: introducing a small amount of Cu into Bi increased the density of states near the Fermi level, improving conductivity, optimizing *OCHO adsorption energy, and lowering the barrier for conversion to HCOOH at Bi–Cu interfacial sites (Fig. 10c and d).

Fig. 10 (a) Schematic illustration of the electrodeposition process for fabricating Bi–Cu alloys; (b) FE of formate for Bi₉Cu₁ sample at different potentials; (c) DFT-calculated free-energy diagram; and (d) schematic representation of *OCHO adsorption and hydrogenation on Bi–Cu interfacial sites, reproduced from ref. 168 with permission from Tsinghua University Press, copyright 2023. Gibbs free energy profiles on Cu–Sn alloy surfaces for (e) CO₂-to-HCOOH; (f) CO₂-to-CO, and (g) HER pathways, reproduced from ref. 92 with permission from Elsevier B.V., copyright 2024.

In addition to Bi, Sn is also distinguished by its high selectivity toward formate, and alloying Sn with other metals has been extensively explored to enhance the efficiency and durability of CO₂RR. However, one of the major challenges for Sn-based catalysts is the instability of the active oxide phase. Although high-valence Sn centers are considered to play a crucial role in formate formation, they are readily reduced to metallic tin (Sn⁰) under negative potentials, leading to activity loss and enhanced HER. To address this issue, an effective approach is interface engineering to stabilize the oxidation state of the active centers. In a recent study by Fan et al., the incorporation of Cu into tin oxide catalysts induced in situ reconstruction under electrolysis conditions, forming a core–shell structure with a Cu₆Sn₅ alloy core and an amorphous SnO_x shell.¹⁶⁹ In situ analyses combined with DFT calculations revealed that the Cu₆Sn₅/SnO_x interface plays a crucial role in stabilizing high-valence Sn species on the surface, preventing their reduction to Sn⁰ even under negative potentials and high current densities. The Sn⁴⁺-rich surface strongly promotes the formation and stabilization of the OCHO intermediate, the direct precursor of formate, whereas Sn⁰ sites preferentially catalyze HER or the CO pathway via COOH. Owing to this stabilization mechanism, the R-CuSnO₃ catalyst exhibited efficient activity across the entire pH range, achieving a maximum FE for HCOO⁻ of 93.4% at −0.9 V in H-cell (0.5 M KHCO₃). In a flow cell, the catalyst maintained 88.4% FE at a current density of 800 mA cm⁻² (corresponding to j_formate = 707 mA cm⁻², with a formate formation rate of 13.2 mmol h⁻¹ cm⁻² in 1 M KOH) and 87.8% FE at 300 mA cm⁻² in a strongly acidic medium (0.1 M H₂SO₄ + 1 M K⁺). Furthermore, operational stability was demonstrated with ∼90% FE maintained over 90 h at −100 mA cm⁻². In a complementary approach, Li et al. exploited dynamic reconstruction in heterogeneous catalysis, where the structure and surface composition of the catalyst evolve under reaction conditions to form the real active phase.¹⁷⁰ Although the initial material was SnO₂ doped with a small amount of Cu, in situ and ex situ analyses revealed that under the reductive potentials of CO₂RR, the catalyst spontaneously reconstructed to form Cu–Sn nanophases, specifically the Cu₆Sn₅ phase. In situ ATR-FTIR measurements recorded a distinct OCHO signal at 1390 cm⁻¹, while the CO signal was nearly absent, confirming that Cu₆Sn₅ is the catalytically active phase favoring formate. DFT calculations indicated that the formation of Cu₆Sn₅ optimizes the electronic structure of Sn, rendering Sn more positively charged (Sn^δ+), thereby lowering the Gibbs free energy barrier for OCHO formation, while simultaneously increasing the barrier for COOH and weakening H binding. As a result, the Cu-doped SnO₂ catalyst after reconstruction achieved FE_formate ∼92% at −0.9 V (H-cell), maintaining >80% in the range −0.8 to −1.1 V with >90% FE stability for 12 h. In the flow-cell (1 M KOH), the catalyst delivered a high current density of 220 mA cm⁻² at −1.2 V, sustaining FE_formate >92% across −0.6 to −1.0 V, and achieved an energy conversion efficiency (ECE) > 50% at 150 mA cm⁻². This indicates that more than half of the input electrical energy was converted into chemical energy stored in formate, a critical threshold for practical application. Moreover, the catalyst maintained stable operation for 14 h at −0.9 V. Overall, whether through stabilizing Sn^δ+ on the oxide layer via interface engineering or regenerating Sn^δ+ within the alloy lattice through dynamic reconstruction, both studies confirm that the persistent presence of Sn^δ+ is the key factor in promoting OCHO formation and enhancing the selectivity toward formate.

In addition to these strategies, another proactive design approach to enhance formate selectivity is the direct tuning of the electronic structure of the Cu–Sn system through alloying. Shan et al. fabricated Cu–Sn thin films with precisely controlled compositions via magnetron sputtering, successfully depositing uniform alloy thin films directly onto carbon paper.⁹² Among the compositions, Cu₄₁Sn₁₁ (Cu–Sn-3 sample, 20.6 at% Sn) exhibited the highest FE for HCOOH (74%) at −1.3 V vs. RHE. Experimental characterization and DFT calculations elucidated the mechanism: alloying with Sn downshifted the Cu d-band center, optimizing the adsorption energy of *OCHO while simultaneously destabilizing *H and *COOH intermediates. More importantly, the Gibbs free energy calculations for CO₂-to-HCOOH, CO₂-to-CO, and HER pathways, as shown in Fig. 10e–g, reveal a dual effect of the Cu–Sn alloy surface. On the Cu–Sn-3 surface, the energy barrier for *OCHO formation was significantly reduced (0.34 eV) compared with pure Cu (0.9 eV) and pure Sn (1.7 eV), whereas *H and *COOH formation became thermodynamically less favorable. This dual effect efficiently suppressed HER and CO evolution, enhancing the selectivity toward formate.

Indium is also a representative p-block metal with an inherent tendency toward formate production in CO₂RR. When designed as alloys, particularly in combination with Sn, the activity and selectivity of In can be significantly enhanced. However, conventional In–Sn catalysts still suffer from the low conductivity of the support and strong competition from HER. To address this limitation, Zhang et al. developed a novel support-engineering strategy by introducing a conductive zeolite framework to host ultrasmall In–Sn nanoclusters.¹⁷¹ As illustrated in Fig. 11a, Sn atoms were first substituted for Al in the Y zeolite framework through ion exchange, converting the insulating AlY structure into a conductive SnY matrix. Subsequent impregnation of In(NO₃)₃ and in situ alloying within the supercages yielded In_0.2Sn_0.8 nanoclusters (∼1.3 nm) uniformly confined inside the zeolite pores. This confinement effectively prevented particle aggregation and maintained excellent structural stability during electrolysis. Fig. 11b presents the calculated DOS, showing that the d-band center of In in InSn@SnY upshift toward the Fermi level relative to pristine InSn, indicating an enhanced electronic coupling between the alloy clusters and the SnY framework. Fig. 11c presents the calculated charge density difference map, revealing significant charge redistribution across the In–Sn/SnY interface. Specifically, the adjacent In atom and interfacial Sn atom in the SnY support exhibit electron depletion of −0.13 e⁻ and −2.17 e⁻, respectively. These electrons are donated to the adjacent Sn atom within the alloy, which shows an accumulation of +0.02 e⁻. This electron transfer triggers an electronic reconfiguration that upshifts the d-band center, strengthening the metal-adsorbate bonding and facilitating *OCHO intermediate stabilization (as illustrated in Fig. 11d). As a result of these synergistic structural and electronic effects, InSn@SnY achieved 98.2% FE for formate, an industrial-level partial current density of 322 mA cm⁻², and stable operation for over 102 h-demonstrating that In, when integrated with advanced support design, can serve as an efficient active center for large-scale formate production.

Fig. 11 (a–d) Formation and electronic structure of In–Sn nanoclusters confined in conductive SnY zeolite: (a) synthesis steps; (b) DOS; (c) charge redistribution; and (d) schematic of d-band upshift, reproduced from ref. 171 with permission from Wiley-VCH GmbH, copyright 2024. (e–g) Structure and preparation of Ga-based liquid metal alloys: (e) liquid-phase alloying process; (f and g) morphology of Ga–In and Ga–Sn droplets, reproduced from ref. 172 with permission from Wiley-VCH GmbH, copyright 2024. (h–n) Structure and mechanism of Pd–Cu nanodendrites: (h) morphology evolution with Pd/Cu ratio; (i) schematic of microfluidic synthesis; (j and k) charge distribution; (l) DOS analysis; and (m and n) performance comparison, reproduced from ref. 173 under the Creative Commons CC BY 4.0 license.

A further step in exploiting the potential of In is the development of SAAs, where individual atoms of a secondary metal are dispersed on the host to maximize atom utilization and precisely tune the electronic structure. Fang et al. reported an Ag₁In system stabilized by an oxygen-pinning (Op) mechanism for formate production.¹⁷⁴ In this system, isolated Ag atoms are anchored by O on the In host, forming covalent bonds that prevent aggregation and maintain structural stability. Mechanistically, Ag atoms act as electronic modulators, altering the local electronic structure of In. In situ ATR-SEIRAS, EIS, and DFT calculations demonstrated that this adjustment enables Op-Ag₁In to simultaneously (i) enhance CO₂ adsorption/activation and (ii) accelerate water dissociation to provide H, thereby facilitating OCHO formation. Consequently, the catalyst achieved ∼92% FE with a partial current density of 13 mA cm⁻² at −0.95 V, 2.23 times higher than pure In. In the flow-cell, Op-Ag₁In delivered 93.5% FE with a stable current density of 70–93 mA cm⁻² over 24 h. Notably, the oxygen-pinning mechanism not only promoted CO₂RR kinetics but also made HER energetically unfavorable by modulating H adsorption. These studies provide compelling evidence that alloying is a powerful tool for tailoring the electronic structures of p-block metals, thereby enhancing intrinsic activity and selectivity toward formate.

Beyond p-block metals, Pd is also an attractive candidate due to its activity at nearly zero overpotential. However, an intrinsic limitation of Pd is its poor durability, as its surface is easily poisoned by trace amounts of CO generated during the reaction. In a recent study, Huang et al. addressed this challenge by developing Pd–Cu nanodendrites with channel-rich structures.¹⁷³ As shown in Fig. 11h and i, the formation mechanism involves electrostatic attraction between [PdCl₄]²⁻ and CTAC to generate Pd-CTA⁺ complexes, which act as nucleation templates for anisotropic growth under mild ascorbic acid (AA) reduction, leading to the formation of channel-rich Pd–Cu alloy nanodendrites (∼25 nm) with high electrochemical surface area (∼97.4 m² g⁻¹, about 10 times that of pure Pd) and a K⁺-rich microenvironment that stabilizes *OCHO intermediates and accelerates CO₂ reduction kinetics. Charge-density difference mapping Fig. 11j and k demonstrates substantial electron transfer from the Pd(200) plane to the adsorbed AA molecule. In contrast, Cu(111) surfaces exhibit only minor charge redistribution. This transfer activates the AA reductant, promoting the reduction of Cu²⁺ and driving alloy formation. Corresponding DOS plots Fig. 11l reveal that the d-band center of Pd(200) (−1.946 eV) is closer to the Fermi level than that of Cu(111) (−2.555 eV), implying higher surface reactivity. This electronic modulation also stabilizes *OCHO intermediates, favoring the formate pathway over the competing *COOH and H₂ evolution routes. As a result, the optimized Pd₁Cu₁ nanodendrites exhibit an outstanding FE for formate of 97.7% at −0.1 V vs. RHE, maintaining FE above 90% across a wide potential window (−0.05 to −0.25 V) and reaching a partial current density exceeding 11 mA cm⁻² (Fig. 11m). The comparison in Fig. 11n further confirms that the Pd₁Cu₁ composition provides the best balance between formate selectivity and suppression of HER or CO pathways, outperforming Pd, Pd₂Cu₁, and Pd₃Cu₁ variants. The synergistic combination of morphology, electronic d-band tuning, and ion-enriched nanoconfinement creates a robust catalytic environment that maximizes CO₂ reduction efficiency. Consequently, the Pd₁Cu₁ nanodendrites maintain >90% formate selectivity for over 15 h at −0.2 V.

In addition to conventional solid-state nanoalloys, a promising direction is the exploration of liquid metal alloys for catalytic applications. These materials, typically based on gallium (Ga), combine metallic properties with fluidic flexibility and can be synthesized simply at room temperature in an energy-efficient manner. In a study by Huang et al., Ga–In and Ga–Sn liquid metal alloy nanoparticles were successfully prepared by dissolving solid In or Sn into liquid Ga, followed by ultrasonication to form stable nanodroplets (Fig. 11e–g).¹⁷² The results showed that these alloys exhibited much higher activity and selectivity for formate compared to the constituent metals. Specifically, Ga_0.75In_0.25 achieved a maximum FE of 82% for formate at −1.05 V, with a partial current density of 55 mA cm⁻² at −1.17 V, 4–6 times higher than Ga nanoparticles or In powder, and maintained ∼72% FE over 12 h at −1.05 V. Similarly, the Ga–Sn system demonstrated impressive performance, with Ga_0.90Sn_0.10 reaching 89% FE and 64 mA cm⁻² at −1.09 V, while sustaining ∼80% FE at an industrial-level current density of 300 mA cm⁻² in a flow cell. Density functional theory calculations revealed that surrounding Ga atoms create a unique electronic and ensemble effect around In or Sn sites, weakening *H and *COOH adsorption (thus suppressing HER and CO pathways), while optimizing OCHO binding, the direct precursor of formate. Consequently, the reaction pathway is strongly directed toward formate production.

In summary, the key to enhancing HCOOH/HCOO⁻ production lies in stabilizing the OCHO intermediate and suppressing HER/CO through various strategies such as alloying, interface engineering, dynamic reconstruction, SAAs, and liquid metal alloys. This confirms that alloying, particularly with p-block metals, is a powerful tool to tune the electronic structure and opens up industrial prospects for formate production from CO₂.

4.1.3. Methane. CH₄ formation, a deeply reduced product requiring eight electrons, originates from the key *CO intermediate. From *CO, subsequent hydrogenation (addition of a proton–electron pair) can yield two major intermediates: *CHO (C–H bond formation) or *COH (O–H bond formation). This branching point critically determines the downstream reaction pathway. Once *CHO or *COH forms, a sequence of further hydrogenation steps leads to the cleavage of the C–O bond and progressively saturates the carbon with hydrogen, ultimately releasing CH₄ (Fig. 7).^70,72,160

To promote deep hydrogenation toward CH₄, an effective strategy is combining Cu with a metal capable of strong *CO adsorption such as Ni. A representative study by Li et al. fabricated Ni–Cu alloys supported on N-doped carbon nanotubes (NCNTs).²² The optimized Ni₁Cu₁-NCNT catalyst exhibited a breakthrough performance in CO₂-to-CH₄ conversion, achieving FE of 99.7% at −1.2 V vs. RHE, among the highest reported for Cu-based alloys. The underlying mechanism was elucidated by both experimental and theoretical approaches. XPS analysis and DFT calculations jointly revealed electron transfer from Cu to Ni within the alloy, shifting the d-band center closer to the Fermi level relative to monometallic Ni-NCNT and Cu-NCNT catalysts. According to d-band theory, this enhanced the binding energies of reaction intermediates on the surface. In situ EC-ATR spectroscopy further confirmed the presence of *COOH and *CHO intermediates, validating that the Ni–Cu alloy created a favorable electronic environment to stabilize key precursors for CH₄ formation.

