Electrochemical Reduction of Carbon Dioxide: Recent Advances on Au-Based Nanocatalysts

Abstract

The electrocatalytic reduction of CO₂ to other high value-added chemicals under ambient conditions is a promising and ecofriendly strategy to achieve sustainable carbon recycling. However, the CO₂ reduction reaction (CO₂RR) is still confronted with a large number of challenges, such as high reaction overpotential and low product selectivity. Therefore, the rapid development of appropriate electrocatalysts is the key to promoting CO₂ electroreduction. Over the past few decades, Au-based nanocatalysts have been demonstrated to be promising for the selective CO₂RR to CO owing to their low reaction overpotential, good product selectivity, high Faraday efficiency and inhibition of the hydrogen evolution reaction. In this respect, this review first introduces the fundamentals of the electrochemical reduction of CO₂ and then focuses on recent accomplishments with respect to Au-based nanocatalysts for CO₂RR. The manipulation of various factors, e.g., the nanoporous structure, nanoparticle size, composition, morphology, support and ligand, allows for the identification of several clues for excellent Au-based nanocatalysts. We hope that this review will offer readers some important insights on Au-based catalyst design and provide new ideas for developing robust electrocatalysts.

1. Introduction

From the beginning of the industrial revolution, the excessive depletion of fossil fuels has resulted in a sharp rise in the concentration of carbon dioxide (CO₂) in the atmosphere to an alarming rate [1]. Correspondingly, the concentration of CO₂ in the global atmosphere has risen from 280 parts per million (ppm) in 1750 to 415 ppm in 2021 and is predicted to reach nearly 600 ppm by the end of this century [2]. This leads to a series of severe environmental problems, such as global warming, extreme weather and desertification [3]. Therefore, one of the most urgent tasks for mankind is how to effectively transform and utilize CO₂ for alleviating the above-mentioned environmental problems and, to a certain extent, humanity’s dependence on fossil energy [4,5].

CO₂ is a linear molecule with two identical C=O double bonds, which are delocalized π bonds with three centers and four electrons [6]. These bonds are similar to triple bonds and are thermodynamically stable [7]. First of all, the activation of CO₂ requires high energy, making the reaction energetically prohibitive. Second, the CO₂ reduction reaction often proceeds without the desirable products [8]. Up until now, considerable research efforts have been devoted to converting CO₂ into other carbon compounds, including electrochemical [9], photochemical [10], thermochemical [11] and biochemical [12] compounds. Surprisingly, the electrochemical carbon dioxide reduction reaction is especially appealing since it not only can be implemented under room temperature and atmospheric pressure but also has good compatibility and complementarity with renewable energy sources [13]. In particular, it can convert excess electrical energy during a specific period of time and store it in chemical bonds [14]. As the price of renewable energy sources (e.g., solar, wind, tidal energy) continues to fall, the solution of carbon recycling on the basis of the electrochemical CO₂ reduction reaction (CO₂RR) holds great promise [15]. As a result, it is imperative to develop remarkable electrocatalysts which significantly speed up the reaction rate and regulate the reaction selectivity toward the target products [16,17]. To date, different types of metal-based electrocatalysts such as Au [18,19,20,21,22,23], Cu [24,25,26,27,28,29,30,31,32,33,34,35,36], Pd [37,38,39,40,41,42,43,44,45,46], Ag [47,48,49,50,51,52,53,54], Bi [32,55,56,57,58], Sn [59,60,61,62,63] and Co [64,65,66,67,68,69,70,71] have been intensively investigated regarding electrochemical CO₂ reduction.

Gold-based catalysts with high Faraday efficiency and low overpotential are of particular interest in this context owing to their selective conversion of CO₂ to CO [72]. This review thus aims to summarize the recent progress on using nanostructured gold as a catalyst to promote the CO₂RR. It starts with the fundamentals of the electrochemical CO₂RR and then concentrates on select representative Au-based nanocatalysts for elucidating their structure–property correlation and catalytic mechanisms in CO₂RR catalysis. Thus, it aims to provide readers with a systematic picture for fundamentally understanding CO₂RR catalytic processes and for future developments in the conversion of CO₂ to CO from the fundamentals to industrialization.

2. Electrochemical Reduction of CO₂

As schematically described in Figure 1, electrochemical CO₂RR systems are composed of an anode, a cathode, an aqueous solution electrolyte saturated with CO₂ and a membrane, in which the oxygen evolution reaction takes place at the anode and the CO₂RR takes place at the cathode.

The electrolyte is favorable to transport CO₂ to the surface of electrocatalysts and can alter the energy barrier of the CO₂RR by the pH value, which has the following two aspects: one is to regulate the overall pH of the system with different salt or salt mixture, and the other is to modulate the local pH through the OH⁻ generated by the CO₂RR. The membrane separates oxidation and reduction products while maintaining the charge balance and forms a closed loop, allowing the transfer of protons to the cathode [16].

The electrochemical reduction of CO₂ is an extremely complex process. As illustrated in

Table 1. Electrocatalytic CO₂ reduction reactions with equilibrium potentials in aqueous solution. Adapted from Ref. [35]. Copyright 2019, American Chemical Society.

Reaction	E0 (V vs. RHE)	Product Name
2H+ + 2e− → H2(g)	0	Hydrogen
CO2 + 2H+ + 2e− → CO(g) + H2O	−0.10	Carbon monoxide
CO2 + 2H+ + 2e− → HCOOH(aq)	−0.12	Formic acid
CO2 + 6H+ + 6e− → CH3OH(aq) + H2O	0.03	Methanol
CO2 + 8H+ + 8e− → CH4(g) + 2H2O	0.17	Methane
2CO2 + 12H+ + 12e− → C2H4(g) + 4H2O	0.08	Ethylene
2CO2 + 12H+ + 12e− → C2H5OH(aq) + 3H2O	0.09	Ethanol
2CO2 + 14H+ + 14e− → C2H6(g) + 4H2O	0.14	Ethane

, by transferring two, four, six or eight electrons, many different carbon-containing compounds can be obtained, such as carbon monoxide (CO), methane (CH₄), ethylene (C₂H₄), hydrogen (H₂), formic acid (HCOOH), methanol (CH₃OH) and ethanol (CH₃CH₂OH).

