Recent Progress in Electrocatalytic Conversion of CO2 to Valuable C2 Products

Converting CO2 into high value‐added chemicals by renewable energy directly alleviates the problem of resource shortages, while also reduces the amount of CO2 in the atmosphere. Among various CO2 reduction methods, CO2 electroreduction has attracted much attention because of its moderate operating conditions and a green and easy‐to‐scale strategy. In this work, the four catalyst enhancement mechanisms: synergistic effect, strain effect, ligand effect, and defect effect are briefly described. Then, the catalytic mechanisms of the C2 products are fully expounded, depending on the number of transferred electrons during the CO2 reduction process. In addition, the membrane electrode is also introduced from the three aspects of ion exchange membrane, catalyst layer, and gas diffusion layer. Finally, the developing direction of CO2 reduction reaction is proposed. This work provides an essential guide for design and preparation of CO2 electroreduction catalysts and fully summarized the intrinsic catalysis mechanism, which is beneficial to the development of CO2 conversion in the future.


DOI: 10.1002/admi.202300186
the most popular methods for converting CO 2 . [2] Remarkably, compared with photocatalytic and thermochemical catalytic conversion methods, electrocatalytic reduction of CO 2 shows the advantages of green, easy-to-scale device, and moderate operation conditions. Furthermore, the voltage required in the conversion process can be obtained by clean energy such as wind energy, solar energy, and so on, and the electrolyte can be recycled. In addition, different products can be obtained by changing the catalyst type and additional voltage. [3] In the past few decades, the CO 2 reduction electrode in the aqueous solution or ionic liquidbased solution were always fabricated by metals, metal oxides, and carbon-based materials. The electrodes prepared with the catalysts directly determine the type and efficiency of the CO 2 conversion product through its morphological characteristics and conductivity. However, catalysts may take place all kinds of variety during reaction such as reduction of active sites, structural changes, and so on. The development of efficient and stable CO 2 electroreduction catalysts has attracted wide attention. [4] Thus, it is necessary to conclude the research progress of CO 2 electroreduction catalysts to guide obtain more achievements.
Among CO 2 electroreduction catalysts, the studies of metalbased materials are extremely in-depth, including transition metals (such as iron, cobalt, nickel, palladium, platinum, copper, silver, and gold) and major metal groups (such as aluminum, indium, thallium, tin, lead), alkali metals, and alkaline earth metals. [5] In addition to metal-based catalysts including metal complexes and metal-organic frameworks, as well as nonmetallic catalysts, are also used in CO 2 electrocatalysis. [6] The electrochemical properties of the catalysts mainly depend on the activity of the electrode material, the availability of the active sites, and the conductivity of the electrode. The active sites of the catalysts could be obtained by adjusting the surface structure of the electrode material, while the activity and stability of the catalysts depend on the chemical properties of the electrode material. So far, researchers have made a lot of efforts to develop low-cost catalysts with high activity, high selectivity, and stability, but most catalysts are still not suitable for practical applications. Therefore, the development of new catalysts with larger specific surface area, higher catalytic activity, and better selectivity are still the key to improving the CO 2 reduction efficiency.
In this work, to promote the development of CO 2 reduction, we reviewed the current researches about electrocatalytic CO 2 www.advancedsciencenews.com www.advmatinterfaces.de reduction. First, four catalyst enhancement mechanisms including synergistic effect, strain effect, ligand effect, and defect effect are briefly introduced. Then, based on the number of transferred electrons in the CO 2 electrocatalytic process, the research progress and application of CO 2 electroreduction to prepare oxalate, acetic acid, ethylene, ethanol, and ethane were enumerated. Finally, CO 2 electroreduction reaction device, membrane electrode, was briefly introduced from the three aspects of ion exchange membrane, catalyst layer, and gas diffusion layer. The advantages and disadvantages of these three components of membrane electrode were elaborated. At the end of this article, we put forward a prospect for the CO 2 electroreduction research based on the present work, including the current electroreduction process with high overpotential, low product selectivity, and other problems. We believed that the deep understanding of the reaction mechanism, new insights into catalysts and electrolytes, and new technologies of electrolytic cells could contribute to the wide application of CO 2 electroreduction in the future.