From a materials design perspective, directing C₁ selectivity (CO, HCOOH, CH₄) in CO₂RR largely depends on the competition between *COOH and *OCHO intermediates. If *COOH is preferentially stabilized, the pathway proceeds via *CO → *CHO, favoring CO production or further deep hydrogenation to CH₄. Conversely, if *OCHO is stabilized, the pathway is directed toward formate.¹⁷⁵ Thus, nanoalloy design strategies focus on tuning electronic and geometric/strain effects to selectively stabilize one of these intermediates, thereby “activating” or “suppressing” the corresponding branch. To demonstrate this principle, a breakthrough study by Go et al. introduced a thermodynamics-based methodology for deliberately controlling and synthesizing distinct phases of Cu–Sn alloys.¹⁷⁶ By adjusting the partial pressure of oxygen (pO₂) during annealing of electrospun polymer nanofibers containing Cu and Sn precursors, they produced a series of catalysts spanning mixed oxides (CuO–SnO₂), oxide-metal composites (Cu–SnO₂), and ordered intermetallic compounds on carbon nanofibers (CNFs), including Cu₄₁Sn₁₁/CNFs, Cu₃Sn/CNFs, and Cu₆Sn₅/CNFs. This study demonstrated that phase control effectively dictates product selectivity in CO₂RR. Specifically, CuO–SnO₂ favored formate, while Cu–SnO₂ exhibited high CO selectivity. Remarkably, intermetallic phases showed distinct product branching: Cu-rich alloys such as Cu₄₁Sn₁₁ and Cu₃Sn significantly enhanced CH₄ formation, with Cu₄₁Sn₁₁ achieving a maximum FE of 39.1%. In contrast, Sn-rich Cu₆Sn₅ preferentially produced formate with FE of 58.6%. DFT calculations clarified the mechanism, showing that the Cu-rich surface of Cu₄₁Sn₁₁ favored the *CO pathway, while Cu₆Sn₅ surfaces stabilized the *HCOOH intermediate.

Among bimetallic systems, Cu–Pd alloys have recently demonstrated remarkable methane selectivity by tuning both geometric and electronic interactions between Cu and Pd sites. Huang et al. reported that introducing Pd into Cu significantly altered the CO₂RR pathway, suppressing C₂H₄ formation while promoting CH₄ evolution.¹⁷⁷ Fig. 12a outlines the synthesis and structure of the Cu₃Pd aerogel, where Pd atoms are homogeneously incorporated into a Cu network to form an ordered fcc-type alloy. The underlying mechanism of Pd-induced methane selectivity is systematically illustrated in Fig. 12b. Pure Cu typically facilitates C–C coupling between adjacent *CO and *CHO intermediates, leading to C₂ products such as ethylene. However, in Cu–Pd alloys, Pd atoms act as spatial separators between Cu sites-introducing a blocking effect that hinders *C–C coupling and instead enhances *C–H hydrogenation pathways. Operando ATR-SEIRAS analysis revealed strong vibrational features corresponding to *COOH, *CHO, and *OCH₃ species without detectable *C–C coupling intermediates, confirming that hydrogenation dominates on Cu–Pd surfaces. Electronic-structure analyses support these observations. PDOS calculations in Fig. 12c show that alloying with Pd shifts the Cu 3d band center closer to the Fermi level by ≈0.74 eV, enhancing electron density and binding strength of adsorbates. Bader charge analysis reveals electron transfer from Cu to Pd, confirming that Pd becomes more electronegative and preferentially adsorbs H⁺, forming reactive *H species that mediate proton-coupled *CO hydrogenation (Fig. 12d). DFT calculations further demonstrated that *H species adsorbed on Pd atoms lower the reaction barrier for *CO → *CHO hydrogenation from 0.82 to 0.70 eV (Fig. 12e), while the energy barrier for *C–C coupling (*CO + *CHO → *COCHO) increased to 1.39 eV compared to 0.78 eV on pure Cu (Fig. 12f). These results quantitatively confirm that the presence of Pd not only modifies the d-band center and charge distribution (Bader analysis shows electron transfer from Cu to Pd, rendering Pd electronegative and favorable for H⁺ adsorption) but also geometrically inhibits C–C bond formation. Consequently, the Cu₃Pd catalyst achieved a maximum CH₄ FE of 43.2% at −1.8 V vs. RHE, with a partial current density of −270 mA cm⁻². These findings highlight that Pd-assisted hydrogen transfer (*H spillover) and site-isolation effects can synergistically steer the CO₂RR pathway toward deep C₁ reduction.

Fig. 12 (a) Formation process and structure of Cu–Pd aerogel alloys; (b) reaction schematic comparing CO₂ reduction and hydrogenation pathways on Cu and Cu–Pd surfaces; (c) PDOS showing electronic modulation after Pd incorporation; (d) charge distribution map illustrating electron transfer between Cu and Pd atoms; (e and f) free-energy profiles comparing *CO hydrogenation and *CO–*CHO coupling steps on Cu and Cu₃Pd, reproduced from ref. 177 under the Creative Commons CC BY license (g) product distribution and FE of Co₁Cu and Cu catalysts; (h and i) in situ Raman spectra showing CO adsorption configurations on Co₁Cu and Cu; (j and k) DFT analyses of *CO adsorption modes and reaction energy barriers on Co₁Cu surfaces, reproduced from ref. 178 with permission from Wiley-VCH GmbH, copyright 2025.

Recent progress in atomic-level alloying strategies has demonstrated that isolating heteroatoms within a Cu matrix can break the conventional trade-off between selectivity and current density for CO₂-to-CH₄ conversion. Li et al. developed a single-atom cobalt–copper alloy (Co₁Cu), where individual Co atoms are uniformly dispersed within a metallic Cu host.¹⁷⁸ This design achieved a CH₄ FE exceeding 60% at −800 mA cm⁻², with a peak partial current density (j_methane = −482.7 mA cm⁻²), outperforming pure Cu (Fig. 12g). EXAFS confirmed the formation of Co–Cu bonds (∼2.42 Å) and the absence of Co–Co coordination, indicating the atomic dispersion of Co atoms. Mechanistic analyses revealed that single Co atoms act as H₂O activation centers, promoting local proton generation and accelerating CO hydrogenation without relying on H-spillover mechanisms. Concurrently, in situ Raman spectroscopy (Fig. 12h and i) identified a shift in the CO adsorption configuration from *linear (CO_L, ∼2080 cm⁻¹) on Cu to *bridge (CO_B, ∼1860 cm⁻¹) on Co₁Cu, where stronger multi-site binding favors deeper hydrogenation to CH₄ rather than C–C coupling or CO desorption. Complementary DFT calculations corroborated these findings, showing that on the Co₁Cu (111)/(100) surfaces, *CO_B formation is thermodynamically preferred over *CO_L formation (Fig. 12j), and the barrier for *CO to *COH hydrogenation decreases. Moreover, the kinetic barrier for *CO–*CO dimerization (C₂₊ formation) increases substantially (Fig. 12k), rationalizing the exceptional CH₄ selectivity.

In addition to conventional alloying, mixed-valence modulation on copper surfaces has recently emerged as a powerful means to control CO₂RR product branching. Sun et al.¹⁷⁹ constructed an Ag@Cu₂O@Cu-BDC double-shell tandem catalyst featuring a Cu²⁺/Cu⁺ valence-state fluctuation buffer zone, which dynamically balanced hydrogenation and C–C coupling pathways. The Ag core produced CO intermediates, which migrated to the Cu²⁺/Cu⁺ interface for further reduction. By tuning the Cu²⁺/Cu⁺ ratio (2.46 → 1.28), product selectivity was switched from CH₄ (59.3%) to C₂H₄ (43.0%). In situ ATR-SEIRAS and DFT analysis revealed that Cu²⁺ sites facilitate *CO → *CHO hydrogenation, whereas Cu⁺ sites promote *CO–*CHO coupling to OCCHO, confirming their complementary roles in steering reaction pathways. Mechanistically, CO molecules generated on the Ag core diffuse outward to the Cu₂O/Cu-BDC interface, where they undergo selective conversion depending on the Cu oxidation state. Cu²⁺-rich domains favor methane production, whereas Cu⁺-dominated sites stabilize C–C coupling for ethylene formation. The Cu-BDC shell functioned as a “valence-state buffer,” preventing Cu⁺ reduction and maintaining catalytic stability for >20 h under continuous electrolysis. This study provides a model for alloy-like electronic regulation via oxidation-state engineering, bridging the concepts of bimetallic synergy and dynamic valence tuning in Cu-based CO₂RR systems.

In summary, achieving high CH₄ selectivity in CO₂RR requires promoting deep hydrogenation while suppressing C–C coupling. The strategies highlighted demonstrate several key approaches, ranging from electronic/valence engineering to optimize intermediate binding, to geometric/adsorption control to block dimerization or favor hydrogenation-prone *CO adsorption. Furthermore, proton kinetic enhancement is employed via *H spillover or direct H₂O activation. The most effective systems synergistically lower the hydrogenation barrier while raising the C–C coupling barrier, thereby funneling the *CO intermediate toward CH₄.

4.1.4. Methanol. Methanol is a vital platform chemical and a high-energy-density liquid fuel. The mechanism of methanol electrosynthesis shares its initial pathways with methane production via the key *CO intermediate. Following the reduction of CO₂ to *CO, subsequent hydrogenation yields intermediates such as *CHO or *COH (Fig. 7). To selectively drive the pathway toward methanol, the *CHO intermediate must undergo further proton-coupled electron transfers to form *CH₂O, followed by *CH₃O or *CH₂OH.^70,160,180 The critical determinant in this process is the successful hydrogenation of these intermediates while strictly preserving the C–O bond. Achieving high selectivity for methanol remains exceptionally challenging due to a severe microkinetic hurdle. The fierce competition between CH₃OH and CH₄ originates at the critical branching point of the *CHOH intermediate: while protonation at the carbon atom (which requires surface-adsorbed hydrogen) preserves the C–O bond to yield CH₃OH, the attack by electrolyte protons on the oxygen atom directly cleaves the C–O bond, releasing H₂O and diverting the pathway toward deep reduction to CH₄.¹⁸¹

To steer the reaction pathway at this critical juncture, researchers have widely employed nanoalloy catalysts to precisely modulate the local electronic microenvironment. These multimetallic systems leverage synergistic interactions to selectively stabilize oxygenated intermediates, thereby promoting the facile desorption of methanol while effectively suppressing C–O bond cleavage. For instance, phase-separated bimetallic systems have demonstrated remarkable synergistic effects. Lu et al. synthesized three-dimensional Pd–Cu bimetallic aerogels comprising crystallized fcc Pd and amorphous Cu domains.¹⁸² The strong interfacial synergy between these phases, characterized by favorable valence state distributions, effectively stabilized the key *CO₂˙⁻ intermediates, enabling the optimal Pd₈₃Cu₁₇ aerogel to achieve a methanol FE of 80.0%. Similarly, highly ordered intermetallic nanoalloys show great promise. Payra et al. developed carbon-supported mixed-phase Pt_xZn (1 < x < 3) nanoalloys that significantly outperformed their phase-pure counterparts (methanol FE of 81.4%).¹⁸³ The heterojunction interfaces facilitated single-electron transfer to activate CO₂, while the specific electronic structure weakened the binding of the *OCH₃ intermediate, promoting its prompt desorption to selectively yield methanol. Furthermore, doping offers a powerful avenue for electronic modulation. By introducing 3% Sn into In₂S₃ microflowers, Bhattacharya et al. induced electron transfer from Sn to In.¹⁸⁴ This created an electron-rich surface that stabilized *CH₂OH intermediates and strongly suppressed HER, representing a significant advance for Indium-based catalysts in alcohol production.

Complementing experimental efforts, integrating DFT with machine learning (ML) has accelerated the rational design of selective nanoalloys across vast compositional spaces. ML frameworks have identified key descriptors, such as Bader charge, metal–ligand bond length, and d-band center, that govern the competition between methanol and methane pathways. For example, data-driven screening revealed that low-concentration Rh-doped Cu alloys induce a 0.18 eV upward shift of the d-band center, strengthening *CHO adsorption and mitigating HER. Extending this approach to highly complex systems, high-throughput screening of 36 750 Cu–Co–Ni–Zn–Sn HEA combinations pinpointed 35 highly selective microstructures. In these computationally optimized nanoalloys, synergistic electronic effects favorably route the reaction through the *HCO intermediate rather than *COH, effectively enhancing methanol selectivity while suppressing unwanted C–C coupling and C–O bond cleavage.

4.1.5. Formaldehyde. In the context of CO₂RR, formaldehyde typically emerges as a transient intermediate within the reaction cascades directed toward methanol or methane rather than a terminal product.¹⁶⁰ The predominant mechanistic pathway involves the hydrogenation of *CO to *CHO, followed by the addition of another proton–electron pair to yield *CH₂O.^160,180,181 The primary bottleneck in the selective synthesis of formaldehyde lies in its exceedingly high intrinsic reactivity. HCHO is notoriously difficult to stabilize as a free product; it typically binds strongly to the catalyst surface and undergoes immediate, deep reduction to CH₃OH or CH₄. Consequently, formaldehyde is largely observed only in trace amounts across most CO₂RR systems, making its effective isolation and accurate quantification a formidable challenge.¹⁸⁵ For this reason, steering the reaction to terminate at formaldehyde remains thermodynamically and kinetically disfavored, explaining the current absence of nanoalloys targeting HCHO. Future breakthroughs to rescue formaldehyde from deep reduction may rely on decoupling intermediate adsorption energies via SAAs to disrupt contiguous active sites, tailoring hydrophobic microenvironments to accelerate HCHO desorption, or employing pulsed electrolysis for precise kinetic control.

4.2. C₂ products

C₂₊ products, encompassing hydrocarbons like ethylene and highly valuable oxygenates such as ethanol and acetic acid, possess significantly higher commercial value and energy density, but their formation pathways are considerably more complex. While the critical rate-limiting step remains the initial C–C coupling, which requires catalysts capable of retaining *CO intermediates to facilitate dimerization without causing surface poisoning, overall selectivity is further governed by the downstream branching of these coupled intermediates. Accordingly, nanoalloy strategies, through electronic effects, ensemble effects, and tandem catalysis, are considered key not only to optimizing the initial C–C coupling but also to dictating the structural evolution of post-coupling intermediates, thereby enabling high selectivity and advancing practical applications.

4.2.1. Ethylene. Among the C₂₊ products of CO₂RR, C₂H₄ is considered the most important owing to its high commercial value and role as an industrial platform chemical. C₂H₄ formation requires a key C–C coupling step, in which *CO acts as the central precursor. However, this pathway directly competes with *CO hydrogenation toward C₁ products such as CH₄. Therefore, maintaining a sufficiently high *CO surface coverage is a prerequisite for enabling C–C coupling, as it increases the probability of two *CO species encountering and reacting. The initial coupling mechanisms are typically proposed as symmetric dimerization (*CO–*CO → *COCO) or asymmetric pathways such as *CO–*COH or *CO–*CHO (yielding *COCOH or *COCHO). Once C₂ intermediates are formed, they undergo successive hydrogenation and dehydration steps, with C–O bond cleavage often steering the pathway toward dehydration and ethylene formation (Fig. 7).^70,72,160 Consequently, the ability to control *CO accumulation and reaction branching at this stage is considered decisive for achieving high C₂H₄ selectivity.