Although the thermodynamic cost of CO₂ reduction is equivalent to that of the hydrogen evolution reaction, a higher energy input (such as a high overpotential) is required to achieve a considerable CO₂ reduction reaction rate [73,74]. When sufficient negative potentials are applied, the competing reaction (HER) becomes a problem. At the same time, CO₂ reduction involves multiple elementary reactions, which further increases the chemical inertness of the reaction and the difficulty in terms of selectivity regulation [75,76]. Under a certain experimental condition, the CO₂RR on different kinds of electrocatalysts can be carried out through different reaction pathways. Although certain reaction pathways share identical intermediates, they respectively generate a variety of products, as displayed in Figure 2 [34].

In particular, the final products depend on the binding strength of these intermediates to the electrocatalyst surface. In order to achieve promising metals from the CO₂ reduction reaction and to produce individual products, the relationship between the binding affinities of the metal surface and the final products was explored from a theoretical and experimental point of view [77,78]. The stable intermediates for the selected reaction path is determined by the linear scaling relationships between the binding strength of reaction intermediates such as *OCHO, *COOH, *CO and *H on the surface of the metal electrocatalyst. The *CO binding strength on the surface of the catalyst is used as an evaluation index for predicting the main product, that is, CO, formate, hydrocarbon or H₂. In order to compare electrochemical CO₂ reduction reaction performance, a volcano plot is usually drawn, which plots overpotential against *CO binding strength, as shown in Figure 3. Based on the Sabatier principle, the interaction between catalyst and adsorbent can neither be too strong nor too weak. Consequently, the volcano plot is consistent with the above-stated principle [79].

At present, it is universally accepted that the rate-determining step of the CO₂RR is usually the first electron transfer to form *CO₂⁻ intermediate (where * refers to the adsorption site), and the reactivity of *CO₂⁻ is the decisive factor in the final product distributions [74]. In the subsequent steps, when the oxygen atom of *CO₂⁻ is bound to the metal electrocatalyst surface, the carbon atom will be protonated to form *OCHO, and then formic acid or formate will be further formed. On the contrary, when the carbon atom of *CO₂⁻ is bound to the metal electrocatalyst surface, the oxygen atom will be protonated to form *COOH. Followed by electron transfer, *CO will be further generated, eventually resulting in CO desorption from the catalyst surface [76]. All in all, the selectivity of the CO₂RR is closely related to the binding capability of *OCHO, *COOH, *CO and *H on the catalyst surface. For instance, Au is thermodynamically conducive to the adsorption of *COOH, and the surface has a weak ability to bind *CO, which makes CO easily desorb from the surface and become the main reduction product [41].

3. Au-Based Nanocatalysts for Electrochemical Reduction of CO₂

Gold (Au) is a precious metal material. It has been widely studied for selectively reducing CO₂ to CO with a lower overpotential and higher Faraday efficiency compared with any other metals in the CO₂RR [80]. At an early stage, Jung et al. used first-principle calculations to recognize the active sites on Au for the CO₂RR and compared the catalytic activity and selectivity of three reaction sites of nanoparticles, i.e., low-index facets, edge sites and corner sites [81]. They found that the corner sites on Au are the most active for the CO₂RR and reducing the size of Au nanoparticles up to 2 nm can lead to an undesirable hydrogen evolution reaction, as illustrated in Figure 4.

These findings were supported by recent experimental observations. Theoretical investigations point out the direction for the further experimental design of efficient Au-based nanocatalysts. Up until now, it has been reported that the Faraday efficiency of Au-based nanocatalysts to generate CO can be as high as 100% in the low overpotential range. Although encouraging progress has been made, developing Au-based catalysts with reduced Au contents while maintaining satisfactory catalytic activity and stability has been less successful. Hence, it is worthwhile mentioning that the method of reducing the content of Au in Au-based nanocatalysts is the key to enhancing electrochemical activity for the CO₂RR. Based on this, in recent years, a variety of effective strategies using the controllable preparation of Au-based nanocatalysts have been developed to improve their electrochemical CO₂RR performance, e.g., the regulation of Au dispersion and the electronic structure to expose more active sites. In the following, we will aim to highlight recently developed Au-based nanocatalysts and concentrate on their morphology-/structure-dependent catalysis for electrochemical CO₂ reduction.

3.1. Nanoporous Effect of Au Catalysts

The notion that metastable grain boundaries can be used as more active sites has been intensively investigated, and this has prompted people to search for porous structures which are capable of maximizing those sites [82]. Among them, the continuous interconnected network of metals plays a vital role in better controlling mass transfer for further catalysis enhancement [83]. Based on this point, Jeon et al. proposed a three-dimensional hierarchically porous gold nanostructure with interconnected macroporous channels ranging from 200 nm to 300 nm and nanopores (approximately 10 nm) prepared by adjacent field nanopatterns, as displayed in Figure 5a–c [84].

On one hand, the interconnected macroporous channels provide effective mass transfer during the electrolysis process; on the other hand, the nanopores provide large active areas. It is noteworthy that hierarchically porous Au is capable of achieving a very high CO selectivity of 85% at a low overpotential of 0.264 V, as illustrated in Figure 5d.

In contrast to dealloyed porous gold, hierarchically porous Au catalysts bring an improvement in mass activity of about 3.96 times, boosting the efficiency of expensive Au, as shown in Figure 5e. In another study, Atwater et al. constructed a nanoporous gold film with pore sizes ranging from 10 nm to 30 nm [21]. The considerably improved electrocatalytic activity was a result of large electrochemical active areas, rich grain boundaries and pH gradients in the nanoporous network. Nanoporous gold films produce a maximum CO Faradaic efficiency of 99% with 6 mA cm⁻² CO partial current density in 50 mM K₂CO₃. Long-term testing suggests that nanoporous gold films can maintain a CO Faraday efficiency of 80% in electrolysis for 110 h. Such a pore diameter distribution allows for the creating and controlling of local pH gradients in the porous network. Gao et al. synthesized an unsupported nanoporous gold leaf catalyst by employing a facile one-step dealloying method [85].