Synergistic Effect
Synergistic effect existing in two components originates from regulating geometric and electronic configurations of active sites, and produces new binding sites, patterns, and different binding strengths with intermediates. [7] Combining materials of different components can combine their physical properties while sometimes they will improve or even produce new properties because of the synergy between atoms. [8] Synergistic effect between metal atoms has the ability to alter the performance of catalysts by changing the adsorption affinity of metal atoms. [9] Jakub et al. found that Ir atoms in Ir 1 /Fe 3 O 4 are more firmly bound to CO compared with the metal Ir by atomic resolution scanning probe microscope, X-ray photoelectron spectroscopy, and other methods. [10] Except for this, improvement of reaction activity also attributes the success to synergistic effects brought by adjusting geometric structure, changing electronic configuration, changing the center of d-band, and other means. [11] Ma et al. obtained that the introduction of additional atoms can lead to the adjustment of the d orbital and the reaction diversity of the electrocatalyst through DFT calculation, and found that FeCo@ and NiCo@GDY are applicable to NRR with HER inhibition capability. [12] Bai et al. believed that asymmetry between Cu 0 and Cu x+ promoted the C-C coupling of the two *CHO species, leading to the conversion of CO 2 into ethanol. Furthermore, the asymmetry was caused by the charge transfer and orbital hybridization between the substrate and Cu atoms. [13] In the process of CO 2 electroreduction, the development of catalysts is subject to high thermal mechanical reaction energy barrier and slow kinetic reaction rate. Therefore, it is particularly important to solve the above problems through the synergy between different atoms (Figure 1).
Synergistic effect between different kinds of non-metallic materials can effectively promote the electroreduction of CO 2 into specific products. Sun's team used a boron phosphide nanoparticle as a CO 2 non-metallic electrocatalyst (Figure 2A-C), which showed methanol Faradaic efficiency of 92.0%. Density functional theory (DFT) calculations displayed that B sites have intense adsorption on CO 2 molecules, and then the electrons contributed by P in BP to B are beneficial to the activation of CO 2 . In the BP(111) crystal plane, the synergy between P and B plays a crucial role in triggering the electrocatalysis of CO 2 . What is more, the *CO and CH 2 O desorption energy barriers are high, which is one of the important factors that CO 2 can be highly selectively reduced to methanol on the BP catalyst. [14] In non-metallic catalysts, nitrogen (N) doped carbon-based materials are always used for CO 2 RR. Dai et al. conducted indepth research on carbon-based metal-free catalysts. They incorporated polyethylenimine (PEI) into nitrogen-doped carbon materials to regulate electrons, as shown in Figure 2D. [15] PEI stabilizes CO 2 •− through H bond which is significant to reduce the Figure 1. Schematic illustration of electrochemical CO 2 RR powered by renewable energy sources such as wind and solar. In this way, high concentration of CO 2 released by human activities can be recycled back and converted to value-added products. Reproduced with permission. [14] Copyright 2019, Wiley-VCH. D) Illustration of the preparation procedure of the PEI functionalized N-doped CNTs. Reproduced with permission. [16] Copyright 2018, Wiley-VCH. E) SEM image of NPCAs. F) TEM and HRTEM images of NPCAs. G) Illustration of the preparation procedure for NPCAs. Reproduced with permission. [19] Copyright 2020, Wiley-VCH.
starting potential. Furthermore, it also enriches CO 2 from solution on electrode at the same time, thereby synergistically providing an effective local environment for the electroreduction of CO 2 to CO 2 •− with N-doped carbon materials. CO 2 molecules with a linear structure require a large amount of energy to produce free radicals. In addition, the thermodynamic energy required for competitive reactions is tightly distributed between them, leading to poor chemical selectivity. Therefore, it is important to create active sites to form CO 2 •− intermediates, while reducing hydrogen production. [16] The N-doped nanodiamond/Si rod array(NDD/Si RA) catalyst prepared by Yu et al. showed excellent CO 2 RR performance. The synergistic effect between carbon and nitrogen derived from the polarization of atoms, which can promote the adsorption of CO 2 and CO 2 •− . This effect could reduce the energy barrier of electrochemical reactions by promoting the adsorption of CO 2 and CO 2 •− . The catalyst has a low starting potential and can quickly convert CO 2 into acetate, improving the selectivity of C 2 products. In addition, the Faraday efficiency can reach 91.2-91.8% when the potential is between −0.8 V (vs RHE) and −1.0 V (vs RHE). [17] The non-metallic nitrogen doped mesoporous carbon catalyst prepared by Sun et al. could efficiently produce ethanol with nearly 100% selectivity at −0.56 V (vs RHE) and a Faraday efficiency of up to 77%. Electrochemical impedance analysis and DFT calculation results confirmed that the design of cylindrical pore structure and the existence of pyridine/pyrrole nitrogen atomic structure make the catalyst have a high electron density surface, which plays an important role in promoting the coupling of *CO, the intermediate product of CO 2 electroreduc-tion, and the generation of ethanol. [18] To further improve the performance of non-metallic catalysts, heteroatom can be doped in non-metallic catalysts. For example, Chen et al. prepared N and P co-doped carbon aerogel electrocatalysts (NPCAs) for CO 2 electroreduction by gelatinization of starch ( Figure 2E-G). Their study showed that P can effectively reduce the occurrence of hydrogen evolution reaction while N increasing the activity of CO 2 electrocatalysis. Through synergy of them, the reaction efficiency can be significantly improved. At the same time NPCAs has a higher electrochemical activity area and total electron conductivity than N-doped carbon aerogels, which facilitates the transfer of electrons from CO 2 to its free radical anions or other key intermediates. [19]