One core strategy for enhancing ethylene production is steering the asymmetric *CO–*CHO coupling pathway, which has a lower energy barrier than *CO–*CO dimerization. Xiao et al. realized this approach by designing a single-atom alloy in which Pr atoms were embedded in a Cu matrix (Pr@Cu).¹⁸⁶ As illustrated in Fig. 13a, the unoptimized Cu catalyst typically drives symmetric *CO–*CO coupling, whereas the introduction of isolated Pr atoms on Cu induces an *asymmetric CO–CHO coupling configuration. With their strong oxygen affinity, Pr atoms played a dual role: enhancing CO₂ activation and promoting H₂O dissociation to supply local protons, thereby enriching *CHO intermediates and favoring *CO–*CHO coupling. The corresponding free-energy diagram (Fig. 13b) reveals that Pr incorporation substantially lowers the activation barrier for *CO–*CHO coupling, from 1.2 eV on pure Cu to 0.07 eV on Pr@Cu, demonstrating the kinetic advantage of asymmetric coupling over conventional *CO–*CO dimerization. As a result, Pr@Cu achieved FE(C₂H₄) of 64.2% at 1200 mA cm⁻² in a flow cell (Fig. 13c), and when integrated into a 100 cm² MEA, the system remained stable for 200 h at 20 A (Fig. 13d), demonstrating industrial-scale potential.

Fig. 13 (a) Schematic comparison between Cu catalyst (symmetric CO–CO coupling) and Pr@Cu single-atom alloy (asymmetric CO–CHO coupling); (b) free-energy profiles showing lower activation barrier for CO–CHO coupling on Pr@Cu; (c) product distribution and current density for Cu and Pr@Cu catalysts with different Pr loadings; (d) long-term stability test in MEA, reproduced from ref. 186 with permission from American Chemical Society, copyright 2025. (e) Tandem mechanism on Cu–Ag Janus heterostructure; (f) DOS plots of Ag, Cu, and Cu–Ag Janus structures, showing d-band modulation with different Ag/Cu ratios, reproduced from ref. 187 with permission from American Chemical Society, copyright 2025. (g) PDOS comparison of Cu and Cu₁₀Sn; (h) free-energy diagram illustrating reduced barriers for CO formation and CO–CO coupling on Cu–Sn; (i) FE of Cu–Sn catalysts with different ratios; (j) schematic of cooperative Cu–Sn interface, reproduced from ref. 188 with permission from Elsevier Inc, copyright 2024.

Heterostructured catalysts, particularly Janus architectures, have shown remarkable effectiveness in maximizing tandem synergy through precise spatial arrangement. Zhang et al. synthesized Janus Cu–Ag structures that exhibited superior CO₂RR performance toward C₂₊ products.¹⁸⁷ In this system, the Ag phase, with weak *CO affinity, acted as the site for CO₂-to-CO conversion, after which *CO “spilled over” to the adjacent Cu phase for C–C coupling (Fig. 13e). Owing to this mechanism, Cu–Ag Janus achieved FE(C₂₊) up to 69.8%, with FE(C₂H₄) of ∼50% and a C₂₊/C₁ ratio 9.2 times higher than pure Cu. In situ ATR-FTIR confirmed the formation of *CHO and *COCHO intermediates, and the electronic structure analysis in (Fig. 13f) further highlights the influence of interfacial coupling on the catalytic properties. The density-of-states distributions show that the d-band center is significantly modulated by the Ag doping. Notably, the optimized CuAg Janus 1 : 0.02 catalyst possesses the highest (most upshifted) d-band center (−3.436 eV), which is significantly higher than that of pure Cu NPs and the other CuAg Janus ratios. This electronic adjustment optimizes the adsorption of key intermediates, thereby favoring C–C bond formation at the heterointerface. Sharing a similar robust design principle, Zheng et al. constructed well-defined Au–Cu Janus nanostructures via a seeded growth strategy.¹⁸⁹ Analogous to the Ag–Cu system, the Au domains act as efficient CO factories, enabling direct CO spillover to the intimately connected Cu counterparts. Both experimental verifications and theoretical simulations revealed that this spatial configuration substantially enhances the local *CO coverage on the Cu surfaces, thereby lowering the energy barrier for C–C coupling and driving deep reduction to yield a remarkable C₂₊ FE of 67%. Together, these studies powerfully highlight that Janus architectures, whether utilizing Ag or Au as the CO, generating phase-offer a highly versatile and universal platform for optimizing multi-step reaction pathways via tandem catalysis.

Beyond noble metals, alloying Cu with p-block metals has emerged as an effective strategy for tuning C₂₊ selectivity. Although Sn is typically associated with C₁ products, Wang et al. demonstrated that incorporating Sn into Cu significantly promoted ethylene formation.¹⁸⁸ DFT calculations revealed that Sn incorporation induces electron withdrawal from Cu, as evidenced by the PDOS analysis in Fig. 13g. The Cu d-band center shifts slightly upward from −2.29 eV to −2.28 eV upon Sn addition, indicating a redistribution of electronic density in which Sn partially extracts electrons from Cu. This electronic adjustment enhances orbital overlap between Cu and the adsorbed CO species, facilitating stronger intermediate stabilization. The resulting increase in stable *CO coverage lowered the C–C coupling barrier (*CO + *CO → *OCCO) from 2.22 eV to only 0.38 eV, while also facilitating C₂H₄ desorption (Fig. 13h). Consequently, the Cu–Sn alloy achieved FE(C₂H₄) of 48.74%, representing a 55% improvement over pure Cu (Fig. 13i). The reaction mechanism, as depicted in Fig. 13j, involves Sn sites acting as electron-withdrawing centers that stabilize CO intermediates, whereas adjacent Cu atoms serve as active sites for CO–CO coupling. This synergistic interaction facilitates C–C bond formation and directs the reaction toward selective C₂H₄ generation.

A compelling demonstration of integrating multiple design strategies is the hollow nanocavity CuPd SAA reported by Zhang et al.¹⁹⁰ This catalyst leveraged both geometric and electronic effects to optimize ethylene production. Geometrically, the nanocavity confinement effect enriched local CO concentration and preferentially stabilized linear *CO (active for C–C coupling) over less active bridge *CO. Electronically, isolated Pd atoms acted as electron sinks, generating Cu^δ+ sites that reduced the C–C coupling barrier, as confirmed by DFT. Together, these effects enabled CuPd SAA to reach FE(C₂H₄) = 75.6% and FE(C₂₊ total = 85.7%) at −0.7 V vs. RHE, with a C₂₊/C₁ ratio 6.5× higher than porous Cu and 12× higher than Cu₂O, while maintaining stability for 20 h of continuous electrolysis.

Advanced surface treatments provide another means of engineering complex active sites. Wang et al. applied H₂ cold plasma treatment to create Cu–Co catalysts with multiple coexisting sites, including Cu⁰, Co⁰, Cu²⁺, and Co²⁺.¹⁹¹ In this tandem mechanism, metallic Co primarily facilitated CO₂-to-CO conversion, after which *CO spilled over to neighboring Cu²⁺ sites for efficient C–C coupling into C₂H₄. Owing to this synergy, the optimized Cu₂Co₁ catalyst achieved FE(C₂H₄) of 70% and remained stable for 200 h.

In addition to experimental advances, theoretical approaches play a vital role in catalyst screening. Wang et al. used DFT to investigate diatomic M/Cu-NG systems (M = Zn, Pd, Sn, Ag, Au) for CO₂-to-ethylene conversion.¹⁹² The screening identified Sn/Cu-NG and Pd/Cu-NG as the most promising, with distinct coupling mechanisms. On Sn/Cu–NG, the reduction proceeds primarily through an asymmetric CO–CHO coupling route, ultimately favoring the COHCHOH intermediate pathway. However, this catalyst faces a high potential-determining step with ΔG = 0.96 eV, which corresponds to the hydrogenation of *CO to *CHO (Fig. 14a). In contrast, Pd/Cu-NG follows a symmetric CO–CO coupling mechanism, leading to the formation of *COCO intermediates. The potential-determining step is this C–C coupling, which has a significantly lower energy barrier of ΔG = 0.71 eV (Fig. 14b). These results suggested Pd/Cu-NG as the more efficient ethylene catalyst, with a lower limiting potential. The electron localization function (ELF) maps in Fig. 14c clearly visualize the charge distribution around the active dual-metal sites. The analysis indicates that Pd/Cu-NG exhibits stronger localized electronic states than Sn/Cu-NG. This strong localization facilitates significant d-orbital hybridization between Pd and Cu, which is beneficial for the C–C coupling process. Conversely, the introduction of Sn makes the electron cloud more delocalized, resulting in weaker orbital hybridization and contributing to its higher C–C coupling barrier.

Fig. 14 (a) Free-energy diagrams for CO₂ reduction on Sn/Cu-NG and (b) on Pd/Cu-NG; (c) electron localization function maps of Pd/Cu-NG and Sn/Cu-NG, reproduced from ref. 192 with permission from American Chemical Society, copyright 2025. (d) Reaction free-energy profile of PtNi@Cu; (e) in situ FTIR spectra of PtNi@Cu; (f) PDOS plots showing d-band modulation among ternary systems; and (g) charge-density difference maps visualizing electron redistribution and cooperative interaction between Pt, Ni, and Cu atoms, reproduced from ref. 193 with permission from Royal Society of Chemistry, copyright 2025.

One major challenge in alloy studies is disentangling compositional effects from morphological influences. To address this, Crandall et al. combined systematic theoretical screening with standardized experimental protocols to explore dilute Cu alloys.¹⁹⁴ DFT was first used to evaluate dopants (Al, B, Ga, Sc, Au, Pd) that could lower the *CO dimerization barrier-the key step in C₂₊ formation. Subsequently, dilute alloy films (∼5 at%) were fabricated via physical vapor deposition (PVD), ensuring nearly identical morphologies to isolate composition effects. Experiments validated predictions: CuAl, CuB, CuGa, and especially CuSc alloys all showed significantly higher C₂₊/C₁ ratios than pure Cu, with CuSc delivering the best performance (C₂₊/C₁ doubled). This increase stemmed from enhanced production of ethylene, ethanol, and propanol rather than dominance of a single product. Notably, scaling to 5 cm² and 100 cm² MEAs confirmed stable performance, with FE(C₂H₄) >40% at 200 mA cm⁻² and only slight decline after 4 h.

Extending beyond binary alloys, ternary alloys offer further opportunities to fine-tune active sites but face challenges due to vast compositional space. To address this, Xiao et al. pioneered a rational design methodology integrating DFT-based screening with three criteria: stability (surface formation energy), activity (CO₂ vs. H₂O adsorption), and selectivity (favoring *COOH over *H).¹⁹³ This identified PtNi@Cu as a promising ethylene catalyst. Computational screening revealed that among the investigated systems, PtNi@Cu exhibits the lowest energy barrier for the COOH* formation step. Moreover, the subsequent reaction pathway shown in Fig. 14d confirms that PtNi@Cu follows the CO*–CHO* and CO*–COH* coupling pathways rather than the conventional CO*–CO* route. Experimental validation confirmed these theoretical predictions, demonstratingthat PtNi@Cu achieves FE(C₂H₄) of 30.9% at −1.2 V vs. RHE. The in situ FTIR spectra (Fig. 14e) further verify the presence of COH* and COCOH* intermediates during electrolysis, supporting the proposed CO*–COH* coupling mechanism. Electronic structure analyses in Fig. 14f shows that PtNi@Cu possesses the highest d-band center compared to the other ternary alloys shown. According to the d-band theory, this higher d-band center allows PtNi@Cu to activate the *CO intermediate more efficiently, thereby reducing the Gibbs free energy of CO₂RR. Finally, the charge-density difference maps in Fig. 14g visualize pronounced charge delocalization between Pt, Ni, and Cu atoms, confirming the cooperative interaction among the three metals. The synergy among Pt, Ni, and Cu enhanced electronic redistribution, stabilized *COH intermediates, and promoted C–C coupling. The significance of this work lies not only in the ∼31% FE achieved but also in the methodological framework it provides, offering a rational “map” for navigating the vast compositional space of ternary alloys-a task nearly impossible through experiments alone.

4.2.2. Ethanol. The production of C₂H₅OH, a deeply reduced product requiring 12 electrons, is regarded as one of the most challenging targets in CO₂RR. Although its pathway originates from the *CO intermediate and involves a C–C coupling step similar to ethylene, the selectivity between the two diverges at subsequent stages. Specifically, after the formation of oxygen-containing C₂ intermediates (e.g., *CH₂CHO), the catalyst may follow two competing routes: (i) deoxygenation via C–O bond cleavage, leading to C₂H₄, or (ii) hydrogenation, preserving the C–O bond to yield ethanol.^70,72,160 An inherent challenge is that the strong stabilization of C₂ intermediates, while necessary for coupling, simultaneously lowers the barrier for deoxygenation. Thus, an effective ethanol-selective catalyst must not only promote C–C coupling but also finely tune proton transfer kinetics in later steps to favor hydrogenation while suppressing deoxygenation.¹⁹⁵

Modern design strategies often focus on creating oxophilic active sites or tailoring the local reaction environment to achieve this.¹⁹⁷ A representative study by Wang et al. demonstrated how atomic-level catalyst engineering can direct C–C coupling pathways in CO₂RR.¹⁹⁶ Using galvanic replacement, they synthesized two model systems: AgCu NW (Ag nanoparticles on Cu nanowires) and Ag₁Cu NW (isolated Ag atoms dispersed on Cu) (Fig. 15a). Structural characterization by TEM and HAADF-STEM confirmed their distinct configurations. The two catalysts exhibited markedly different product selectivity: AgCu NW favored C₂H₄ with a maximum FE of 54.9% at 156.0 mA cm⁻², while Ag₁Cu NW displayed superior selectivity for ethanol, achieving 56.3% FE at 172.8 mA cm⁻². At the core of this study is the elucidation of the mechanism governing product branching, achieved by integrating advanced operando spectroscopic analyses with theoretical calculations. In particular, operando Raman and ATR-SEIRAS measurements (Fig. 15b–e) offer direct evidence of the distinct C–C coupling pathways occurring on two catalyst systems. For AgCu NW, strong *CO-related vibrations at 282, 362, and 1950–2150 cm⁻¹ indicate a high *CO coverage that favors *CO–*CO coupling toward ethylene (Fig. 15b). Complementary ATR-SEIRAS spectra (Fig. 15d) reveal a dominant *COCO band at 1561 cm⁻¹, together with direct detection of the *COCO intermediate, clear evidence that C₂H₄ forms via the symmetric *CO–*CO coupling route. In contrast, Ag₁Cu NW exhibits distinct operando Raman signals of *OH species (∼537 cm⁻¹), implying that isolated Ag sites promote H₂O dissociation and generate a proton-rich microenvironment that facilitates *CO hydrogenation to *CHO (Fig. 15c). Corresponding ATR-SEIRAS spectra display *CHO (1244 cm⁻¹) and *COCHO (1083 cm⁻¹) bands, confirming the *CO–*CHO coupling pathway responsible for ethanol formation (Fig. 15e). The potential-dependent evolution of these bands further supports their mechanistic assignment. DFT calculations (Fig. 15f) reveal that the hydrogen-assisted *CO → *CHO step on HO-Ag₁Cu sites possesses the lowest energy barrier (0.83 eV). Consistent with these energetic profiles, the corresponding schematic (Fig. 15g) visualizes the asymmetric *CO–*CHO coupling sequence, where *CO and *CHO intermediates couple asymmetrically at paired Cu atoms adjacent to single Ag sites, producing *COCHO and subsequently ethanol. Collectively, these operando and theoretical findings demonstrate how atomic-level catalyst engineering-specifically the introduction of isolated Ag sites-modulates the local reaction environment, lowers the *CO hydrogenation barrier, and steers the C–C coupling pathway from symmetric (*CO–*CO → C₂H₄) to asymmetric (*CO–*CHO → CH₃CH₂OH) routes.