Nanoporous gold leaf catalysts exhibit a mass activity of 20.51 A g⁻¹ with a 90% CO Faraday efficiency compared with nonporous gold foil at −1.2 V (vs. Ag/AgCl). Similarly, Qian et al. reported a nanoporous Au electrocatalyst synthesized by electrochemical dealloying [86]. A Faraday efficiency of up to 98% at an overpotential of 390 mV was obtained for CO₂ conversion to CO. In contrast to bulk materials, dealloyed nanoporous gold electrocatalysts with high activity are attributed to abundant step/kink sites that are active for electrochemical reactions [87,88]. Despite the enhancement in selectivity, nanoporous structures have their own restrictions on catalysis owing to the difficulty in realizing fast mass transfer in the pores and in reducing the dimensions of the catalyst to further increase its catalytic performance [22].

3.2. Size Effect of Au Catalysts

As is widely known, as catalysts, nanoparticles provide an effective platform to yield more active sites for the CO₂RR without mass transfer problems. On the surface of a nanoparticle, atoms on the corner sites, edge sites and the crystal planes have different coordination numbers and chemical potentials. Therefore, in principle, the catalysis of nanoparticles can be tuned by controlling their sizes and surface structures [89]. On the basis of this, the size of metal nanoparticles is of great importance for the activity and selectivity of catalysts. From the perspective of geometric structures, with the decrease of particle size, the proportion of low coordination atoms will gradually increase and be fully exposed [90]. Furthermore, it is commonly known that the under-coordinated sites on the surface can be functioned as active sites of heterogeneous catalysis, which can notably alter the proportion and distribution of active centers of nanocatalysts [91]. In terms of electronic structure, the electronic energy level of metal particles will also be significantly changed as a result of the quantum size effect, which may speed up the charge transfer rate between catalytic materials and reactants, further changing the binding energy between them. To elucidate the size effect of Au nanosized catalysts, Sun et al. synthesized four monodisperse Au nanoparticles with particle sizes of 4, 6, 8 and 10 nm, respectively, as shown in Figure 6a [92]. They found that when the Au size is 8 nm, the CO Faraday efficiency is its highest, but the current density of CO is at its highest at an Au size of 4 nm, as can be seen from Figure 6b. The smaller the size of Au nanoparticles, the higher the mass activity. At −0.9 V versus RHE, 4-nm Au nanoparticles show a maximum 14 A g⁻¹ CO partial mass activity in CO₂-saturated 0.5 M KHCO₃, as displayed in Figure 6c. However, H₂ is the dominant product for smaller Au nanoparticles. At −0.67 V versus RHE, a maximum CO Faraday efficiency of 90% is observed for 8 nm Au nanoparticles. The researchers inferred that the edges of the nanoparticles are most selective for CO, while the corners are more likely to evolve hydrogen. For nanoparticles of different sizes, the proportion of edges and angles are different. There are many edges and corners in smaller nanoparticles, causing the highest current density of CO at 4 nm. Unfortunately, the selectivity of CO is not high, because hydrogen evolution is also enhanced. With the rapid decrease in angle content, 8-nm Au nanoparticles with a certain number of edges exhibit the optimal CO selectivity and better current density. Researchers used the first principle calculation to identify the relationships between particle size and catalytic activity [81]. They suggested that reducing the size of Au nanoparticles to 2 nm will instead increase the hydrogen evolution reaction due to the quantum size effect of ultrafine particles. For example, Strasser et al. also studied the size effect of a series of Au nanoparticles with a size of 1–8 nm and found similar results to those reported previously, as illustrated in Figure 6d [93]. Decreasing the size of Au nanoparticles from 8 nm to 1 nm results in a nearly 13-fold enhancement in the current density.

Concerning, 1 nm Au nanoparticles, they show a maximum 60 mA cm⁻² CO partial current density in CO₂-saturated 0.1 M KHCO₃, while the CO partial current density for 8 nm Au nanoparticles is much lower (25 mA cm⁻²), as exhibited in Figure 6e. Density function theory (DFT) calculations show that smaller Au nanoparticles have a high density of low coordination surface sites. In addition, this conclusion further demonstrates that H₂ can preferentially be evolved for very small, free-ligand Au nanoparticles. The above-mentioned studies mean that the proportion of CO and H₂ can be adjusted by tailoring the size of the catalyst.

To elucidate the relationships between ligand-capped Au nanoclusters and their performance in the CO₂RR, Kauffman et al. synthesized negatively charged Au₂₅(SC₂H₄Ph)₁₈⁻ nanoclusters with a size of ca. 1 nm, abbreviated as Au₂₅⁻ [90]. Electrochemical evaluations illustrate that Au₂₅⁻ is a wonderful CO₂RR electrocatalyst and exhibits remarkable CO₂RR performance in terms of a lower overpotential, higher CO Faraday efficiency and larger current density compared with 2-nm, 5-nm and bulk Au electrocatalysts. At −1 V versus RHE, a current density of 17 mA cm⁻² for Au₂₅⁻ is achieved in CO₂-saturated 0.1 M KHCO₃. Considerable research effort has been devoted to the size-dependence of electrocatalysts. It was found that Au₂₅⁻ nanoclusters produce an approximately 300 mV reduction in CO onset potential and a seven-fold increase in CO production relative to 2 nm Au nanoparticles. It is speculated that the negative charge on Au₂₅⁻ favors *COOH intermediate adsorption and promotes electron transfer, resulting in a remarkable enhancement in CO₂ reduction.