Strain Effect
Strain effects are typically generated through lattice mismatches or structural defects. [20] The strain effect manipulates the electronic structure by adjusting the adsorption strength of the dband center and the intermediate, thus affecting the catalytic performance, which has received extensive attention. The catalytic reactions mainly rely on individual atoms, especially unsaturated sites, such as angle atoms and edge atoms. Thus, singleatom catalysts (SACs), containing many distinctive metal centers, have been widely used to boost the reaction kinetics of CO 2 conversion. [21] SACs are the smallest heterogeneous catalysts with no binding between dispersed metal atoms and an Figure 3. A) Reduction of CO 2 on a single-atom catalysts. Reproduced with permission. [24] Copyright 2019, American Chemical Society. B) Reduction of CO 2 on an indium single-atom catalyst. C) Illustration of In-N-C synthesis. Reproduced with permission. [26] Copyright 2021, American Chemical Society. D) XRD patterns of Pd-NC and NC. E) HAADF-STEM image and F) elemental mapping of Pd-NC. Reproduced with permission. [27] Copyright 2020, Wiley-VCH.
atomic utilization rate of 100%. In the heterogeneous catalysis system, the performance of the catalyst is often affected by the strain effect. Moreover, the strain effect between single-atom sites and support atoms could be changed by regulating the local coordination structure and pore size of the support materials, effectively improving the excellent catalytic activity, selectivity, and stability. [22] Wei et al. discovered the atomization effect of precious metals (such as Pd, Pt, Au), which could be redistributed by sintering the noble metal nanocatalyst to redisperse the single atom position, providing a valuable reference for highperformance heat-stable catalysts. [23] Carbon materials, oxides, and nitrides are commonly used as carriers in single-atom catalysts. [22] Nitrogen doping in carbonaceous materials can form M-N bonds to improve the stability of SACs. For instance, Qin et al. obtained a Fe-N-C catalyst containing uniformly dispersed iron single atoms using a ZIF-8 molecular sieve as a template and used it in CO 2 RR. Their study showed that the Fe centers in the graphite layer were poisoned by adsorbed *CO, thus they were not the active sites of CO 2 conversion. Strain effect originated from the defective structure of the nanoporous graphite layer balanced the binding energy of *COOH and *CO. Therefore, the reaction could proceed at a lower overpotential, while also alleviate the toxicity of *CO intermediates to Fe centers ( Figure 3A). [24] Liao et al. prepared a stable and conductive 2D phthalocyanine covalent organic framework and used it as an electrocatalytic catalyst for CO 2 reduction to acetic acid. At −0.8 V, the Faraday efficiency of the single product is 90.3(2)% and the current density is 12.5 mA cm −2 . Phthalocyanine group has more obvious conjugation effect and more nitrogen atoms, which can effectively increase the electron density of coordination metal atoms and promote d electron transfer. The isolated active site of copper phthalocyanine with high electron density is conducive to the generation of *CH 3 . *CH 3 and CO 2 generate acetic acid, avoiding coupling with *CO or *CHO to generate ethylene and ethanol. [25] Lu et al. prepared an indium single-atom catalyst (In-N-C) by pyrolysis of In/Zn imidazolic acid molecular sieve framework for CO 2 electroreduction ( Figure 3C). The strain effect between the dispersed In atoms and the N atoms effectively boosts the electron interaction on the carbon skeleton and reduces the formation barrier of *OCHO intermediates to facilitate CO 2 conversion ( Figure 3B). [26] Compared with conventional indium metal catalysts, the In-N structure exhibited higher electrocatalytic activity and selectivity. Besides, Chen et al. mixed glucose and dicyanamide with palladium salt solution, then evaporated and annealed to obtain a Pd single-atom catalyst with Pd-N 4 sites, which exhibited high catalytic activity for CO 2 electroreduction. The Pd-N 4 sites were the most likely to produce CO without forming palladium hydride. Moreover, the well-dispersed Pd-N 4 single atom active site stabilized the adsorbed CO 2 intermediates, thereby enhancing the CO 2 RR capacity at low overpotential ( Figure 3D-F). [27] As a bridge connecting homogeneous and heterogeneous catalysts, the catalytic performance of SACs is mainly affected by two aspects. On the one hand, the catalytic performance is susceptible to the influence of local coordination environments such as coordination atoms and coordination numbers, which is the characteristic of homogeneous catalysis. On the other hand, a support is required to stabilize the single atom, which significantly influences the position of the single atom and the surrounding geometric environment. [22] In addition to focusing on the development of the support and the dispersion of metal atoms on the support, [28] it is also necessary to optimize the preparation method and understand strain effect between single atom and support.