Fig. 15 (a) Schematic of catalyst synthesis via galvanic replacement; (b) operando Raman spectra for AgCu NW and (c) Ag₁Cu NW; (d) operando ATR-SEIRAS spectra identifying the *COCO intermediate on AgCu NW and (e) the *CHO and *COCHO intermediates on Ag₁Cu NW; (f) DFT free energy diagram for the *CO to *CHO hydrogenation step, showing a reduced barrier on the HO-Ag₁Cu NW site; and (g) schematics of the proposed asymmetric *CO-*CHO coupling pathway to ethanol, reproduced from ref. 196 under CC-BY-NC-ND.

Another breakthrough strategy was introduced by Kim et al., addressing the thermodynamic immiscibility barrier in alloying.¹⁹⁸ Instead of conventional alloying prone to phase segregation, they applied a metallurgical approach by incorporating sacrificial Al into Cu–Ag–Al ternary alloys to form metastable intermetallics, followed by selective leaching of Al. This yielded a supersaturated Ag–Cu solid solution, with Cu clusters uniformly dispersed in ordered Ag lattices-a structure unattainable by standard methods. Operando Raman spectroscopy and DFT calculations revealed that this solid solution generated a high density of step/edge defects, altering *CO adsorption from terrace-preferred LFB atop sites (favoring C₂H₄) to defect-preferred HFB atop sites, increasing *CO binding strength. Consequently, production shifted from ethylene on pure Cu (69.6 ± 1.3% FE) toward ethanol (40.4 ± 2.4% FE) on the supersaturated alloy. The combined effects of Cu–Ag interfacial synergy and defect stabilization lowered the dimerization barrier and stabilized intermediates along the ethanol pathway.

One of the remarkable recent advances in ethanol synthesis at industrially relevant current densities comes from combining intermetallic compounds with facet engineering strategies. In a notable study, Peng et al. successfully synthesized intermetallic Cu₂Mg with preferential exposure of the (111) facet.¹⁹⁹ Ordered Cu₃–Mg sites on this facet were identified as key to ethanol selectivity. Electron transfer from Mg to Cu created electron-rich Cu centers capable of strongly stabilizing the crucial *CHCHOH intermediate, the direct precursor to ethanol, while the competing *CCH intermediate (toward C₂H₄) was energetically unfavorable. As a result, Cu₂Mg(111) achieved a remarkable 76.2% FE for ethanol at 600 mA cm⁻², with a partial current density of 720 mA cm⁻²-among the highest reported.

Beyond metal–metal interactions, advanced catalyst designs integrate additional factors. Guo et al.²⁰⁰ proposed a dual-modification strategy on Cu–Zn alloys: (1) alloying with Zn to tune Cu's electronic structure, and (2) introducing the organic stabilizer 3,5-diamino-1,2,4-triazole (DAT). Zn incorporation enhanced FE and partial current density for ethanol by 325% and 600%, respectively, compared to pure Cu. Further addition of DAT boosted them by 194% and 292%, yielding CuZn-DAT with ethanol selectivity up to 87% among C₂ products. Mechanistic studies showed that Zn stabilized *CO and lowered the barrier for C–C coupling while suppressing HER, whereas DAT coordinated with CuZn sites to modulate adsorption of *CO and *CHO, synergistically steering the reaction toward ethanol.

The transport of the *CO intermediate has also been revealed as a key factor. Wang et al.²⁰¹ conceptually distinguished two migration modes in tandem catalysts (Fig. 16a). In conventional “out-migration,” CO desorbs from Ag sites into the electrolyte and then re-adsorbs onto Cu, resulting in poor utilization efficiency. In contrast, in “inter-migration,” CO transfers directly between adjacent Ag and Cu atoms without leaving the surface. This eliminates interfacial barriers and maintains high CO surface coverage, thereby facilitating efficient C–C coupling. To favor the latter, they designed an AgCu-SAA, where Ag atoms were atomically dispersed in Cu (Fig. 16b). This eliminated physical phase boundaries, enabling direct *CO transfer. HR-TEM image, AC-HAADF-STEM image, and EDS element mapping images (Fig. 16c–e) confirm the atomic dispersion of Ag on Cu, with no detectable Ag–Ag coordination, while XANES/EXAFS analyses (Fig. 16f–h) reveal strong Ag–Cu interactions and the absence of Ag clusters, supporting the single-atom alloy configuration. The catalytic implications are shown in Fig. 16i–k. AgCu-SAA exhibits the highest CO conversion rate and C₂₊ FE among Cu and AgCu-Nano catalysts, achieving 83.4% FE(C₂₊) at 900 mA cm⁻². This superior performance arises from the enhanced CO retention and migration efficiency at atomic Ag–Cu interfaces. Operando Raman spectra (Fig. 16l–n) further confirm distinct adsorption modes. Compared with Cu and AgCu-Nano, AgCu-SAA shows stronger bands for bridge and low-frequency atop CO adsorption, indicating greater surface CO coverage and stronger CO–Cu interaction. Calculations reveal that the CO adsorption energy on AgCu-SAA lies between that of Cu and AgCu-Nano, achieving an optimal balance between retention and mobility (Fig. 16o and p). Energy-barrier profiles demonstrate that for AgCu-Nano, CO desorption into the electrolyte (out-migration) is thermodynamically preferred, whereas for AgCu-SAA, the barrier for CO transfer from Ag to adjacent Cu sites is comparably low, validating the inter-migration mechanism. The corresponding energy profiles (Fig. 16q and r) visualize the facile CO transport along the metal–metal interface, emphasizing that atomic Ag sites act as “relay stations” for CO delivery toward C–C coupling centers. Collectively, these combined operando and computational results reveal that atomic-level Ag–Cu coordination enables direct CO inter-migration, boosting local CO concentration and stabilizing key intermediates for C–C bond formation.

Fig. 16 (a) Schematic comparison of out-migration vs. inter-migration modes of CO transfer between Ag and Cu; (b) synthesis route of AgCu-SAA; (c–e) HR-TEM, AC-HAADF-STEM, and EDS mappings confirming atomic dispersion of Ag; (f–h) Ag K-edge XANES spectra, Ag K-edge FT-EXAFS spectra, and Wavelet transform image of the AgCu-SAA; (i–k) CO conversion rate, C₂₊ FE, and LSV curves over Cu, AgCu-SAA, and AgCu-nano; (l–n) in situ Raman spectra revealing strong bridge and low-frequency atop CO adsorption on AgCu-SAA; (o and p) DFT-computed free-energy diagrams and CO adsorption energies over Cu, AgCu-Nano, and AgCu-SAA; (q and r) calculated energy barriers and pathways for CO transfer from Ag to Cu, reproduced from ref. 201 with permission from American Chemical Society, copyright 2025.

While tandem catalysis remains pivotal, emerging evidence suggests that classical tandem models cannot fully explain selectivity differences between ethylene and ethanol on alloys of similar composition. A deeper mechanism involves dynamic restructuring of Cu under the influence of secondary metals.²⁰² For example, Au not only provides CO but also acts as a structural modulator, driving Cu into amorphous surface layers enriched in undercoordinated sites during reaction. These dynamic sites, rather than static tandem effects, govern the branching toward more complex products. This paradigm shift-from phase-separated tandem models to dynamic atomic-scale restructuring-offers new directions for tailoring product selectivity. Direct evidence comes from Rollier et al.,²⁰³ who investigated Cu–Ag catalysts for CO reduction. They identified Cu_0.9Ag_0.1 as optimal, with C₂₊ selectivity up to 63%, outperforming pure Cu. Operando and in situ characterization showed that the oxide precursors reconstructed into Ag-core Cu-shell nanoparticles, with WAXS confirming a Cu–Ag alloy phase at the interface. These interfacial sites, featuring strong electronic effects, were pinpointed as the active centers driving enhanced C–C coupling and suppressing HER.

4.2.3. Acetic acid/acetate. Acetic acid (or acetate) is a high-value C₂ oxygenate, serving as a vital precursor for various polymers and industrial solvents. Its formation is complex and faces intense competition from ethylene and ethanol production, as these products share common initial C₂ C–C coupling intermediates. While acetate can form through specific pathways like the nucleophilic attack of *CH₃ species on *CO₂˙⁻,²⁰⁴ it is frequently proposed to originate directly from the ethylene/ethanol pathways. This occurs either via the isomerization of the *OCH₂COH intermediate into a three-membered ring compound followed by further reduction,²⁰⁵ or through the dehydration of (OH)CCOH to form a ketene intermediate (*C=C=O) that converts to acetate at a high local pH.²⁰⁶ Alternatively, acetate can also be formed via a Cannizzaro-type disproportionation of acetaldehyde under highly alkaline conditions.^204,205 Because acetate synthesis fiercely competes with ethylene and ethanol, achieving high FE on monometallic catalysts remains elusive. Breaking these inherent selectivity trade-offs requires maintaining a strongly alkaline local environment and high *CO surface coverage. To navigate these complex requirements, researchers have effectively employed spatial decoupling strategies such as tandem catalysis.

A prominent example is the distinct architecture developed by Hu et al., comprising copper nanoparticles encapsulated by single-nickel-atom-modified carbon frameworks (Ni SACs-Cu NPs).²⁰⁷ In this system, atomically dispersed Ni sites efficiently reduce CO₂ to *CO, generating a localized, high-concentration *CO environment. These intermediates readily migrate to the adjacent Cu(111) facets, thermodynamically circumventing the formate pathway and favoring specific C–C coupling toward acetate. This synergistic tandem strategy achieved an impressive acetate selectivity of over 60% among liquid products at high current densities (50–200 mA cm⁻²). Beyond macro-spatial decoupling, precise atomic engineering within sub-nanometer bimetallic systems has proven highly effective in driving the crucial C–C coupling steps. For instance, an entropy-derived CuPd sub-1nm alloy engineered on CuO/phosphomolybdic acid subnanosheets leverages subnanometer confinement to create fully exposed Cu–Pd pairs.²⁰⁸ These unique bimetallic sites synergistically enhance *CO surface coverage. Furthermore, in situ Raman spectroscopy and ab initio molecular dynamics revealed that high vibrational entropy induces energetic oscillations within the Cu–Pd pairs, dynamically facilitating the close proximity and subsequent asymmetric C–C coupling of adsorbed *CO and *COH to form the crucial COH–C=O species. This unique subnanoalloy achieved a peak FE of 46.5% for acetate and maintained robust stability over 20 hours.

The importance of tailoring the local electronic environment to lower the energy barriers for such asymmetric coupling is a universal principle that extends to valence-engineered composite systems. Zhang et al. demonstrated that precisely tuning the Cu(II)/Cu(I) ratio optimizes the synergistic interplay between mixed-valence copper sites.²⁰⁹ This electronic modulation effectively steers the reaction toward asymmetric *CO–COH coupling while suppressing HER, yielding an acetate FE of 35.78%, a mechanistic advantage that strongly parallels the synergistic effects harnessed in bimetallic nanoalloys. Furthermore, isolating secondary metals at the atomic level offers a powerful avenue to steer selectivity. Sun et al. demonstrated the potential of precise bimetallic interface engineering with an atomically dispersed Cu–Au alloy (CuAu_1%).²¹⁰ The incorporation of isolated Au atoms onto a Cu nanoparticle host creates abundant atomic Cu–Au interfaces that fundamentally alter the geometric and electronic structure. DFT calculations revealed that this atomic dispersion effectively weakens the binding strength of *CO+*CO intermediates, thereby facilitating direct C–C coupling and favorably routing the pathway toward acetate. In an alkaline flow cell, this CuAu_1% nanoalloy achieved a maximum acetate FE of 39% and an outstanding partial current density of 217 mA cm⁻², underscoring the efficacy of single-atom alloying for targeted C₂₊ oxygenate synthesis.

In summary, the successful conversion of CO₂ into specific C₂ products highlights the transformative power of nanoalloy engineering in breaking inherent monometallic scaling limitations. By synergistically integrating tandem catalysis, precise electronic modulation, and atomic-level confinement, these advanced multimetallic architectures enable unprecedented control over both the critical C–C coupling step and the subsequent branching pathways toward targeted hydrocarbons and oxygenates.

4.3. Toward C₃₊ products

The electrochemical production of C₃₊ compounds such as n-propanol is an ambitious goal due to their higher energy density and economic value, and remains the most formidable challenge in CO₂RR, requiring multiple complex coupling steps. Although the precise mechanism is still under debates, recent studies have begun to provide deeper insights.²¹¹ While symmetric *CO–*CO dimerization has long been considered the canonical pathway for C–C bond formation, Zheng et al.²¹² reported, through DFT simulations on Pd₃Au alloys, that chain propagation is instead dominated by sequential asymmetric couplings. In this route, C₂ intermediates (*CCH₂, *CCH₃) couple further with C₁ fragments (*CH_x) to yield C₃ products. In particular, the *CCH₂ + *CH₃ step exhibits a relatively low barrier, favoring the formation of propylene and propane. Stable C₃ intermediates can even engage in further coupling with C₁ units to extend into C₄ hydrocarbons. By contrast, oxygenated intermediates such as CO or COH present significantly higher barriers, explaining why Pd₃Au tends to produce hydrocarbons rather than oxygenates.

A comprehensive study by Bae et al.²¹³ on Cu–Zn alloy electrodes demonstrated the formation of a broad spectrum of products, including propanol, isopropanol, and hydrocarbons up to C₄. A key finding was the resemblance of CO₂RR to the classical Fischer–Tropsch (F–T) synthesis. Using Anderson–Schulz–Flory distribution analysis, the authors found that the production of C₂₊ alkanes and C₃₊ alkenes followed a linear trend, indicating stepwise chain-growth akin to *CO or *CH_x insertion in F–T synthesis. As illustrated in Fig. 17a, the process begins with CO₂ adsorption and reduction to *CO intermediates on Cu–Zn alloy surfaces, followed by sequential proton–electron transfers that generate *CH_x species. The coupling between *CO and *CH_x intermediates enables C–C bond formation, leading to the formation of C₂ products. Subsequent chain propagation occurs through repetitive *CH₂ or *CO insertion steps, yielding C₃₊ hydrocarbons such as propanol, isopropanol, and butenes. Interestingly, ethylene deviated from this linearity, suggesting it is produced through a parallel pathway, most likely the well-established *CO–*CO dimerization mechanism. This provides a new perspective in which catalyst design must not only enable initial C–C coupling but also promote Fischer–Tropsch-like chain propagation.