Identical conclusions were obtained in subsequent studies, which revealed that the conversion rate of CO₂ into CO on Au₂₅⁻ nanoclusters is higher than that of neutral Au₂₅⁰ nanoclusters and positively charged Au₂₅⁺ nanoclusters. This is consistent with the observations of Cao, Kim and co-workers [94]. Enhanced CO₂ reduction behavior can be ascribed to the negative charge from electron-donating ligands. At −1 V versus RHE, long-term testing showed stable CO₂RR activity and CO Faraday efficiency over a 36-h electrolysis experiment in CO₂-saturated 0.1 M KHCO₃. An average mass activity of approximately 1656 A g⁻¹ and a CO Faraday efficiency of approximately 87% was observed for Au₂₅⁻.

3.3. Morphological Effect of Au Catalystss

It is worthwhile mentioning that the highly tunable morphology of nanomaterials provides a variety of nanostructures with unique surface atomic compositions. Crystals can often be grown into specific morphologies by controlling the kinetics and thermodynamics of the nanoparticle growth [95]. Subsequently, different crystal planes will be exposed, affecting the ratio of corner sites, edge sites and surface atoms on the plane. It is generally believed that morphology regulation has lead to a broad new idea in the design of Au-based nanocatalysts [96]. Preparing Au nanoparticles with high index facets by changing their morphology is an appealing way to improve their catalytic performance for the CO₂RR. For example, Nam et al. synthesized concave rhombic Au decahedrons with thiol ligands, as depicted in Figure 7a [97]. The concave rhombic Au nanoparticles are capable of attaining a maximum Faradaic efficiency of 90% in CO₂-saturated 0.5 M KHCO₃. Compared to nonconcave Au nanoparticles, Au nanocubes and Au films, concave rhombic Au nanoparticles exhibit lower overpotentials, a higher Faradaic efficiency and a larger current density, as indicated in Figure 7c. The enhanced CO₂ electrocatalytic activity predominantly originates from the high density of edge sites and high index facets on the surface, which is consistent with studies reported by Zhu et al., as illustrated in Figure 7b [92].

In addition, Xia et al. studied the CO₂RR on 50-nm gold colloid and 50-nm tetrahedron, as can be seen from Figure 7d,e [98]. A maximum CO Faraday efficiency of 88.8% on Au tris-octahedron at −0.6 V was observed. In comparison, Au colloids produce q 59.04% CO Faraday efficiency at −0.7 V, as exhibited in Figure 7f. Density function theory calculations show that the facets on an Au (221) tetrahedron are more conducive to stabilizing *COOH intermediate than those of Au (111), resulting in a lower overpotential and higher Faradaic efficiency. The edge atoms with a low coordination number on Au tetrahedrons binds CO₂ more strongly than in-plane atoms, making it easier to hydrogenate it to *COOH.

It is worth noting that Au nanoparticles with low index facets, such as Au (100) and Au (111) cannot significantly improve CO₂RR activity, while Au nanoparticles with high density edge sites and high index facets favor the adsorption of *COOH and desorption of *CO, leading to a remarkable performance in terms of the CO₂RR. Sun et al. developed a simple seed-mediated growth method to prepare a series of 2-nm Au nanowires (NWs) [99]. At −0.55 V versus RHE, a mass activity of approximately 6.5 A g⁻¹ in CO₂-saturated 0.5 M KHCO₃ was obtained for 500 nm Au NWs. It is unexpected that this value exceeds previously reported values compared with 4–10 nm Au nanoparticles in the similar potential range. Structural modeling indicates that the enhanced CO₂RR performance of Au NWs is a result of a higher fraction of edge sites, facilitating the adsorption of *COOH intermediates and desorption of *CO intermediates.

In another study, Liu et al. investigated the use of Au nanoneedles to electrochemically convert CO₂ to CO [100]. They achieved an onset potential of only 0.07 V and a FE of 95% for over 8 h. The Au nanoneedles exhibited a greater CO current density than Au nanorods, oxide-derived noble metal catalysts and Au nanoparticles. The tips of Au nanoneedles exhibit a high electric field, which can concentrate electrolyte cations, resulting in higher local CO₂ concentrations near the active reaction surface. Gracias et al. reported that biaxially compressed nanofolded Au catalysts with better a Faradaic efficiency convert CO₂ into CO more efficiently than planar catalysts [101]. Their tight folded morphology shortens the mass transport of electrolyte matter, producing an increase in the local pH. The high CO selectivity is ascribed to a low HER activity.

3.4. Support Effect of Au Catalysts

To the best of our knowledge, it is necessary to load Au nanoparticles on conductive materials in order to disperse the active nanoparticles to achieve superior performance [102,103]. The support materials are also used to stabilize nanoparticles for preventing uncontrollable aggregation and for forming active interfaces [104,105]. Large defects such as grain boundaries are thought to effectively enhance electrocatalytic CO₂ reduction performance [106]. Kanan et al. used a vapor deposition method to prepare Au nanoparticles loaded on carbon nanotubes (Au/CNT) and showed a linear correlation between CO₂RR activity and Au–Au boundary density [107]. The higher the grain boundary density, the higher the electrochemical CO₂ reduction current density, mass activity and CO Faraday efficiency. In 0.5 M CO₂ saturated NaHCO₃ solution, the maximum mass activity at −0.7 V was 35 A g⁻¹ and the CO Faraday efficiency was 90%. Long term stability tests showed that CO₂RR activity and CO Faraday efficiency remained unchanged during 12-h electrolysis, but the mass activity decreased slightly due to particle sintering. It is well known that the synergy between metal catalysts and their carrier with abundant active sites can not only enhance the catalytic activity but also significantly improve the catalytic stability. Fischer et al. use a bottom-up method to synthesize a composite material consisting of gold nanoparticles embedded in a graphene nanoribbon (GNR) matrix [108]. Electrochemical tests showed that the electronic and structural effects of graphene nanoribbon (GNR) matrix composites boosted the electrochemical active surface area (ECSA) of Au nanoparticles and improved the dispersion of the same. Therefore, the nanocomposite improves the overall catalytic yield (100 times higher than that of carbon-supported Au nanoparticles). The maximum CO Faraday efficiency was more than 90% and the catalyst had a good stability over 24 h.