Ligand Effect
Ligand effect is generated by introducing a second or third metal into metal nanocrystals, leading to an increase or decrease in the center of the d-band. [29] The ligand effect can improve the reaction kinetics by adjusting the binding energy of the intermediate on the catalyst surface. [30] Copper-based catalysts have shown unique advantages in CO 2 RR. Due to the strong interaction between Cu active centers and *CO intermediates, Cu catalysts exhibit highly selective electrocatalytic performance for the reduction of CO 2 to hydrocarbons. Zhang et al. used metalorganic complexes with a well-defined structure as electrocatalysts to reveal the previously neglected effect of cuprophilic interaction on the reduction of electrocatalytic CO 2 to specific hydrocarbons (CH 4 ) for the first time. [31] The study showed that the presence of intramolecular cuprophilic interaction could effectively reduce the free energy of determination steps in the CO 2 -CH 4 conversion process, thereby promoting the generation of CH 4 . Their work provided a new design idea for the synthesis of more efficient Cu-based catalysts in the future. Scientists also studied the activity and selectivity of the catalytic products from many factors such as the surface morphology, internal structure, particle size, and valence state of the copper-based catalyst. For example, Gao et al. mimicked the unique hierarchy of the hy-drophobic leaves to prepare copper electrodes with multi-level high curvature structures on the gas diffusion layer (GDL). The electrode had favorable hydrophobicity and gas affinity, which effectively captured CO 2 molecules and enriched alkali metal cations in the catalytic process. Moreover, it also constructed a stable gas-solid-liquid three-phase interface, alleviated the overflow of electrolyte under large current density and achieved stable catalytic CO 2 conversion for a long time. [32] Besides, adding an internal structure with a large specific surface area and organic framework to the copper-based catalyst also improves the structure of the catalysts. For example, Zhang et al. synthesized a copper-based conductive metal-organic backbone electrocatalyst (Cu-DBC) by solvothermal method at 85°C (Figure 4B,C). The highly conjugated organic ligand imparted the unique redox properties and conductivity of Cu-DBC. The uniformly distributed Cu-O 4 sites were beneficial to prepare CH 4 , which was one of the best Cu-based electrocatalysts for CO 2 reduction to CH 4 . [33] In addition to copper-based catalysts, metals such as bismuth, nickel, indium, and cobalt are also widely used in CO 2 electroreduction. Zhang et al. synthesized the catalyst through the thermal decomposition of the bismuth-funded organic backbone (Bi-MOF) and dicyanamide (DCD). The catalyst obtained unique Bi-N 4 sites on the porous carbon grid ( Figure 4A), which exhibited high reductivity during CO 2 RR. [34] Zhao et al. impregnated a series of ionic liquids with high CO 2 solubility into the pores of the columnar nickel-nitrogen-carbon (Ni-N-C) catalyst to alter the interaction of the CO 2 -Ni sites and formed a solid/liquid interface with high CO 2 concentrations ( Figure 4D). The best Ni-N-C/[Bmim][PF 6 ] composites performed better than Ni-N-C catalysts with their maximum FE CO 1.5 times higher than Ni-N-C catalysts and keeping stable within 30 h. Their study demonstrated the great potential of Ni-N-C/ILs composites in future industrial CO 2 RR applications. [35] The ligand effect generated between different metal atoms changes the surface adsorption energy and activation energy of the catalyst. Therefore, alloy catalysts fabricated with other metals on the basis of the metal element catalyst effectively optimize the selectivity of the product. Wang et al. rapidly prepared Cu based bimetallic catalyst using sodium borohydride reduction precipitation method. The result indicates that the introduction of some metal components can exhibit excellent performance of CO 2 electrocatalysis. It is also found that MgCu 3 (with a molar ratio of 1:3 for Mg and Cu) has