Fig. 17 Strategies for CO₂RR toward C₃₊ products. (a) Fischer–Tropsch-like chain growth on Cu–Zn alloys, reproduced from ref. 213 with permission from American Chemical Society, copyright 2024. (b and c) Product selectivity and Ru composition effects on Pt_1−xRu_x/C; (d) reaction schemes on Pt/C and Pt_0.9Ru_0.1/C, reproduced from ref. 214 under the Creative Commons CC-BY 4.0 license. (e) Tandem CO₂ → C₂H₄ → C₄H₈ cascade integrating electrochemical and thermochemical steps, reproduced from ref. 215 with permission from American Chemical Society, copyright 2024.

A landmark work by Matsuda et al.²¹⁴ demonstrated the production of not only CH₄ and ethanol but also acetone on Pt_0.9Ru_0.1 catalysts. Beyond overcoming Pt's conventional limitation of producing mainly H₂ and suffering from CO poisoning, Pt_0.9Ru_0.1/C in MEA delivered simultaneous C₁–C₃ products with a total FE of 31.7% at an ultralow overpotential of just 0.18 V (Fig. 17b). This was the first report of ethanol and acetone generation on Pt–Ru systems. Fig. 17c further compares product distributions as a function of Ru content, showing that the composition Pt_0.9Ru_0.1 provides the optimal balance between CO adsorption and hydrogenation activity, leading to the highest overall multicarbon yield. Fig. 17d schematically depicts the mechanistic difference between Pt/C and Pt_0.9Ru_0.1/C surfaces under varying CO₂ concentrations. On pure Pt, strong CO binding blocks active sites and limits reduction to methane under low CO₂ availability, while under high CO₂ concentrations the surface becomes poisoned. In contrast, the incorporation of Ru weakens Pt–CO interaction, allowing CO intermediates to accumulate and migrate, facilitating Langmuir–Hinshelwood-type coupling between adjacent CO species. This coupling leads to the formation of C–C bonds and subsequent hydrogenation into ethanol and acetone, marking a distinct pathway beyond methane formation.

A promising approach is tandem cascade systems integrating electrochemical and thermochemical steps. Lee et al.²¹⁵ achieved this by producing 1-butene from CO₂. Their setup consisted of two sequential reactors: (i) a high-performance electrolyzer reducing CO₂ to C₂H₄, and (ii) a thermochemical reactor for C₂H₄ dimerization (Fig. 17e). The C₂H₄ stream from the first reactor was fed directly, after dehydration, into the second reactor, where a Ru–Ni metal–organic framework catalyzed its conversion to 1-butene with 97% selectivity at 25 °C and 1 atm. Overall, the integrated CO₂ → C₂H₄ → C₄H₈ cascade achieved ∼47% selectivity at 156 mA cm⁻². Although the second step exhibited near-perfect efficiency, overall performance depended on the cumulative yield of both reactors and transfer losses, underscoring both the promise and the challenge of cascade strategies for accessing C₃₊ products.

Recent advances toward C₃₊ products highlight three complementary strategies: (i) Fischer–Tropsch-like chain growth enabling stepwise *CO/*CH_x insertion, (ii) surface electronic tuning in nanoalloy systems that promotes CO accumulation and C–C coupling, and (iii) tandem cascade systems combining electrochemical and thermochemical stages for selective C₄ formation. Despite these advances, the direct electrochemical synthesis of higher-order products remains fundamentally constrained by competing pathways, limited intermediate coverage, and the delicate balance between hydrogenation and coupling kinetics. Future progress will depend on multi-site catalyst design that enables both localized intermediate confinement and sequential C–C insertion, possibly through hybrid systems coupling electrocatalysis with chemical upgrading. Achieving such integrated control will be crucial for transforming CO₂RR from C₂-level selectivity to scalable C₃₊ fuel and chemical production.

Finally, to quantitatively benchmark recent progress, Table 5 compares the key electrochemical metrics (FE, current density, applied potential, and durability) of representative nanoalloy systems. This side-by-side comparison reveals that while current designs excel in product selectivity, achieving concurrent high current densities and long-term stability remains the primary challenge for practical CO₂RR deployment.

Table 5 Comparison of CO₂ electroreduction performance across various alloy catalyst systems

No.	Alloy catalyst system	Main product	FE	Current density	Applied potential	Stability	Ref.
1	Cu3Sb (intermetallic alloy)	CO	97.9%	41.9 mA cm^−2	−0.6 V vs. RHE	>12 hours	161
2	Cu1Fe2/NC (Cu–Fe alloy on N-carbon)	CO	98.91%		−0.7 V vs. RHE	24 hours	163
3	NP-Ag5Cu5Au5 (nanoporous alloy)	CO	∼96%	147 mA cm^−2 (partial j)	−1.173 V vs. RHE	8 hours	162
4	AgCu nanoalloy	CO	83.2%	7.52 mA cm^−2 (partial j)	−0.5 V vs. RHE	>3 hours	90
5	Cu0.07Zn (nanosheets alloy)	Syngas (CO/H2)	∼100%	20 mA cm^−2	−0.95 V vs. RHE	42.2 hours	165
6	Cu–Zn-675 (bimetallic Cu–Zn)	Syngas (CO/H2)	∼90%	5.98 mA cm^−2	−0.81 V vs. RHE	7.5 hours	164
7	InSn@SnY (In0.2Sn0.8 alloy on zeolite)	HCOO^−/HCOOH	98.2%	322 mA cm^−2 (partial j)	−1.33 V vs. RHE	102 hours	171
8	Cu6Sn5@SnOx core–shell	HCOO^−/HCOOH	88.4%	707.2 mA cm^−2 (partial j)	−1.53 V vs. RHE	>90 hours	169
9	Op-Ag1In (SAA)	HCOO^−/HCOOH	93.54%	70 mA cm^−2	−0.9 V vs. RHE	24 hours	174
10	Pd1Cu1 (nanodendrites alloy)	HCOO^−/HCOOH	>90%	∼31.7 mA cm^−2	−0.2 V vs. RHE	15 hours	173
11	Ga0.75In0.25 LMP (liquid metal alloy)	HCOO^−/HCOOH	77.1%	150 mA cm^−2	−1.05 V vs. RHE	>12 hours	172
12	Cu41Sn11	HCOO^−/HCOOH	74%	∼20 mA cm^−2	−1.3 V vs. RHE	>10 hours	92
13	Cu6Sn5/CNFs (Sn-rich phase)	HCOO^−/HCOOH	58.6%		−2.31 V vs. RHE		176
14	Ni1Cu1-NCNT (alloy on N-carbon)	CH4	99.7%	11.54 mA cm^−2 (partial j)	−1.2 V vs. RHE	High stability at −1.2 V	22
15	Co1Cu (Co–Cu SAA)	CH4	62.3%	482.7 mA cm^−2 (partial j)	∼ −0.85 V vs. RHE	10 hours	178
16	Cu3Pd (bimetallic alloy)	CH4	43.23%	269.68 mA cm^−2 (partial j)	−1.8 V vs. RHE		177
17	Cu41Sn11/CNFs (Cu–Sn phase)	CH4	39.1%		−1.36 V vs. RHE		176
18	Cu3Sn/CNFs (Cu-rich phase)	CH4	34.7%		−1.50 V vs. RHE		176
19	PtxZn/C (intermetallic alloy)	CH3OH	81.4%		−0.90 V vs. RHE	16 hours	183
20	Sn2/In2S3 (3% Sn)	CH3OH	31%	7.97 mA cm^−2 (partial j)	−1.331 V vs. RHE	50 hours	184
21	Pr@Cu-2 (Pr–Cu SAA)	C2H4	64.2%	1200 mA cm^−2	−1.6 V vs. RHE	>200 hours	186
22	CuPd SAA (nanocavity enriched)	C2H4	75.6% (C2+ 85.7%)		−0.7 V vs. RHE	20 hours	190
23	Cu2Co1	C2H4	70%	10.2 to 11.7 mA cm^−2	−1.0 V vs. RHE	200 hours	191
24	AgCu NW (bimetallic nanowires)	C2H4	54.9%	156.0 mA cm^−2	−1.1 V vs. RHE	5 hours	196
25	CuAg Janus 1 : 0.02	C2H4	50% (C2+ ∼70%)	67.6 mA cm^−2	−1.2 V vs. RHE	∼4 hours	187
26	Cu10–Sn (Cu–Sn nanosheets)	C2H4	48.74%	11.99 mA cm^−2 (partial j)	−1.29 V vs. RHE	>8 hours	188
27	Au–Cu Janus nanostructures	C2H4 & C2H5OH	67%	290 mA cm^−2 (partial j)	−0.75 V vs. RHE	12 hours	189
28	PtNi@Cu (ternary alloy)	C2H4	30.9%	24.2 mA cm^−2	−1.2 V vs. RHE	14 hours	193
29	Cu2Mg(111)	C2H5OH	76.2 ± 4.8%	600 mA cm^−2	−0.6 V vs. RHE	15 hours	199
30	Ag1Cu NW (Ag-SAA on Cu nanowire)	C2H5OH	56.3%	172.8 mA cm^−2	−1.0 V vs. RHE	5 hours at ∼250 mA cm^−2	196
31	CuZn-DAT	C2H5OH	19.9%	13.45 mA cm^−2 (partial j)	−1.15 V vs. RHE	8 hours	200
32	CuPd-1.5-alloy (Sub-1nm alloy)	CH3COO^−	46.5 ± 2.1%	∼25–36 mA cm^−2	−0.7 V vs. RHE	∼20 hours	208
33	CuAu1% (atomically dispersed alloy)	CH3COO^−	39%	217 mA cm^−2 (partial j)		∼140 min at 200 mA cm^−2	210

---

5. Conclusions

In this review, we have presented a comprehensive product-centered classification for CO₂ electroreduction, demonstrating how specific nanoalloy architectures can be tailored to overcome the intrinsic limitations of monometallic catalysts. Our analysis highlights that catalytic performance is governed by a synergistic interplay between thermodynamic optimization and geometric control. On one hand, electronic and strain effects act as thermodynamic levers, tuning the d-band center of the active metal to optimize the binding strength of key intermediates, thereby enhancing intrinsic turnover frequency and suppressing the competing hydrogen evolution reaction. On the other hand, ensemble and tandem effects serve as geometric levers, where the spatial arrangement of atoms dictates the reaction pathway. Isolated sites spatially restrict C–C coupling to favor C₁ products, whereas contiguous multi-atom ensembles or proximal tandem interfaces facilitate the dimerization of *CO intermediates, effectively steering the pathway toward C₂₊ hydrocarbons and oxygenates.

As CO₂RR research progresses toward practical applications, it is crucial to critically evaluate the inherent strengths and outstanding unresolved issues of current nanoalloy design strategies (summarized in Table 6). While single-atom/dilute alloying and strain engineering are highly effective in breaking LSRs and modulating d-band centers, they intrinsically suffer from thermodynamic metastability. Specifically, the aggregation of single atoms and the relaxation of lattice strain under highly negative cathodic potentials remain significant barriers to industrial lifespans. Conversely, intermetallic phase control offers a robust solution to these stability issues by preventing atomic reconstruction and phase segregation via high mixing enthalpies, albeit being limited to specific thermodynamically favorable metal pairs. Meanwhile, tandem/Janus and nanocavity structuring excellent geometrical levers to enhance local *CO coverage and C–C coupling; however, translating their complex syntheses into scalable mass-production remains a formidable technical challenge. Future catalyst development must aim to bridge these gaps by combining the atomic precision of these strategies with scalable and robust fabrication techniques.

Table 6 Systematic evaluation of nanoalloy design strategies for CO₂RR

Nanoalloy design strategy	Advantages	Limitations	Applicable scenarios	Examples
Single-atom/dilute alloying	- Maximizes atom utilization	- Single-atom sites and dilute components are highly susceptible to reconstruction and aggregation under dynamic, highly negative cathodic potentials	- Tuning specific reaction pathways for target products	- Pr single-atom on Cu^186
	- Effectively breaks inherent LSRs of intermediates	- Precise synthesis control at bulk scale remains challenging	- Suppressing the competing HER	- Single Co atom on Cu^178
	- For example, Pr single-atom on Cu facilitates unique asymmetric CO–CHO coupling,^186 and dilute alloys can modulate CO binding without blocking active sites.^194			- Cu-based dilute alloys^194
Intermetallic and phase control	- Highly ordered atomic arrangements provide uniform active ensembles, separating geometric effects from electronic ones	- Restricted to specific metal pairs that can thermodynamically form stable intermetallic phases	- Continuous, long-term stable electrolysis operations in flow-cells or MEAs where catalyst degradation is a severe issue	- (111) facet-oriented Cu2Mg (ref. 199)
	- High mixing enthalpy suppresses the phase segregation and catalyst restructuring during reaction, leading to exceptional long-term stability	- Synthesis often requires strict processing windows (e.g., precise temperature or oxygen partial pressure control)		- Ordered Pd3Bi vs. solid solution^149
				- Intermetallic CuAu^156
Tandem/Janus structuring	- Couples distinct active sites in close proximity to facilitate intermediate spillover	- Synthesis requires complex, multi-step procedures (such as seeded growth strategies)	- Cascade CO2 electroreduction aiming for complex multi-carbon	- Au–Cu Janus nanostructures^189
	- This increases local *CO coverage and significantly promotes C–C coupling toward C2+ products	- Precisely controlling the interface boundary and domain size at a mass–production scale is extremely difficult		- Cu–Ag tandem catalysis^93,203
Strain engineering	- Introduces lattice mismatch to modulate the d-band center of the active metal	- Lattice strain is thermodynamically metastable	- Boosting intrinsic activity and fine-tuning the binding strength of specific intermediates	- Sn-doped strained CuAg film^128
	- This precisely optimizes the adsorption energy of key intermediates without altering the bulk composition	- Strain relaxation during long-term electrochemical cycling inevitably leads to irreversible performance degradation and structural collapse		- Strain-engineered AuPd^129
				- Tensile-strained Cu-Bi^127
Nanocavity and hierarchical porous structuring	- Generates a strong confinement effect that enriches the local concentration of intermediates (increasing *CO coverage), effectively promoting C–C coupling	- Highly porous or dendritic morphologies are prone to collapse or severe surface reconstruction under strong cathodic potentials over time	- Designing high-surface-area electrodes for deep reduction to C2+ products in aqueous or flow-cell systems	- Nanocavity enriched CuPd^190
	- Enhances mass transport of reactants/products	- Synthesis may involve complex dealloying or templating steps		- Channel-rich Pd–Cu nanodendrites^173

---

6. Future perspectives

Looking forward, bridging the gap between laboratory-scale discovery and industrial deployment requires a paradigm shift focused on deepening the fundamental understanding of dynamic interfaces. While static models are well-established, the behavior of the alloy–electrolyte interface under reaction conditions remains a critical missing mechanism. Since nanoalloys are particularly prone to potential-driven atomic segregation, elemental leaching, and structural reconstruction under harsh cathodic conditions, future research must decouple the dynamic restructuring of metastable active phases from the bulk structure. To unravel these complex mechanisms during CO₂RR, the extensive deployment of advanced in situ and operando characterization techniques is indispensable. A robust analytical framework for this can be drawn from a recent systematic review of in situ methodologies used to investigate interfacial failure mechanisms in zinc anodes.²¹⁶ By analogously applying these in situ/operando spectroscopies and microscopies, researchers can track transient chemical valences, intermediate adsorption, and morphological evolution in real-time. Ultimately, this shift from post-mortem static observations to dynamic mechanistic tracking will be the key to stabilizing active metastable phases and designing “self-healing” alloys.