Recently, a synthetic method was proposed to selectively adjust the H₂ to CO ratio in the range of 0.3–3, which may be suitable for syngas production. Hwang et al. loaded different amounts of Au nanoparticles (6–8 nm) on layered titanate nanosheets to alter the electronic structure of the hybrid materials, which is a key factor regulating the stability of the reaction intermediates, resulting in an adjustable H₂/CO ratio of syngas, as displayed in Figure 8a–c [109].

In recent years, supported Au catalysts have become a research hotspot because of their good synergistic effects. For example, Au-CeO_x exhibits a higher Faraday efficiency than Au or CeO_x alone, as shown in Figure 8d–g [110]. Density function theory calculations show that the Au-CeO_x interface is the active site for CO₂ activation. The synergistic effect between Au and CeO_x is favorable to *COOH intermediate stability, thus promoting CO₂ reduction.

3.5. Ligand Effect of Au Catalysts

The surface modification of metal catalysts by molecular ligands is also a promising means to enhance their electrochemical activity [111], regulating the microenvironment near the metal surface and improving interfacial charge transfer [112]. Liu et al. reported a design principle for tuning heterogeneous nanoparticles using capped chelated organic ligands, as exhibited in Figure 9a,b [113]. Ligands can maintain the exposed metal active sites to the greatest extent by forming an almost bare metal surface, thereby regulating CO₂RR performance and improving its catalytic stability. Relative to Au nanoparticles coated with oleylamine, Au nanoparticles functionalized with tetradentate porphyrin ligands show a 110-fold increase in current density in terms of the electrochemical reduction of CO₂ to CO at an overpotential of 340 mV and a Faraday efficiency of 93%, as displayed in Figure 9c [114]. These catalysts show excellent stability after electrolysis for 72 h. DFT calculations further confirm the chelation for stabilizing the nanoparticle interface and adjusting the catalytic activity.

In order to uncover the effect of a molecular ligand carbon chain structure on the catalytic performance of metal catalysts, Wallace et al. modified Au nanoparticles with linear amines (propylene amine (PA), hexylamine (HA), oleylamine (OLA), ethylenediamine (EDA) and branched polyamines (PEI)), as shown in Figure 9e,f [106]. They showed that linear amines are beneficial to CO production, while branched polyamines hinder CO formation, as illustrated in Figure 9g.

Theoretical calculations show that Au nanoparticles with linear amine are favorable to CO production due to the adsorption of amine on the corner sites with low coordination numbers, which makes it is easy to stabilize *COOH intermediate. However, Au nanoparticles with branched amine easily suppress CO generation because of the high coverage of amine on Au NPs.

Yang et al. found that N-heterocyclic carboenylated Au nanoparticles exhibit a higher Faraday efficiency and CO current densities in CO₂RR than pure Au nanoparticles at an overpotential of 0.46 V [94]. Due to substantial progress in recent decades, sole Au nanocatalysts have been developed as efficient CO₂RR nanocatalysts via the manipulation of the nanoporous structure, nanoparticle size, morphology, support and ligands.

Table 2. Summary of the catalytic performance of representative sole Au-based nanocatalysts for the CO₂RR.

Catalysts	Electrolyte	Major Products and Maximum FE	Stability	Refs
Mesoporous Au	0.1 M KHCO3	CO (75%) at −0.4 VRHE	-	[114]
Nanoporous Au film	0.05 M K2CO3	CO (80%) at −0.5 VRHE	110 h at −0.5 VRHE	[21]
Porous Au film	0.1 M KHCO3	CO (90.5%) at −0.5 VRHE	4% decrease after 9 h at −0.5 VRHE	[83]
Au25 clusters	0.1 M KHCO3	CO (>95%) at −1.0 VRHE	-	[89]
Monodispersed Au (8 nm)	0.5 M KHCO3	CO (90%) at −0.67 VRHE	-	[92]
Compressed nanofolded Au	0.5 M KHCO3	CO (~98%) at −0.4 VRHE	-	[101]
Au NWs (2 nm)	0.5 M KHCO3	CO (~94%) at −0.35 VRHE	6 h at −0.35 VRHE	[99]
Concave rhombic dodecahedral Au nanocatalyst	0.5 M KHCO3	CO (93%) at −0.57 VRHE	-	[97]
Au/CeOx-47.3	0.1 M KHCO3	CO (90.1%) at −0.5 VRHE	-	[110]
Au/CNT	0.5 M NaHCO3	CO (94%) at −0.5 VRHE	12 h at −0.5 VRHE	[107]
Au/C3N4	0.5 M KHCO3	CO (90%) at −0.45 VRHE	15 h at −0.7 VRHE	[115]
Au/TiNS	0.10 M KOH containing 1 mM Pb(OAc)2	CO (80%) at −0.65 VRHE	-	[109]
GNR-Au NPs	0.5 M KHCO3	CO (88%) at −0.47 VRHE	24 h at −0.47 VRHE	[108]
PVA modified Au NPs	0.5 M KHCO3	CO (95%) at −0.55 VRHE	24 h	[116]
Oleylamine modified rGO-Au NPs	0.1 M KHCO3	CO (70%) at −0.65 VRHE	10 h	[105]
NHC modified Au NPs	0.1 M KHCO3	CO (80%) at −0.9 VRHE	11 h	[117]
N-heterocycliccarbene- functionalized Au NPs	0.1 M KHCO3	CO (83%) at −0.57 VRHE	-	[94]

systematically summarizes the catalytic activity of representative single Au nanocatalysts for the CO₂RR.

3.6. Alloy Effect of Au Catalysts

In the Au-based material system, alloying by the addition of foreign atoms cannot only reduce the content of precious Au metal but also regulate the geometric and electronic structure of the parent nanostructures [118]. So far, the alloy regulation strategy for Au-based catalysts has made some progress. In contrast to sole Au nanostructures, the Au-based alloy catalysts show superior electrocatalytic activity and stability for the CO₂RR [119].