considerable selectivity for the production of C 2 H 4 . The corresponding ethylene Faraday efficiency could reach 80%. [36] Hoang et al. prepared copper-silver bimetallic catalysts with different silver content by electrochemical deposition, which had high selectivity for C 2 compounds. [37] What is more, the Cu 3 Ag 1 catalyst prepared by Lv et al. ( Figure 4F) using electrodeposition of Cu matrix was 126-fold higher than that of the only Cu matrix catalyst for CO 2 conversion to ethanol, and the activity was increased by 25-fold. [38] In addition to silver, addition of tin and zinc is supposed to validly improve the selectivity of products. Sun et al. regulated the activity and selectivity of CuO by controlling the content of metallic tin. Compared with undoped CuO, the current density of CuO catalysts containing metal Sn increased by about 2.3 times. And the Faraday efficiency of ethylene is close to 48.5% ± 1.2%, while the maximum value of the original CuO is 26.8% ± 2.2%. [39] In addition, Ren et al. synthesized the CuZn bimetallic catalyst by  6 ] preparation. Reproduced with permission. [35] Copyright 2022, Elsevier. E) SEM image of NiZn alloy. Reproduced with permission. [42] Copyright 2021, Elsevier. F) Schematic diagram of the electrodeposition Cu-based system obtains Cu 3 Ag 1 . Reproduced with permission. [38] Copyright 2020, Wiley-VCH.
atomic layer deposition (ALD) depositing ZnO on CuO, which showed a biased mechanism for the formation of ethanol, and the ethanol/ethylene ratio increased by more than five times. In situ Raman spectroscopy indicated that Zn modified *CO binding site of Cu, where free *CO combined with adsorbed *CH 3 and formed a *COCH 3 intermediate that could only be reduced to ethanol. [40] Moreover, Gong et al. prepared CuZn particles with different Zn doping concentrations. They obtained the surface structure of the catalyst model by combining the neural network potential energy surface and molecular dynamics methods. They found that the two types of active sites, CuZn sites surrounded by stoichiometric Cu and Zn atoms and CuZn sites mainly surrounded by Zn atoms, can effectively promote C-C coupling to improve the selectivity of C 2+ products. These new insights have resolved the debate over the active centers on CuZn catalysts, providing theoretical guidance and research ideas for the development of high selectivity Cu based catalysts based on highthroughput screening in the future. [41] Zhang et al. designed a new Ni-Zn alloy catalyst for CO 2 electrocatalytic production. Unlike the Zn catalyst which mainly produced CO and the Ni catalyst that produced H 2 ( Figure 4E), the Ni-Zn alloy catalyst achieved a FE HCOO-of 36% ± 0.7 at −0.9 V (vs RHE). Ligand effect between Ni and Zn has an intermediate d-band center and valence band, leading to a decrease in H* bonding strength and a favorable pathway for HCOOH formation. [42] Although many researches have been done in the field of CO 2 electroreduction catalysts, a deeper understanding of the reaction mechanism of ligand effects between alloy catalysts is needed. For example, how CO 2 molecules adsorb with the active sites of different alloy catalysts, how to control the pore channels of catalysts to promote the transport of protons and electrons, and how to smoothly desorb the target products to facilitate the separation of various products later. The exploration of the theoretical reaction mechanism of ligand effect in alloy catalysts will be the key to the preparation of high-performance catalysts.
engineering. [45] Among them, defect engineering is the most effective method of improving the intrinsic activity of catalysts at the atomic level. Because the vacancy defect between their anions (such as oxygen, sulfur, nitrogen) and cations (such as iron, titanium, cobalt) can be used as active sites to promote electrocatalysis, and the bonds formed between the elements favor the adsorption of reaction intermediates. In addition, defects could boost the formation of specific intermediates, decrease the overpotential, and elevate the stability of catalysts.