Beyond advanced experimental characterization, the rational design of complex nanoalloys increasingly relies on theoretical computations. While traditional DFT provides fundamental insights, exploring the vast compositional space of multimetallic nanoalloys requires high-throughput approaches. To transcend these computational limits, ML has emerged as a powerful tool. Fast-forward prediction frameworks, integrating comprehensive material databases, structural descriptors, and deep learning algorithms, have successfully accelerated the discovery of advanced energy materials.²¹⁷ Adapting these robust ML methodologies provides critical methodological references for the theoretical design and rapid performance prediction of CO₂RR nanoalloys. Such frameworks can rapidly screen multimetallic combinations, effectively guiding targeted experimental synthesis and significantly reducing empirical iterations.

Furthermore, regarding technological development, the focus must shift from half-cell activity to full-cell system performance to realize practical deployment. Actual industrialization demands the simultaneous achievement of stringent benchmarks: current densities of at least 1 A cm⁻², FEs exceeding 80%, operational stability beyond 1000 hours, and high single-pass CO₂ conversion.²¹⁸ As research transitions to MEAs and porous solid electrolyte reactors, overcoming mass transport limitations at the triple-phase boundary becomes paramount. Consequently, future nanoalloy designs must incorporate hydrophobicity engineering to prevent electrode wetting and flooding under high-throughput conditions.^218,219 Scaling up from laboratory electrodes (<5 cm²) to industrial stacks (>1 m²) introduces complex engineering hurdles. Addressing nonuniform fluid distribution and edge effects requires sophisticated flow field architectures combined with advanced ion-exchange membranes to regulate local pH, suppress hydrogen evolution, and mitigate CO₂ crossover.²²⁰ Finally, enhancing economic feasibility is a critical prerequisite. Adapting CO₂RR systems to operate efficiently with low-concentration CO₂ streams, such as industrial flue gas (∼13% CO₂), will drastically reduce operational expenses for high-value C₂₊ production.

As CO₂RR technologies advance toward commercialization, their direct integration with intermittent renewable energy sources, such as photovoltaics and wind power, presents a critical frontier. However, the variable nature of these power supplies introduces multifaceted challenges. Primarily, frequent start–stop cycles severely compromise both catalyst and electrolyzer durability. In zero-gap architectures, discontinuous operation degrades the hydrophobicity of the gas diffusion electrode, triggering flooding that restricts CO₂ mass transport and exacerbates parasitic hydrogen evolution.²²¹ At the atomic level, these transient potentials induce extreme structural stress; power lapses to open-circuit voltages can trigger rapid catalyst dissolution and restructuring, whereas sudden power surges promote carbonate precipitation that blocks active sites.²²² Bridging this electrical gap necessitates robust power electronics or hybrid energy storage buffers, both of which substantially inflate the capital expenditures of the integrated plant. Finally, the diurnal nature of renewable sources introduces ambient temperature swings that destabilize the local reaction environment. Such temperature variations directly perturb CO₂ solubility, interfacial pH, and activation energies, while also accelerating the physical deformation of ion-exchange membranes. Consequently, deploying thermo-mechanically resilient components and adaptive regulation systems is indispensable to mitigate these environmental swings and ensure practical viability.

Conflicts of interest

There are no conflicts to declare.

Data availability

No primary research results, software or code have been included and no new data were generated or analysed as part of this review.

Acknowledgements

Authors thank HUTECH University for support in this study. Mr Khanh Quang Nguyen thanks Vingroup Innovation Foundation (VINIF), VinUniversity for funding in the PhD Scholarship Programme of code VINIF.2025.TS19.