On one hand, most Au-based nanoalloy catalysts mainly produce C1 products, especially CO. Yang et al. synthesized AuCu alloy nanoparticles with an average size of 10–11 nm using Au(CH₃COO)₂ and Cu(CH₃COO)₂ as precursors and octadecene as a reducing agent, with them being composed of Cu, AuCu₃, AuCu, Au₃Cu and Au, respectively [120]. With the increase in Cu content, more products were obtained for Au-Cu alloy nanocatalysts. At −0.73 V versus RHE, the Au₃Cu alloy nanoparticles exhibited an excellent mass activity of 200 A g⁻¹ and a maximum CO Faradaic efficiency. The researchers suggest that the preference of CO formation on AuCu alloy nanoparticles is ascribed to the synergistic, electronic and geometric effects stabilizing *COOH intermediates better than *CO, favoring CO formation. In another study, Smith et al. prepared Au-Pt bimetallic films with different molar ratios for electrocatalytic CO₂ reduction, revealing the relationship between the formation of the main product, CO/HCOOH, and catalyst composition [121]. When it comes to pure gold, the maximum Faraday efficiency of CO is 77% at −0.6 V. With an increase in Pt content, the Faraday efficiency of CO decreases. In order to further understand the relationship between alloy components and CO₂RR performance, Faraday efficiency and CO partial current density at different potentials were studied. It can be seen that with the increase in Au content, the Faraday efficiency of CO increases and the initial potential is more positive, indicating that the composition of Au-based nanoalloy has a large influence on its CO₂RR performance. Specifically, as the Au content increases, the d band center deviates from the Fermi level, thereby reducing the binding energy of *COOH and *CO and significantly improving the CO₂RR performance. In another study, Janaky et al. successfully prepared Au-Sn nanoparticles as an electrocatalyst for the CO₂RR, and the product only consisted of HCOOH, CO and H₂, as can be seen from Figure 10a [122]. When the Au/Sn atomic ratio reaches 1/2, the Au-Sn nanoalloy has the lowest overpotential and the best stability. In particular, in terms of long-term stability, a current density of 8 mA cm⁻² is achieved at −1.0 V versus RHE for 10 h, as shown in Figure 10b,c.

On the other hand, the combination of other metals with Au can also promote the formation of C2+ products in the CO₂RR. As a representative study for Au-based nanoalloys, Koper et al. reported that the Pd-Au alloy can electrochemically reduce CO₂ to C1–C5 and other hydrocarbons via the electrochemical deposition method [123]. It is well known that *CO is a key reaction intermediate in the formation of hydrocarbons. At the same time, it is easier to reduce *CO to hydrocarbons than to reduce CO₂ to hydrocarbons. By alloying Au (weakly bound to *CO) and Pd (strongly bound to *CO), precise control over the binding strength of *CO intermediate on the catalyst surface in the CO₂RR process is realized. The CO₂ reduction products of Pd-Au alloy through an online electrochemical mass spectrometry (OLEMS) test include hydrogen, C1 (CH₄, HCOOH and methanol), C2 (C₂H₄, C₂H₆, ethanol and acetic acid), C3 (C₃H₆ and C₃H₈), C4 (1-butane, isobutane and butane) and C5 (2-methylbutane, pentane and pentene). This work shows that by precisely regulating the binding energy of *CO intermediate on the metal nanocatalyst surface, higher hydrocarbons (for example, C3 to C5) can be obtained from CO₂ reduction. Chorkendorffs et al. used AuCd alloy to tune the adsorption energy of the intermediate and alter the reaction path [124]. It is worthwhile emphasizing that surface reconstruction should be taken into account for bimetallic catalysts. Although there are few studies on this subject, there is no doubt that doping two or more elements can also better modulate the reaction path.

Figure 10. (a) TEM image of Au₁Sn₂ catalysts. (b) FE as a function of composition at −1.0 V vs. RHE. (c) Stability at various potentials. Reproduced with permission from Ref. [122]. Copyright 2018, American Chemical Society. (d) TEM image of MDA NP. (e) Faradaic efficiencies of CO produced by MDA NP and Au NP. (f) CO Current densities of MDA NP. Reprinted with permission from Ref. [125]. Copyright 2020, American Chemical Society.

Figure 10. (a) TEM image of Au₁Sn₂ catalysts. (b) FE as a function of composition at −1.0 V vs. RHE. (c) Stability at various potentials. Reproduced with permission from Ref. [122]. Copyright 2018, American Chemical Society. (d) TEM image of MDA NP. (e) Faradaic efficiencies of CO produced by MDA NP and Au NP. (f) CO Current densities of MDA NP. Reprinted with permission from Ref. [125]. Copyright 2020, American Chemical Society.

Wang et al. synthesized Mo-doped Au nanoparticles by using a solvothermal method for the electrochemical CO₂RR, as displayed in Figure 10d [125]. By introducing Mo into Au, charge transfer occurs between the two elements and electron-rich Au is formed near Mo atoms. A CO Faraday efficiency of 97.5% at −0.4 V was obtained for Au-Mo alloy particles, as illustrated in Figure 10e. The partial current density of Au-Mo alloy particles is 75 times higher relative to pure Au nanoparticles and can last for 50 h, as shown in Figure 10f. Theoretical calculations showed that Mo can lower the reaction energy barrier through extra Mo-O bonds and help stabilize *COOH intermediate. These results indicate that the synergy between electronic and geometric effects can improve the CO₂ reduction performance of the catalyst, thus providing a clue to the basic mechanism of the alloy electrocatalyst.