Oxygen vacancy is one of important defect engineering for CO 2 electroreduction because of their low formation energy. For example, Xia et al. successfully synthesized a novel carbon nanorod encapsulated bismuth oxide catalyst (Bi 2 O 3 @C) by carbonized BiBTC nanorods and oxidation treatment in an argon atmosphere ( Figure 5A), which was applied to CO 2 electroreduction. In this catalyst system, Bi 2 O 3 could improve reaction kinetics and formic acid selectivity because the Bi-O structure facilitated the adsorption of OHCO* intermediates and the formation of formate esters. Meanwhile, carbon materials also contributed to increase the activity and current density of formic acid production. [46] Bai et al. synthesized Cu 2 O microcrystals with different morphologies and surface shapes using wet chemical reduction method, namely O-Cu 2 O, D-Cu 2 O, and C-Cu 2 O. The Faraday efficiency of C 2+ products (ethylene and ethanol) on D-Cu 2 O/Cu reached 70%. Their research found that the improve-ment in catalytic performance is mainly due to the presence of oxygen vacancies, the presence of high refractive index surfaces. The high performance of CO 2 RR is related to the surface reconstruction of Cu 2 O. During the CO 2 RR process, Cu 0 and Cu + sites are generated, stabilizing the adsorption of *CO, and enhancing C−C coupling. [47] On this basis, increasing the number of oxygen vacancies could further enhance the reaction rate. Hu et al. improved the oxygen vacancy and electron density of iron by doping alkaline earth elements (Ca, Sr, and Ba) in LaFeO 3 -cathode. Without changing the phase structure, oxygen desorption, and CO 2 adsorption capacity could be enhanced. [48] Besides, Wang et al. developed a Cu-doped CeO 2 electrocatalyst ( Figure 5B-D), where the strong interaction between CeO 2 and Cu results in a high degree of dispersion of Cu to increase oxygen vacancies in adjacent locations. The efficient catalytic point reduced a single CO 2 molecule to CH 4 with an efficiency of 58%. [49] In addition to metal oxides, metal sulfides and nitrides are also used in CO 2 electroreduction catalysts. For instance, Yu et al. synthesized a defect controllable copper sulfide nanocrystal through colloidal nucleation method, and then successfully developed a new core-shell vacancy copper nanocatalyst using in situ electrochemical reduction method. These catalysts enhance the selectivity of liquid alcohols. Compared with the bare copper nanoparticles catalyst, the alcohol/ethylene ratio has been increased by more than six times, and a C 2+ alcohol current density of 126 ± 5 mA cm −2 has been achieved, with a Faraday efficiency of 32 ± 1%. [50] Moreover, TiN@Co 5.47 N catalyst was constructed by plasma nitriding and exquisite atomic layer deposition Co x N process ( Figure 5E). The material had continuous catalytic longterm stability within 1500 h, with the overpotential increased by 1.3%. This was mainly attributed to the cooperative electron interaction between TiN and Co 5.47 N. Meanwhile, the formation of active CoTi layered double hydroxide (CoTi DH) layer at the interface/surface of TiN@Co 5.47 N during the electrocatalytic process is also important to boost the corresponding activity. [51] Defect effect between anions and cations in metal compounds modulates the electronic structure and surface characteristics, thereby changing the active site to improve the performance of the catalyst. Although metal compounds have been developed in the field of catalysts, the research still needs to focus on constructing different types of defects and bonds between different elements in a convenient and low-cost way, and more precise control of the density of vacancies and their distribution are also indispensable. [52] Synergistic effects, ligand effects, strain effects, and defect effects could act together or independently in electrocatalytic reactions, resulting in catalysts exhibiting excellent performance. Therefore, it is necessary to have a deeper understanding of the enhancement mechanism of catalysts. Among the catalysts reported, transition metal catalysts have been widely used in CO 2 electroreduction both theoretically and experimentally. However, most studies have shown that the correlation between the binding energy of various reaction intermediates and how to separate intermediates limits the development of catalysts for large-scale commercialization of CO 2 electroreduction reactions. [53] An efficient solution is that converting metal catalysts into individual atoms, which alter their binding energies with intermediates due to changes in the electronic environment. [54] However, when designing a single-atom catalyst, the coordination and pore size of the support material should be considered. [54] The internal structure of the catalyst is able to regulate by introducing covalent organic frameworks into the metal catalysts. Besides metal catalysts, non-metal catalysts are supposed to use for CO 2 electroreduction. Table 1 lists the performance of different catalysts used for CO 2 electroreduction. Moreover, the products of CO 2 electroreduction are not only related to the types of catalysts, but also they are connected with the preparation methods of the catalysts.