References

1. United Nations Framework Convention on Climate Change (UNFCCC), Report of the Conference of the Parties on its Twenty-First Session, Held in Paris from 30 November to 13 December 2015. Addendum. Part Two: Action Taken by the Conference of the Parties at its Twenty-First Session, United Nations, 2016.
2. The Intergovernmental Panel on Climate Change (IPCC), Global Warming of 1.5°C: IPCC Special Report on Impacts of Global Warming of 1.5 °C above Pre-industrial Levels in Context of Strengthening Response to Climate Change, Sustainable Development, and Efforts to Eradicate Poverty, Cambridge University Press, 1st edn, 2022.
3. Felipe de Carvalho, UN News, 2025.
4. M. Babayev and A. A. Corrêa do Lago, The Baku to Belém Roadmap to 1.3T, United Nations Framework Convention on Climate Change (UNFCCC), 2025.
5. N. Höhne, M. J. Gidden, M. Den Elzen, F. Hans, C. Fyson, A. Geiges, M. L. Jeffery, S. Gonzales-Zuñiga, S. Mooldijk, W. Hare and J. Rogelj, Nat. Clim. Change, 2021, 11, 820–822.
6. H. L. Van Soest, M. G. J. Den Elzen and D. P. Van Vuuren, Nat. Commun., 2021, 12, 2140.
7. P. Guo, K. Liu, X. Liu, R. Liu and Z. Yin, Energy Fuels, 2024, 38, 5659–5675.
8. M. Aresta, A. Dibenedetto and A. Angelini, Chem. Rev., 2014, 114, 1709–1742.
9. K. Li, X. An, K. H. Park, M. Khraisheh and J. Tang, Catal. Today, 2014, 224, 3–12.
10. C. Hepburn, E. Adlen, J. Beddington, E. A. Carter, S. Fuss, N. Mac Dowell, J. C. Minx, P. Smith and C. K. Williams, Nature, 2019, 575, 87–97.
11. M. Hubert, A. M. Esposito, D. Peterson, E. Miller and J. Stanford, in Energy Earth Shots, US Department of Energy, 2024.
12. S. Hughes, S. Henry, M. Turner, A. Zoelle, N. Kuehn, M. Adams, A. J. Eggleston, M. Woods and G. A. Hackett, Quality Guidelines for Energy System Studies: Performing a Techno-Economic Analysis for Carbon Conversion Technologies, National Energy Technology Laboratory (NETL), Pittsburgh, PA, Morgantown, WV, and Albany, OR (United States), 2023.
13. W. Zhang, Y. Hu, L. Ma, G. Zhu, Y. Wang, X. Xue, R. Chen, S. Yang and Z. Jin, Adv. Sci., 2018, 5, 1700275.
14. S. Zhao, R. Jin and R. Jin, ACS Energy Lett., 2018, 3, 452–462.
15. M. S. Yesupatham, B. Honnappa, N. Agamendran, S. Y. Kumar, G. Chellasamy, S. Govindaraju, K. Yun, N. C. S. Selvam, A. Maruthapillai, W. Li and K. Sekar, Adv. Sustainable Syst., 2024, 8, 2300549.
16. S. Cao, H. Chen, Y. Hu, J. Li, C. Yang, Z. Chen, S. Wei, S. Liu, Z. Wang, D. Sun and X. Lu, Chem. Eng. J., 2023, 461, 141936.
17. L. Liu, F. Wang, X. Chu, L. Zhang, S. Zhang, X. Wang, G. Che, S. Song and H. Zhang, Adv. Energy Mater., 2024, 14, 2301575.
18. Q. Wang, T. Luo, X. Cao, Y. Gong, Y. Liu, Y. Xiao, H. Li, F. Gröbmeyer, Y.-R. Lu, T.-S. Chan, C. Ma, K. Liu, J. Fu, S. Zhang, C. Liu, Z. Lin, L. Chai, E. Cortes and M. Liu, Nat. Commun., 2025, 16, 2985.
19. R. Nankya, Y. Xu, A. Elgazzar, P. Zhu, T. Wi, C. Qiu, Y. Feng, F. Che and H. Wang, Angew. Chem., Int. Ed., 2024, 63, e202403671.
20. X. Long, Y. Peng, Z. Yu, Y. Zhang, X. Jiang and H. Yang, Appl. Catal., A, 2025, 701, 120330.
21. A. Elgazzar, P. Zhu, F.-Y. Chen, S. Hao, T.-U. Wi, C. Qiu, V. Okatenko and H. Wang, ACS Energy Lett., 2025, 10, 450–458.
22. J. Li, C. Jing and J. Wang, Nano Res., 2024, 17, 6922–6930.
23. H. Wu, B. Tian, W. Xu, K. K. Abdalla, Y. Kuang, J. Li and X. Sun, J. Am. Chem. Soc., 2024, 146, 22266–22275.
24. S.-C. Wang, X. Ji, R. Hou, L. Qi, P. Jing, X. Xu, B. Liu and J. Zhang, ACS Mater. Lett., 2024, 6, 4728–4737.
25. G.-R. Zhu, Z.-L. Wang, H.-C. Hu, X.-H. Guan and H.-B. Zhu, J. Alloys Compd., 2025, 1026, 180410.
26. Z. Lv, C. Wang, W. Liu, R. Liu, Y. Liu, X. Feng, W. Yang and B. Wang, Adv. Energy Mater., 2024, 14, 2402551.
27. C. Zhan, F. Dattila, C. Rettenmaier, A. Herzog, M. Herran, T. Wagner, F. Scholten, A. Bergmann, N. López and B. Roldan Cuenya, Nat. Energy, 2024, 9, 1485–1496.
28. X. Zhao, G. Hao, Q. Fang, X. Lang, D. Li, D. Zhong, J. Li and Q. Zhao, Appl. Surf. Sci., 2024, 664, 160202.
29. J. Zhou, B. He, P. Huang, D. Wang, Z. Zhuang, J. Xu, C. Pan, Y. Dong, D. Wang, Y. Wang, H. Huang, J. Zhang and Y. Zhu, Angew. Chem., 2025, 137, e202418459.
30. Q. Li, J. Wu, C. Yang, S. Li, C. Long, Z. Zhuang, Q. Li, Z. Guo, X. Huang, Z. Tang, H. Li, D. Wang and Y. Li, J. Am. Chem. Soc., 2025, 147, 6688–6697.
31. C. Chen, M. Wang, Y. Zhang, X. Yang, K. Zou, W. Lin, M. He, H. Wu and B. Han, Green Chem., 2025, 27, 5449–5454.
32. Y. Yan, K. Liu, C. Yang, Y. Chen, X. Lv, C. Hu, L. Zhang and G. Zheng, Small, 2024, 20, 2406345.
33. Y.-J. Zhang, V. Sethuraman, R. Michalsky and A. A. Peterson, ACS Catal., 2014, 4, 3742–3748.
34. A. Goyal, G. Marcandalli, V. A. Mints and M. T. M. Koper, J. Am. Chem. Soc., 2020, 142, 4154–4161.
35. S. Back, J. Lim, N.-Y. Kim, Y.-H. Kim and Y. Jung, Chem. Sci., 2017, 8, 1090–1096.
36. Z. Zhao and G. Lu, Chem. Sci., 2022, 13, 3880–3887.
37. Y. Wang, L. Cao, N. J. Libretto, X. Li, C. Li, Y. Wan, C. He, J. Lee, J. Gregg, H. Zong, D. Su, J. T. Miller, T. Mueller and C. Wang, J. Am. Chem. Soc., 2019, 141, 16635–16642.
38. F. Liu, C. Wu and S. Yang, J. Phys. Chem. C, 2017, 121, 22139–22146.
39. L. Liu, H. Akhoundzadeh, M. Li and H. Huang, Small Methods, 2023, 7, 2300482.
40. O. G. Njoku, P. Fronczak, K. Smeltz and I. T. McCrum, npj Comput. Mater., 2025, 11, 113.
41. X. Ma, Y. Shen, S. Yao, C. An, W. Zhang, J. Zhu, R. Si, C. Guo and C. An, J. Mater. Chem. A, 2020, 8, 3344–3350.
42. Q. Shao, X. Han, K. Li, W. Guan, Y. Ling and Z. Lv, J. Colloid Interface Sci., 2025, 693, 137564.
43. R. Wang and J. Lee, Small, 2024, 2401546.
44. X. Wu, Q. Yan, H. Wang, D. Wu, H. Zhou, H. Li, S. Yang, T. Ma and H. Zhang, Adv. Mater. Interfaces, 2024, 34, 2404535.
45. Y. Zheng, H. Liu and Y. Zhang, ChemSusChem, 2020, 13, 6090–6123.
46. H. S. Moon, S. A. Jaffer, R. K. Miao, E. H. Sargent and D. Sinton, Nat. Rev. Mater., 2026, 11, 244–254.
47. S. Kim, H. I. Lee, S. Kwon, J. Lee, I. Jeon, H.-S. Oh, M. Grätzel, W. A. Goddard and C. Choi, ACS Nano, 2026, 20, 3205–3227.
48. H. B. Yang, S.-F. Hung, S. Liu, K. Yuan, S. Miao, L. Zhang, X. Huang, H.-Y. Wang, W. Cai, R. Chen, J. Gao, X. Yang, W. Chen, Y. Huang, H. M. Chen, C. M. Li, T. Zhang and B. Liu, Nat. Energy, 2018, 3, 140–147.
49. A. R. Woldu, Z. Huang, P. Zhao, L. Hu and D. Astruc, Coord. Chem. Rev., 2022, 454, 214340.
50. P.-F. Sui, M.-R. Gao, Y.-C. Wang and J.-L. Luo, Acc. Mater. Res., 2024, 5, 699–711.
51. J. Han, X. Bai, X. Xu, X. Bai, A. Husile, S. Zhang, L. Qi and J. Guan, Chem. Sci., 2024, 15, 7870–7907.
52. H. Jiang, R. Luo, Y. Li and W. Chen, EcoMat, 2022, 4, e12199.
53. C.-C. Weng, X.-W. Lv, J.-T. Ren, T.-Y. Ma and Z.-Y. Yuan, Electrochem. Energy Rev., 2022, 5, 19.
54. W. Zhang, A. Yu, H. Mao, G. Feng, C. Li, G. Wang, J. Chang, D. Halat, Z. Li, W. Yu, Y. Shi, S. Liu, D. W. Fox, H. Zhuang, A. Cai, B. Wu, F. Joshua, J. R. Martinez, L. Zhai, M. D. Gu, X. Shan, J. A. Reimer, Y. Cui and Y. Yang, J. Am. Chem. Soc., 2024, 146, 21335–21347.
55. T. Ahmad, S. Liu, M. Sajid, K. Li, M. Ali, L. Liu and W. Chen, Nano Res. Energy, 2022, 1, e9120021.
56. L. Zaza, K. Rossi and R. Buonsanti, ACS Energy Lett., 2022, 7, 1284–1291.
57. P. A. Kempler and A. C. Nielander, Nat. Commun., 2023, 14, 1158.
58. J. Qiao, Y. Liu, F. Hong and J. Zhang, Chem. Soc. Rev., 2014, 43, 631–675.
59. O. S. Bushuyev, P. De Luna, C. T. Dinh, L. Tao, G. Saur, J. Van De Lagemaat, S. O. Kelley and E. H. Sargent, Joule, 2018, 2, 825–832.
60. Y. Zang, P. Wei, H. Li, D. Gao and G. Wang, Electrochem. Energy Rev., 2022, 5, 29.
61. Y. Hori, K. Kikuchi and S. Suzuki, Chem. Lett., 1985, 14, 1695–1698.
62. M. Ma, H. A. Hansen, M. Valenti, Z. Wang, A. Cao, M. Dong and W. A. Smith, Nano Energy, 2017, 42, 51–57.
63. T.-H. Wang, C.-Y. Lin, Y.-C. Huang and C.-Y. Li, Electrochim. Acta, 2023, 437, 141500.
64. Y. Wang, C. Li, Z. Fan, Y. Chen, X. Li, L. Cao, C. Wang, L. Wang, D. Su, H. Zhang, T. Mueller and C. Wang, Nano Lett., 2020, 20, 8074–8080.
65. S. Liu, H. Tao, L. Zeng, Q. Liu, Z. Xu, Q. Liu and J.-L. Luo, J. Am. Chem. Soc., 2017, 139, 2160–2163.
66. R. Daiyan, X. Lu, Y. H. Ng and R. Amal, ChemistrySelect, 2017, 2, 879–884.
67. S. Liu, X.-Z. Wang, H. Tao, T. Li, Q. Liu, Z. Xu, X.-Z. Fu and J.-L. Luo, Nano Energy, 2018, 45, 456–462.
68. J. T. Feaster, C. Shi, E. R. Cave, T. Hatsukade, D. N. Abram, K. P. Kuhl, C. Hahn, J. K. Nørskov and T. F. Jaramillo, ACS Catal., 2017, 7, 4822–4827.
69. C. W. Lee, K. D. Yang, D. Nam, J. H. Jang, N. H. Cho, S. W. Im and K. T. Nam, Adv. Mater., 2018, 30, 1704717.
70. J. Zhang, J. Ding, Y. Liu, C. Su, H. Yang, Y. Huang and B. Liu, Joule, 2023, 7, 1700–1744.
71. Y. Zhang, Y. Xiong, Y. Wang, Q. Wang and J. Fan, Nanoscale Horiz., 2025, 10, 2211–2238.
72. T. K. Todorova, M. W. Schreiber and M. Fontecave, ACS Catal., 2020, 10, 1754–1768.
73. L. Li, X. Li, Y. Sun and Y. Xie, Chem. Soc. Rev., 2022, 51, 1234–1252.
74. A. Rendón-Calle, S. Builes and F. Calle-Vallejo, Curr. Opin. Electrochem., 2018, 9, 158–165.
75. G. M. Tomboc, S. Choi, T. Kwon, Y. J. Hwang and K. Lee, Adv. Mater., 2020, 32, 1908398.
76. D. Ma, C. Zhi, Y. Zhang, J. Chen, Y. Zhang and J.-W. Shi, ACS Nano, 2024, 18, 21714–21746.
77. L. Li, K. Yuan and Y. Chen, Acc. Mater. Res., 2022, 3, 584–596.
78. T. N. Nguyen, M. Salehi, Q. V. Le, A. Seifitokaldani and C. T. Dinh, ACS Catal., 2020, 10, 10068–10095.
79. S. Lu, M. Mazur, K. Guo, D. C. Stoian, M. Gu, W. M. Tucho and Z. Yu, Small, 2024, 20, 2309251.
80. Y. Ouyang, L. Shi, X. Bai, Q. Li and J. Wang, Chem. Sci., 2020, 11, 1807–1813.
81. J. Greeley, Annu. Rev. Chem. Biomol. Eng., 2016, 7, 605–635.
82. M. M. Montemore and J. W. Medlin, Catal. Sci. Technol., 2014, 4, 3748–3761.
83. J. Pérez-Ramírez and N. López, Nat. Catal., 2019, 2, 971–976.
84. C. T. Campbell, Acc. Chem. Res., 2019, 52, 984–993.
85. X. Guan, C. Zhao, X. Liu, S. Liu, W. Gao and Q. Jiang, J. Phys. Chem. C, 2020, 124, 25898–25906.
86. S. V. Parmar, D. Kaur and V. Avasare, J. Mater. Chem. A, 2024, 12, 18774–18814.
87. L. Wang, Z. Liu, R. Li, H. Gao, P. Yang, W. Wang, X. Xue, S. Feng, L. Yu and G. Wang, J. Mater. Chem. A, 2023, 11, 24836–24853.
88. V. R. Manikam, K. Y. Cheong and K. A. Razak, Mater. Sci. Eng., B, 2011, 176, 187–203.
89. P. Lu, J. Zhou, Y. Hu, J. Yin, Y. Wang, J. Yu, Y. Ma, Z. Zhu, Z. Zeng and Z. Fan, J. Mater. Chem. A, 2021, 9, 19025–19053.
90. M. Śliwa, H. Zhang, J. Gao, B. O. Stephens, A. J. Patera, D. Raciti, P. D. Hanrahan, Z. A. Warecki, D. L. Foley, K. J. Livi, T. H. Brintlinger, M. L. Taheri, A. S. Hall and T. J. Kempa, Nano Lett., 2024, 24, 13911–13918.
91. R. Ferrando, J. Jellinek and R. L. Johnston, Chem. Rev., 2008, 108, 845–910.
92. X. Shan, H. Zhang, Q. Liu, J. Yu, J. Zhu, R. Li and J. Wang, Appl. Surf. Sci., 2025, 686, 162200.
93. M. Van Der Veer, N. Daems, P. Cool and T. Breugelmans, Green Chem., 2025, 27, 6039–6055.
94. Y. Zhang, F. Chen, X. Yang, Y. Guo, X. Zhang, H. Dong, W. Wang, F. Lu, Z. Lu, H. Liu, H. Liu and Y. Cheng, Small, 2025, e06942.
95. A. S. Nugraha, V. Malgras, J. Kim, J. Bo, C. Li, M. Iqbal, Y. Yamauchi and T. Asahi, Small Methods, 2018, 2, 1800283.
96. L. Mattarozzi, S. Cattarin, N. Comisso, A. Gambirasi, P. Guerriero, M. Musiani, L. Vázquez-Gómez and E. Verlato, Electrochim. Acta, 2014, 140, 337–344.
97. L. Mattarozzi, S. Cattarin, N. Comisso, R. Gerbasi, P. Guerriero, M. Musiani and E. Verlato, Electrochim. Acta, 2017, 230, 365–372.
98. A. Kloke, C. Köhler, R. Gerwig, R. Zengerle and S. Kerzenmacher, Adv. Mater., 2012, 24, 2916–2921.
99. S. Thota, S. Chen and J. Zhao, Chem. Commun., 2016, 52, 5593–5596.
100. L. Castilla-Amorós, P. Schouwink, E. Oveisi, V. Okatenko and R. Buonsanti, J. Am. Chem. Soc., 2022, 144, 18286–18295.
101. H. Yun, W. Choi, D. Shin, H.-S. Oh and Y. J. Hwang, ACS Catal., 2023, 13, 9302–9312.
102. J. Mo, D. Lou, J. Li, X. Tao, Z. Zheng and W. Liu, J. Colloid Interface Sci., 2025, 697, 137934.
103. C. Chen, D. Wu, Z. Li, R. Zhang, C. Kuai, X. Zhao, C. Dong, S. Qiao, H. Liu and X. Du, Adv. Energy Mater., 2019, 9, 1803913.
104. Z. Lin, J. Yue, L. Liang, B. Tang, B. Liu, L. Ren, Y. Li and L. Jiang, Appl. Surf. Sci., 2020, 504, 144461.
105. V. Amendola, S. Scaramuzza, F. Carraro and E. Cattaruzza, J. Colloid Interface Sci., 2017, 489, 18–27.
106. Y. Yao, Z. Huang, P. Xie, S. D. Lacey, R. J. Jacob, H. Xie, F. Chen, A. Nie, T. Pu, M. Rehwoldt, D. Yu, M. R. Zachariah, C. Wang, R. Shahbazian-Yassar, J. Li and L. Hu, Science, 2018, 359, 1489–1494.
107. M. Geng, Y. Zhu, J. Guan, R. Zhang, Q. Zou, L. Wang, B. Guo and M. Zhang, J. Alloys Compd., 2024, 1005, 176180.
108. H. Lee, J.-S. Lee, G.-W. Kwak, J. Kim, K.-M. Kim, D. G. Kang, G.-N. Yun, H.-T. Kim, S.-J. Choi and S.-J. Kim, ACS Nano, 2025, 19, 18095–18107.
109. S. Zhang, Y. Jiang, J. Zhao and Y. Du, Chem. Soc. Rev., 2026, 55, 1494–1513.
110. F.-M. Li, X.-Q. Gao, S.-N. Li, Y. Chen and J.-M. Lee, NPG Asia Mater., 2015, 7, e219.
111. D. G. Shchukin, D. Radziuk and H. Möhwald, Annu. Rev. Mater. Res., 2010, 40, 345–362.
112. X. Xu and J. Guan, Adv. Mater. Interfaces, 2025, 2505823.
113. C. Yang, Y. Gao, T. Ma, M. Bai, C. He, X. Ren, X. Luo, C. Wu, S. Li and C. Cheng, Adv. Mater., 2023, 35, 2301836.
114. B. Hammer and J. K. Norskov, Nature, 1995, 376, 238–240.