3.7. Core–Shell Effect of Au Catalysts

In recent decades, the reasonable design and controllable preparation of core–shell nanoparticles with different sizes have become frontier research topics in the field of nanoscience [87,126]. In these nanostructures, the core–shell interaction can significantly affect the surface electronic structure and alter the adsorption energy of reaction intermediates, thereby achieving better catalytic activity and improving the utilization efficiency of Au to a greater extent [127,128]. In the electrochemical CO₂ reduction reaction, the core–shell structure can be formed by dispersing Au active components on the surface of other metal nanoparticles. Gong et al. synthesized core–shell Pd-Au catalysts using seed growth methods, as shown in Figure 11a,b [129]. Using the galvanic replacement reaction, the standard reduction potential of gold ions is higher than that of palladium ions; gold ions can be easily reduced with the oxidation of palladium atoms to form core–shell Pd-Au with Pd as the core and PdAu as the shell. Pd@PdAu catalysts with different thicknesses in terms of the PdAu shells were obtained by changing the feeding ratio of metal precursors.

Pd@Pd₃Au₇ nanocrystals exhibit superior CO₂RR performance, with a CO Faraday efficiency of 94% at −0.5 V versus RHE and a CO Faraday efficiency of nearly 100% at −0.6 V to −0.9 V, as displayed in Figure 11c,d. At −0.7 V, the Pd@Pd₃Au₇ catalyst can stably operate for 8 h without Faraday efficiency attenuation. Such enhanced electrocatalytic performance is ascribed to the ligand effect and the electronic effect between the Pd core and PdAu shell. By optimizing the binding energy of *COOH and *CO intermediates, selectively reducing CO₂ to CO can be achieved.

Wang et al. first used theoretical calculations to establish a screening principle based on the binding energy of *CO and *COOH intermediates and selected Fe as the element to prepare core–shell catalysts [130]. The calculations show that Au-Fe core–shell catalysts with AuFe alloy as the core and Au as the shell have lower initial reaction overpotential and CO desorption energy, as exhibited in Figure 11e. Then, Au-Fe nanocatalysts were prepared by solvothermal synthesis. The Faraday efficiency of this catalyst to convert CO₂ to CO at −0.4 V can be as high as 97.6%, suggesting that the hydrogen evolution reaction is almost completely inhibited, as can be seen from Figure 11f. The mass activity (48.2 mA mg⁻¹) is 100 times higher than that of Au nanoparticles (0.5 mA mg⁻¹), as illustrated in Figure 11g. The nanocatalyst exhibits high stability for 90 h with almost no attenuation in terms of current density.

Li et al. fabricated core–shell Ag@Au nanowires with nano-thick Au film as the shell and Ag nanowires as the core via a galvanic-free method as electrocatalysts for CO₂ reduction [127]. At −1.0 V, Ag@Au nanowires with a relative Au content displayed a CO Faradaic efficiency of nearly 100% in 0.1 M KCl. The enhanced catalytic performance was mainly ascribed to the synergistic effects between Au shell and Ag core and the strong interaction between CO₂ and Cl ions in electrolyte. Up until now, Au-based multiple metal nanocatalysts acquired via the manipulation of the alloy and core–shell structure have also been exploited as outstanding CO₂RR nanocatalysts.

Table 3. Summary of the catalytic performance of representative Au-based nanocatalysts for the CO₂RR.

Catalysts	Electrolyte	Major Products and Maximum FE	Stability	Refs
Ordered Au Cu NPs	0.1 M KHCO3	CO (~80%) at −0.77 VRHE	12 h at −0.76 VRHE	[131]
Pd5@Au95 NPs	0.1 M KHCO3	CO (~80%) at −0.5 VRHE	-	[132]
Twisted Pd0.8Au nanowires	0.5 M KHCO3	CO (94.3%) at −0.6 VRHE	15.7% decrease after 8 h at −0.6 VRHE	[133]
Au3Cu nanocubes	0.5 M KHCO3	CO (90.2%) at −0.38 VRHE	30 h at −0.38 VRHE	[134]
Au75Pd25 NPs	0.1 M KHCO3	CO (~80%) at −0.5 VRHE	-	[20]
Au1Ni1-CNFs	0.1 M KHCO3	CO (92%) at −0.98 VRHE	16 h at −0.98 VRHE	[119]
AuFe-CSNP	0.5 M KHCO3	CO (97.6%) at −0.4 VRHE	90 h at −0.5 VRHE	[130]
Pd@Pd3Au7 nanocubes	0.1 M KHCO3	CO (94%) at −0.5 VRHE	8 h at −0.7 VRHE	[129]
Au94Pd6 NPs	0.1 M HClO4	CO (94.7%) at −0.6 VRHE	12 h at −0.6 VRHE	[135]
Pd-Au electrode	0.1 M KH2PO4/0.1 M K2HPO4	CO (30.9%) at −0.6 VRHE	-	[123]
Mo-doped Au nanoparticles	0.5 M KHCO3	CO (97.5%) at −0.4 VRHE	50 h at −0.4 VRHE	[125]
Au-bipy-Cu	0.1 M KHCO3	CO (90.6%) at −0.9 VRHE	-	[136]
AgPd-edged Au nanoprisms	0.1 M HClO4	HCOOH (49%) at −0.18 VRHE	12 h at −0.18 VRHE	[137]
Au1Sn2 NPs	0.1 M NaHCO3	HCOOH (42%) at −1.1 VRHE	Slight increase after 10 h at −1.0 VRHE	[122]
Cu63.9Au36.1/NCF	0.5 M KHCO3	CH3OH (15.9%) and C2H5OH (12%) at −1.1 VSCE	-	[138]
Cu3Au nanowire arrays	0.1 M KHCO3	C2H5OH (48%) at −0.5 VRHE	14% decrease after 8 h at −0.7 VRHE	[139]
AuCu/Cu-SCA	0.5 M KHCO3	C2H5OH (29 ± 4%) and C2H4 (16 ± 4%) at −1.0 VRHE	24 h at −1.0 VRHE	[19]
AuAgPtPdCu NPs	0.5 M K2SO4	CH4 (38.0%) and C2H4 (29.5%) at −0.9 V Ag/AgCl	Steady decrease after 1 h at −0.8 VAg/AgCl	[140]

CNFs, carbon nanofibers; CSNP, core–shell nanoparticle; NCF, nanoporous Cu film; SCE, saturated calomel electrode; SCA, submicron arrays.

summarizes the catalytic activity of representative Au-based multiple metal nanocatalysts for the CO₂RR in detail.