Application of CO 2 Electroreduction
The type and concentration of electrolytes have a significant impact on the selectivity of CO 2 electroreduction products. KCl, K 2 SO 4 , KClO 4 , and dilute HCO 3 − solutions tend to generate ethylene and alcohol, while methane is preferentially generated in concentrated HCO 3 − . [55] The high concentration of protons under acidic conditions could reduce polarization loss, but it affects the stability of the working electrode and also increases the occurrence of hydrogen evolution reactions. [56] Under alkaline conditions, CO 2 is more likely to be converted into multi carbon products, as higher pH can lower the energy barrier for activating CO 2 and inhibit the occurrence of hydrogen evolution reactions. [57] Multiphase electrochemical reactions occur on the surface of the working electrode, so the catalytic performance is determined by both the electrode and electrolyte. As an important part of CO 2 electroreduction, the study of electrode microenvironment could help better control the reaction. The electric double layer (EDL) formed between the electrode and electrolyte is the prerequisite for understanding the electrode microenvironment. [58] The structure of EDL could be described by Gouy-Chapman-Stern (GCS) model, which is composed of inner Hermann von Helmholtz layer, outer Hermann von Helmholtz layer, and diffusion layer. Among them, the inner Hermann von Helmholtz layer is the chemisorption of ions on the surface of the working electrode, the outer Hermann von Helmholtz layer is generated by the electrostatic attraction of ions, and the arrangement of ions in the diffusion layer is loose. [59] The concentration of solvated ions in EDL plays an important role in electrochemical reactions. Because of the negative pressure of its working electrode, the solvated cations accumulate on the EDL, enhancing the cation effect during CO 2 RR. The large-sized alkaline cations also have a significant impact on the selectivity of CO 2 electrocatalytic products. When cations such as K + , Rb + , and Cs + have larger relative sizes, their hydrates can better hydrolyze in CO 2 saturated 0.1 m HCO 3 − , thereby reducing the pH near the catalyst. The K + on the catalyst surface can stabilize the *COOH and *CO intermediates because the electron density on the carbon of the *COOH intermediate will increase in the presence of K + , thereby strengthening the C-Au bond. [60] Converting CO 2 into useful chemical products is a compelling strategy for mitigating environmental problems caused by increasing atmospheric CO 2 concentrations, while also benefiting energy storage. CO 2 can be activated under the driving of electric field and the reaction temperature is mild. What is more, it can be combined with other renewable energy sources, so CO 2 electroreduction has a high application potential. There are three main steps in CO 2 electroreduction: 1) Chemisorption of CO 2 on the surface of the catalyst (cathode); 2) electron transfer and proton migration by breaking carbon-oxygen bonds or forming carbon-hydrogen bonds; 3) desorption of products from the electrode surface and diffusion into the electrolyte. [61] In the electrolysis process, CO 2 can be reduced to different carbon oxides and hydrocarbons through two-electron, four-electron, six-electron, or even eight-electron transfer. CO 2 molecule accepts an electron from the electrode surface to form CO 2 − radical, which is the first and rate-determining step of CO 2 electroreduction. The CO 2 − radical is more prone to protonation. After accepting an electron and a proton, the single electron on the intermediate will acquire a second electron and combine with a proton to form formic acid. If a dehydration reaction occurs next, *CO is formed, resulting in the generation of C 1 product, such as formaldehyde, methanol and methane. In addition to C 1 products, CO 2 electroreduction can also form multi-carbon products. C-C coupling in this process is the rate-determining step in the formation of multi-carbon products. Table 2 summarizes the standard potentials for reducing CO 2 . At high potential, the reduction of *CO to -CHO is the first step of the reaction. Followed by another *CO to generate -COCHO intermediate, which is the precursor of C 2 and even C 2+ products. In the CO 2 electrocatalysis process, C 2 products can be generated by different numbers of electrons transferred. Herein, the C 2 products are classified according to the numbers of electrons transferred during the reaction process, and the corresponding researches are briefly introduced. 17

Two-Electron Process of CO 2 Electroreduction to Prepare Oxalic acid