115. D. Wei, Y. Wang, C. Dong, Z. Zhang, X. Wang, Y. Huang, Y. Shi, X. Zhao, J. Wang, R. Long, Y. Xiong, F. Dong, M. Li and S. Shen, Angew. Chem., 2023, 135, e202217369.
116. K. Liu, M. Ma, L. Wu, M. Valenti, D. Cardenas-Morcoso, J. P. Hofmann, J. Bisquert, S. Gimenez and W. A. Smith, ACS Appl. Mater. Interfaces, 2019, 11, 16546–16555.
117. J. Wang, F. Zhu, B. Chen, S. Deng, B. Hu, H. Liu, M. Wu, J. Hao, L. Li and W. Shi, Chin. J. Catal., 2023, 49, 132–140.
118. X. Sun, R. B. Araujo, E. C. Dos Santos, Y. Sang, H. Liu and X. Yu, Chem. Soc. Rev., 2024, 53, 7392–7425.
119. C. Yang, B. H. Ko, S. Hwang, Z. Liu, Y. Yao, W. Luc, M. Cui, A. S. Malkani, T. Li, X. Wang, J. Dai, B. Xu, G. Wang, D. Su, F. Jiao and L. Hu, Sci. Adv., 2020, 6, eaaz6844.
120. W. Zhang, J. Zeng, H. Liu, Z. Shi, Y. Tang and Q. Gao, J. Catal., 2019, 372, 277–286.
121. J. Zeng, W. Zhang, Y. Yang, D. Li, X. Yu and Q. Gao, ACS Appl. Mater. Interfaces, 2019, 11, 33074–33081.
122. X. Wang, N. Fu, J.-C. Liu, K. Yu, Z. Li, Z. Xu, X. Liang, P. Zhu, C. Ye, A. Zhou, A. Li, L. Zheng, L.-M. Liu, C. Chen, D. Wang, Q. Peng and Y. Li, J. Am. Chem. Soc., 2022, 144, 23223–23229.
123. S. Zhang, W. Ruan and J. Guan, Adv. Energy Mater., 2025, 15, 2404057.
124. Z. Sun, C. Li, J. Lin, T. Guo, S. Song, Y. Hu, Z. Zhang, W. Yan, Y. Wang, Z. Wei, F. Zhang, K. Zheng, D. Wang, Z. Li, S. Wang and W. Chen, ACS Nano, 2024, 18, 13286–13297.
125. N. Guo, H. Xue, A. Bao, Z. Wang, J. Sun, T. Song, X. Ge, W. Zhang, K. Huang, F. He and Q. Wang, Angew. Chem., 2020, 132, 13882–13888.
126. M. Luo and S. Guo, Nat. Rev. Mater., 2017, 2, 17059.
127. Y. Wei, X. Xu, D. Shi, Y. Jiang, C. Zheng, L. Tan, Z. Liu, S. Zhong and Y. Yu, Small, 2024, 20, 2401017.
128. X. Du, Y. Qin, B. Gao, J. H. Jang, C. Xiao, Y. Li, S. Ding, Z. Song, Y. Su and K. T. Nam, J. Mater. Chem. A, 2022, 10, 7082–7089.
129. S. Ji, Q. Wang, J. Huang, R. Guo, Y. Li, Y. Zheng, R. Li, Y. Zhao, C. Wu and M. Jin, Chem. Eng. J., 2025, 520, 166114.
130. J. Hao, Z. Zhuang, J. Hao, K. Cao, Y. Hu, W. Wu, S. Lu, C. Wang, N. Zhang, D. Wang, M. Du and H. Zhu, ACS Nano, 2022, 16, 3251–3263.
131. M. Mavrikakis, B. Hammer and J. K. Nørskov, Phys. Rev. Lett., 1998, 81, 2819–2822.
132. X. Chen, R. Lu, C. Li, W. Luo, R. Yu, J. Zhu, L. Lv, Y. Dai, S. Gong, Y. Zhou, W. Xiong, J. Wu, H. Cai, X. Wu, Z. Deng, B. Xing, L. Su, F. Wang, F. Chao, W. Chen, C. Xia, Z. Wang and L. Mai, Nat. Commun., 2025, 16, 1927.
133. L. Guo, J. Zhou, F. Liu, X. Meng, Y. Ma, F. Hao, Y. Xiong and Z. Fan, ACS Nano, 2024, 18, 9823–9851.
134. X.-D. Liang, N. Tian, S.-N. Hu, Z.-Y. Zhou and S.-G. Sun, Mater. Rep.: Energy, 2023, 3, 100191.
135. Y. Xing, X. Kong, X. Guo, Y. Liu, Q. Li, Y. Zhang, Y. Sheng, X. Yang, Z. Geng and J. Zeng, Adv. Sci., 2020, 7, 1902989.
136. J. Feng, L. Wu, S. Liu, L. Xu, X. Song, L. Zhang, Q. Zhu, X. Kang, X. Sun and B. Han, J. Am. Chem. Soc., 2023, 145, 9857–9866.
137. P. Liu and J. K. Nørskov, Phys. Chem. Chem. Phys., 2001, 3, 3814–3818.
138. F. Maroun, F. Ozanam, O. M. Magnussen and R. J. Behm, Science, 2001, 293, 1811–1814.
139. L. Bian, Y. Bai, J.-Y. Chen, H.-K. Guo, S. Liu, H. Tian, N. Tian and Z.-L. Wang, ACS Nano, 2025, 19, 9304–9316.
140. Y. Yan, H. Zhou, T. Li, D. Wang, P. Schaaf, G. Guo and X. Wang, Small, 2025, 2501125.
141. Q. Zeng, S. Tian, H. Liu, L. Xu, P. Cui, D. Chen, J. Wang and J. Yang, Adv. Mater. Interfaces, 2023, 33, 2307444.
142. C. Chen, Y. Li, S. Yu, S. Louisia, J. Jin, M. Li, M. B. Ross and P. Yang, Joule, 2020, 4, 1688–1699.
143. D. Meng, M. Zhang, D. Si, M. Mao, Y. Hou, Y. Huang and R. Cao, Angew. Chem., Int. Ed., 2021, 60, 25485–25492.
144. X. Kang, Z. He, F. Wang, Y. Liu and L. Guo, Adv. Mater. Interfaces, 2025, 35, 2419802.
145. W. Xu, H. Shang, J. Guan, X. Yang, X. Jin, L. Tao and Z. Shao, Adv. Mater. Interfaces, 2025, 35, 2412812.
146. Q. Qin, H. Suo, L. Chen, Y. Wang, J. Wang, H. Liu, S. Dou, M. Lao and W. Lai, Adv. Mater. Int., 2024, 11, 2301049.
147. B. Chen, L. Gong, N. Li, H. Pan, Y. Liu, K. Wang and J. Jiang, Adv. Mater. Interfaces, 2024, 34, 2310029.
148. R. Yang, L. Xia, W. Jiang, Y. Cheng, K. Wang, T. Chen, F. Li, X. Zhao, B. Wang, Y. Zhou, F. P. G. De Arquer and M. Huang, Adv. Energy Mater., 2025, 15, 2405964.
149. T. Gong, G. Qiu, M.-R. He, O. V. Safonova, W.-C. Yang, D. Raciti, C. Oses and A. S. Hall, J. Am. Chem. Soc., 2025, 147, 510–518.
150. P. An, L. Wei, H. Li, B. Yang, K. Liu, J. Fu, H. Li, H. Liu, J. Hu, Y.-R. Lu, H. Pan, T.-S. Chan, N. Zhang and M. Liu, J. Mater. Chem. A, 2020, 8, 15936–15941.
151. B. Sheng, D. Cao, Z. Qi, H. Shou, Y. Xia, X. Su, S. Chen, C. Wu, H. Liu, P. J. Chimtali, Y. Chu, C. Liu, X. Wu and L. Song, Mater. Today, 2025, 83, 54–63.
152. T. Liu, H. Luo, T. Ouyang and Z. Liu, Adv. Mater. Interfaces, 2025, 35, 2415367.
153. Z. Zhang, L. Wang, X. Yang, M. Chu, L. Huang, J. Fan, W. Xie, D. Liu, W. Yuan and Y. Situ, ACS Sustainable Chem. Eng., 2024, 12, 18566–18576.
154. W. Lai, Y. Qiao, Y. Wang and H. Huang, Adv. Mater., 2023, 35, 2306288.
155. H. Wu, H. Yu, Y. Chow, P. A. Webley and J. Zhang, Adv. Mater., 2024, 36, 2403217.
156. S. Kuang, M. Li, X. Chen, H. Chi, J. Lin, Z. Hu, S. Hu, S. Zhang and X. Ma, Chin. Chem. Lett., 2023, 34, 108013.
157. Y. Jia, Y. Ma, W.-E. Yang, Y. Zhu, X. Zhang, S. Xie, T.-W. Kuo, S. Dai and K.-W. Wang, Chem. Eng. J., 2024, 488, 150946.
158. G. Mei, Z. Sun, Z. Fang, Y. Dan, X. Lu, J. Tang, W. Guo, Y. Zhai and Z. Zhu, Small, 2025, 21, 2504086.
159. S. Liang, L. Huang, Y. Gao, Q. Wang and B. Liu, Adv. Sci., 2021, 8, 2102886.
160. S. Nitopi, E. Bertheussen, S. B. Scott, X. Liu, A. K. Engstfeld, S. Horch, B. Seger, I. E. L. Stephens, K. Chan, C. Hahn, J. K. Nørskov, T. F. Jaramillo and I. Chorkendorff, Chem. Rev., 2019, 119, 7610–7672.
161. H. Huang, K. Yue, K. Lei, B. Y. Xia and Y. Yan, J. Phys. Chem. C, 2024, 128, 12067–12074.
162. Y. Wang, R. Xie, N. Ci, Z. Zhu, C. Li, G. Chai, H.-J. Qiu and Y. Zhang, J. Energy Chem., 2024, 99, 450–457.
163. Y. Song, Y. Hua, X. Liu and Z. Gao, J. Alloys Compd., 2025, 1022, 179927.
164. Y. Guo, Y. Luo, H. Wang, X. Chen, X. Liu, B. Guo, J. Sun, R. Wang, W. Li and C. Zhao, Fuel, 2025, 392, 134821.
165. R. Zou, R. Luo, F. Liu, Y. Wang, M. S. Riaz, S. Shen, G. Yang, Z. Tang, H. Huang, G. Li and Q. Bi, Small, 2025, 21, 2501030.
166. X. Linghu, J. Chen, L. Jiang and T. Wang, Nano Mater. Sci., 2024.
167. Y. Kuang, H. Rabiee, L. Ge, T. E. Rufford, Z. Yuan, J. Bell and H. Wang, Energy Environ. Mater., 2023, 6, e12596.
168. W. Wu, J. Zhu, Y. Tong, S. Xiang and P. Chen, Nano Res., 2024, 17, 3684–3692.
169. T. Fan, J. Zhang, X. Zhang, M. Wang, X. Yi, Y. Lum and Z. Chen, Nano Energy, 2024, 130, 110135.
170. B. Li, J. Chen, L. Wang, D. Xia, S. Mao, L. Xi, S. Ying, H. Zhang and Y. Wang, Appl. Catal., B, 2025, 363, 124784.
171. Z. Zhang, M. Li, S. Yang, Q. Ma, J. Dang, R. Feng, Z. Bai, D. Liu, M. Feng and Z. Chen, Adv. Mater., 2024, 2407266.
172. H. Huang, X. Ding, X. Mao, Y. Yan, F. Lv, B. Pan, W. Huang, L. Wang, N. Han and Y. Li, Adv. Mater. Interfaces, 2024, 34, 2408966.
173. X. Huang, X. Yu, H. Liu, S. Xia, Y. Jia, Z. Chen, J. Xu and Y. Xia, Adv. Mater. Interfaces, 2025, e13814.
174. C. Fang, L. Huang, W. Gao, X. Jiang, H. Liu, R. Hu, X. Li, J. Yu and W. Zhou, Adv. Energy Mater., 2024, 14, 2400813.
175. A. Huang, T. Zhou, J. Zhang, Y. Zhang, Y. Wu, Y. Wang and W. Luo, ChemCatChem, 2024, 16, e202400504.
176. S. Go, W. Kwon, D. Hong, T. Lee, S.-H. Oh, D. Bae, J.-H. Kim, S. Lim, Y.-C. Joo and D.-H. Nam, Nanoscale Horiz., 2024, 9, 2295–2305.
177. J. Huang, Y. Yang, X. Liang, B. Chen, Y. Shen, Y. Chen, J. Yang, Y. Yu, F. Huang, H. He, P. Chen, L. Zhou and A. Guan, Adv. Sci., 2025, 12, 2417247.
178. J. Li, M. Wei, B. Ji, S. Hu, J. Xue, D. Zhao, H. Wang, C. Liu, Y. Ye, J. Xu, J. Zeng, R. Ye, Y. Zheng, T. Zheng and C. Xia, Angew. Chem., 2025, 137, e202417008.
179. H. Sun, Y. Wen, Y. Li, R. Fan, Y. Lu, C. Feng, H. Liu, Y. Wang, Z. Ma, Y. Zhang, R. Amal and Z. Yan, Appl. Catal., B, 2026, 381, 125891.
180. A. Husile, Z. Wang and J. Guan, Chem. Sci., 2025, 16, 5413–5446.
181. X. Zhang, S.-X. Guo, K. A. Gandionco, A. M. Bond and J. Zhang, Mater. Today Adv., 2020, 7, 100074.
182. L. Lu, X. Sun, J. Ma, D. Yang, H. Wu, B. Zhang, J. Zhang and B. Han, Angew. Chem., 2018, 130, 14345–14349.
183. S. Payra, S. Shenoy, C. Chakraborty, K. Tarafder and S. Roy, ACS Appl. Mater. Interfaces, 2020, 12, 19402–19414.
184. G. Bhattacharya, R. Manna, P. Sardar, S. Rahut, S. Kumar and A. N. Samanta, J. Environ. Chem. Eng., 2025, 13, 117451.
185. S. Desmons, J. Bonin, M. Robert and S. Bontemps, Chem. Sci., 2024, 15, 15023–15086.
186. Y. Xiao, F. Yu, C. Xia, D. Zhu, J. Chen, N. Liu, Y. Zhao, R. Qi, W. Guo, B. You, T. Yao, Y. Pang, Z. Wang, H. Wang, F. Song and B. Y. Xia, J. Am. Chem. Soc., 2025, 147, 15654–15665.
187. S. Zhang, B. Zhang, S. Yang, T. Shao, X. Li, R. Cao and M. Cao, ACS Appl. Nano Mater., 2025, 8, 1893–1902.
188. H. Wang, F. Zhang, Y. Li, Y. Pang, X. Zhao, Z. Song, W. Wang, J. Sun and Y. Mao, J. Colloid Interface Sci., 2025, 678, 506–514.
189. Y. Zheng, J. Zhang, Z. Ma, G. Zhang, H. Zhang, X. Fu, Y. Ma, F. Liu, M. Liu and H. Huang, Small, 2022, 18, 2201695.
190. Z.-Y. Zhang, H.-B. Wang, F.-F. Zhang, J.-W. Li, X.-Z. Hu, S.-W. Yan, Y.-M. Bai, X. Zhang, G.-R. Shen, P.-F. Yin, J. Yang, C.-K. Dong, J. Mao, H. Liu and X.-W. Du, Rare Met., 2024, 43, 1513–1523.
191. J. Wang, W. Li, J. Peng, S. Shang, X. Fu, G. He and Q. Zhang, Vacuum, 2024, 222, 113017.
192. N. Wang, J. Li, S. Gan, X. Chen, G. Wang, Q. Chen, T. Zhou and B. Li, Langmuir, 2025, 41, 19821–19833.
193. Y. Xiao, Y. Yang, W. Liu and J. Liu, J. Mater. Chem. A, 2025, 13, 26812–26822.
194. B. S. Crandall, Z. Qi, A. C. Foucher, S. E. Weitzner, S. A. Akhade, X. Liu, A. R. Kashi, A. K. Buckley, S. Ma, E. A. Stach, J. B. Varley, F. Jiao and J. Biener, Small, 2024, 20, 2401656.
195. J. Du, J. Liu, Z. Liu, N. Tian, Y. Zhang and X. Gu, Chem. Commun., 2025, 61, 13316–13326.
196. S. Wang, F. Li, J. Zhao, Y. Zeng, Y. Li, Z.-Y. Lin, T.-J. Lee, S. Liu, X. Ren, W. Wang, Y. Chen, S.-F. Hung, Y.-R. Lu, Y. Cui, X. Yang, X. Li, Y. Huang and B. Liu, Nat. Commun., 2024, 15, 10247.
197. Z. Wang, Y. Li, Z. Ma, D. Wang and X. Ren, iScience, 2024, 27, 110437.
198. J.-Y. Kim, H. S. Ahn, I. Kim, D. Hong, T. Lee, J. Jo, H. Kim, M. K. Kwak, H. G. Kim, G. Kang, S. Go, W. H. Ryu, G.-D. Lee, M. Kim, D.-H. Nam, E. S. Park and Y.-C. Joo, Nat. Synth., 2023, 3, 452–465.
199. C. Peng, J. Ma, G. Luo, S. Yan, J. Zhang, Y. Chen, N. Chen, Z. Wang, W. Wei, T. Sham, Y. Zheng, M. Kuang and G. Zheng, Angew. Chem., Int. Ed., 2024, 63, e202316907.
200. T. Guo, C. Ma, X. Wang, Z. Men, B. Liu and A. H. Bedane, Energy Fuels, 2025, 39, 9532–9545.
201. M. Wang, M. Fang, Y. Liu, C. Chen, Y. Zhang, S. Jia, H. Wu, M. He and B. Han, J. Am. Chem. Soc., 2025, 147, 16450–16458.
202. D. Luo, W. Dai, K. Wu, S. Liu, C. Tang, Y. Sun, F. Dong and C. Long, Nanoscale, 2025, 17, 9057–9071.
203. F. A. Rollier, V. Muravev, N. Kosinov, T. Wissink, D. Anastasiadou, B. Ligt, L. Barthe, M. Costa Figueiredo and E. J. M. Hensen, J. Mater. Chem. A, 2025, 13, 2285–2300.
204. E. Ruiz-López, J. Gandara-Loe, F. Baena-Moreno, T. R. Reina and J. A. Odriozola, Renewable Sustainable Energy Rev., 2022, 161, 112329.
205. A. J. Garza, A. T. Bell and M. Head-Gordon, ACS Catal., 2018, 8, 1490–1499.
206. A. Shekhawat, M. A. A. Mahbub and W. Schuhmann, ChemElectroChem, 2026, 13, e202500413.
207. X. Hu, J. Li, Z. Zhou and L. Wen, ACS Mater. Lett., 2023, 5, 85–94.
208. S. Nie, L. Wu, Q. Liu and X. Wang, J. Am. Chem. Soc., 2024, 146, 29364–29372.
209. K. Zhang, S. Chen, Y. Ge, X. Zhang, Z. Zhang, C. Ye, G. Pan and L. Hu, ACS Appl. Nano Mater., 2025, 8, 22285–22294.
210. Q. Sun, Y. Zhao, X. Tan, C. Jia, Z. Su, Q. Meyer, M. I. Ahmed and C. Zhao, ACS Catal., 2023, 13, 5689–5696.
211. X. Zhong, H.-J. Peng, C. Xia and X. Liu, J. Mater. Chem. A, 2024, 12, 19663–19684.
212. M. Zheng, C. Lv and X. Zhou, Chem.–Eur. J., 2025, 31, e202404354.
213. S. Bae, S. Y. Hwang, G. Yun, Y. Gwon, S. Y. Kim, C. K. Rhee and Y. Sohn, Energy Fuels, 2024, 38, 15497–15514.
214. S. Matsuda, R. Ishibashi and M. Umeda, ACS Omega, 2025, 10, 9630–9638.
215. M. G. Lee, S. Kandambeth, X.-Y. Li, O. Shekhah, A. Ozden, J. Wicks, P. Ou, S. Wang, R. Dorakhan, S. Park, P. M. Bhatt, V. S. Kale, D. Sinton, M. Eddaoudi and E. H. Sargent, J. Am. Chem. Soc., 2024, 146, 14267–14277.
216. D. Zhang, H. Lu, C. Duan, X. Yuan, Z. Zhu, Y. Qin, Y. Song and Y. Jin, Energy Mater. Adv., 2025, 6, 0141.
217. X. Liu, K. Yang, L. Zhang, W. Wang, S. Zhou, B. Wu, M. Xiong, S. Yang and R. Tan, Energy Mater. Adv., 2024, 5, 0131.
218. H. Zheng, K. Shi, H. Dong, Y. Yang, P. Yin, B. Shen and J. Wang, Energy Fuels, 2025, 39, 9316–9344.
219. J. Peng, L. Sun, Y. Li, Q. Zhang, X. Ren, X. Li, J. Zhang, X. Sun, Z. Song and L. Zhang, Electrochem. Energy Rev., 2025, 8, 16.
220. L. Xue, Y. Qi, Z. Li, H. Yang, R. Liu and B. Zhang, J. Energy Chem., 2025, 110, 88–108.
221. J. Chen, D. Du, Q. Huang and J. Luo, ACS Energy Lett., 2025, 10, 4027–4034.
222. E. R. Cofell, Z. Park, U. O. Nwabara, L. C. Harris, S. S. Bhargava, A. A. Gewirth and P. J. A. Kenis, ACS Appl. Energy Mater., 2022, 5, 12013–12021.

---