4. Conclusions and Perspectives

For the foreseeable future, fossil fuels will continue to provide the majority of our energy. Electrochemical CO₂ reduction to high value-added chemicals and useful fuels and the use of chemical bonds to store renewable energy provide new strategies for achieving carbon neutrality in human society. Based on its technical and economic feasibility, the electrochemical conversion of CO₂ to CO is regarded to be the most commercially available CO₂ reduction pathway. At present, progress has been made concerning Au-based nanocatalysts for electrochemical CO₂ reduction. However, metal Au is expensive and its reserves are limited, greatly restricting its further application. As a result, it is of great importance to actively develop Au-based catalysts with cost-effectiveness and high performance to commercialize CO₂ reduction technology. At present, for CO generation, a number of gold-based nanocatalysts have reached the expected Faraday efficiency and selectivity, and big steps have been made towards practical application.

Recent experiments and calculations using Au-based nanostructured catalysts for the electrochemical reduction of carbon dioxide were reviewed. It was found that the porous structure, size, morphology, composition, support and capping ligand all affect catalytic activity and product selectivity [141]. On one hand, by reducing the size of Au nanocatalysts or designing nanoporous structures with abundant surfaces, the utilization of Au atoms and the intrinsic catalytic performance of Au can be increased, thus reducing Au loading [142]. On the other hand, constructing Au or Au alloy nanoparticles with controllable surfaces is an appealing way to improve the catalytic performance and selectivity of Au-based nanocatalysts for CO₂ reduction, such as precisely adjusting the exposed crystal faces and surface electronic structure [143,144].

Multi-metal systems with synergistic effects among different metals can be optimized in terms of the CO Faraday efficiency and partial current density. In addition, the catalyst carrier has a significant impact on the performance and durability of the catalyst, which is key in the context of improving the industrial feasibility of the technology. From the viewpoint of the economy, industrial applications usually require lower synthetic costs. Therefore, the development of simple synthetic strategies to accurately prepare Au-based nanocrystals has been a hot topic. Electrochemical CO₂ reduction reactions on Au-based nanocatalysts involve complex chemical reaction processes. The types and selectivity of the target products are determined by the binding energy of the nanocatalyst on the reaction intermediates (such as *CO, *OCHO, *COOH, *H, etc.).

In-situ spectroscopic characterization can offer clear information about the structure and surface state of the Au-based nanocatalyst as well as the binding configurations of the adsorbed intermediates on the Au-based nanocatalyst’s surface during the reaction. Therefore, in-situ infrared spectroscopy, in-situ Raman spectroscopy, in-situ XRD, in-situ TEM and in-situ synchrotron radiation are effectively combined to track the evolution processes of intermediate species adsorbed on the Au surface during CO₂ reduction and to investigate the structural evolution and chemical state of Au. It is helpful to further understand the mechanism of the Au-based catalyst for CO₂ reduction. The development of high sensitivity and high resolution in-situ characterization techniques and advanced density functional theory models provides important guidelines for precisely preparing robust Au-based nanocatalysts for CO₂ reduction.

The catalyst is the core of catalytic reactions. In the future, we should pay more attention to the stability of catalysts. Electrolyte contamination, morphological transformation, catalyst degradation and CO poisoning may all cause catalyst deactivation. To commercialize these catalysts, it is imperative to achieve long term stability under the conditions of the reaction rate required by the business. At present, the application of CO₂ electrolytic cells is still at an early stage in terms of the design of commercial technology for CO₂ conversion. Although many works have proved that we can achieve high CO₂ reduction current densities (>100 mA cm⁻²), the stability of these systems (<1000 h) remains an urgent problem to be solved. The structures of electrolytic cells, gas diffusion electrodes, electrolytes and ion exchange membranes greatly change the overall performance of the CO₂RR. However, current research only focuses on the design of the catalyst, which deserves our attention.

From what have been reported until now, we can conclude that there remains significant room for enhancement in terms of Au-based catalysis for the CO₂RR. Currently, the accumulation in Au-based composite nanomaterials creates opportunities as well as challenges when it comes to applying these materials in electrochemical CO₂ reduction reactions. In particular, the as-synthesized Au-based nanocomposites may possess superior performance to promote electrochemical CO₂ reduction reaction for good selectivity due to their electronic coupling effect and the lateral strain effect for different components. Accordingly, the application of Au-based nanocomposites in electrochemical CO₂ reduction reactions is a hot topic. For example, there have been many reports in the literature that semiconductors (metal sulfides, etc.) exhibit excellent performance in electrocatalytic carbon dioxide reductions. Therefore, the heterogeneous nanocatalysts that combine precious metal gold and semiconductors may show extraordinary activity in electrocatalytic carbon dioxide reductions. In particular, optimizing the composition and structure of composite materials is an effective way to further improve electrocatalytic performance.

In addition, in order to develop robust and relatively cheap catalysts and facilitate the industrialization of CO₂ electroreduction, other influencing factors should be optimized. Taking into account the low solubility of CO₂ in aqueous solutions and the characteristics of heterogeneous catalytic reactions, flow cells equipped with gas diffusion electrodes can be adopted to solve the above problems. Meanwhile, ionic liquids and solid polymer electrolytes have emerged as new appealing electrolytes to suppress the competitive hydrogen evolution reaction. Finally, one vital problem that needs to be accounted for is the requirement to make the syntheses of Au-based catalysts ecofriendly for scaling-up production. Therefore, it is necessary to use cost-effective raw materials and simple synthesis procedures.

In conclusion, recent studies have shown that it is possible to use Au-based electrocatalysts for the commercial application of CO₂ electroreduction. In spite of facing considerable challenges, it is generally accepted that, year by year, in-depth research in this field will encourage the development of robust and economically feasible Au-based electrocatalysts, thus realizing carbon neutrality in the near future.