In CO 2 electroreduction to C 2 products, oxalic acid is the only two-electron product. Oxalic acid is mainly served as reducing agent and bleaching agent, and is able to prepare ethylene glycol and smelt metals. [62] In industrial production, oxalic acid is mainly formed by the hydrolysis of oxalate esters, which is a cumbersome process. Electrocatalytic oxalic acid prepared by CO 2 electric reduction can reduce the production cost by simplifying the reaction steps while reducing the CO 2 content in the air. For example, Subramanian et al. used zinc as anode and stainless steel 304 L as cathode to convert CO 2 into zinc oxalate. The average Faradaic efficiency of zinc oxalate in this process was 73.9%. Then zinc oxalate was served as raw material to synthesize oxalic acid, which the extraction rate of oxalic acid was 58.1%. In addition, their research showed that the increase of reaction pressure and the decrease of solvent water content boosted the Faradaic efficiency of CO 2 to oxalate conversion, and the yield of zinc oxalate reached the maximum when the pressure was increased to 2 bar. [63] What is more, Bouwman et al. oxidized a dinuclear copper(I) complex by CO 2 in air to a tetranuclear copper(II) complex containing two CO 2 -derived oxalic acid bridging groups ( Figure 6A). Their study expounded that the binding of CO 2 with Cu(I) provides a low-energy pathway for the formation of two Cu(II)(CO 2 − ) radical intermediates. Due to the better thermodynamic properties of copper(II) oxalate, the subsequent C-C coupling on Cu(II) was more favorable for the formation of oxalic acid. [64] Based on this mechanism, Lung et al. verified the selective formation of Cu(II)-oxalate by calculation, and also proposed the formation mechanism of oxalic acid: Two Cu(I) molecules reduced one CO 2 molecule to Cu 2 (CO 2 − ) intermediate, and metalmediated nucleophile-like attacked the second CO 2 molecule to form Cu(II)-oxalate ( Figure 6B). [65] However, Andrew et al. indicated that oxalic acid was formed at the surface anion site by CO in aqueous electrolyte ( Figure 6C), without the participation of CO 2 − intermediate. Moreover, this process was highly sensitive to the hydrogen bonding environment. They prepared Cr 2 O 3 -Ga 2 O 3 alloy thin film catalysts wrapped on glassy carbon. At the potential of −1.48 V (vs Ag/AgCl) ( Figure 6D), the Faradaic efficiency of oxalic acid reached 59% in KCl aqueous solution with pH 4.1 ( Figure 6E). At the same time, the ratio of Cr 2 O 3 :Ga 2 O 3 was 3:1 ( Figure 6F). [66] Early catalysts for the electroreduction of CO 2 to oxalic acid were platinum, [67] mercury, and titanium, [68] as well as metal complex catalysts, [69] while the overpotential during conversion was −3.0 V (vs Ag/Ag + ). [67] Therefore, subsequent studies have focused on improving the product selectivity while reducing the overpotential of the reaction. Cheng et al. prepared a series of Pb-Sn mixed oxides supported on carbon black by hydrothermal method to generate oxalic acid by electroreduction of CO 2 in propylene carbonate solution. Among them, the Faradaic efficiency of oxalic acid reached 85.1% and the current density reached 2.0 mA cm −2 under the potential of −1.9 V (vs Ag/Ag + ) for the PbSnO 3 /C catalyst. During the reaction, the formation of formate was decreased due to the lower water content of propylene carbonate. [70] The small overpotential benefited from the synergy between the reduced lead and the oxidized tin in the PbSnO 3 /C catalyst, which stabilized the CO 2 intermediate.
At present, there are still deficiencies in the production of oxalic acid by CO 2 electroreduction. For example, the overpotential of oxalic acid is large, the selectivity of the product is low, the reaction is accompanied by the hydrogen evolution reaction, and the reaction process needs to be simplified. Therefore, future research should focus on how to modify the catalyst to reduce the overpotential of the reaction, discover many aprotic solvents to suppress the hydrogen evolution reaction, and explore other pathways for the conversion of CO 2 to oxalic acid.

Eight-Electron Process CO 2 Electroreduction to Produce Acetic Acid
The industrial synthesis of acetic acid is a multi-step reaction starting from natural gas through carbonylation of syngas, which consumes a lot of fossil energy. [71] Therefore, as an eight-electron product of CO 2 electroreduction, the production of acetic acid is of great significance in reducing costs and simplifying the industrial synthesis.
Genovese et al. deposited copper nanoparticles on carbon nanotubes as cathodes of flow electrolysis cells in CO 2 electroreduction, under normal temperature and pressure, with the Faradaic efficiency of acetic acid reaching 56%. They explained that the formation of acetic acid was attributed to the high concentration of CO 2 . Because a large number of CO 2 − radicals on the cathode surface would form acetic acid with the adsorbed -CH 3 (Figure 7A). [71] Based on this mechanism, Zang et al. used Mo 8 -modified copper nanoparticles in titania nanoarrays (Mo 8 @Cu/TNA) as a catalyst for CO 2 electroreduction ( Figure 7B). At a low potential of −1.13 V (vs RHE), the Faradaic efficiency of acetic acid reached 48.68% ( Figure 7C). Their study showed that the Cu-O-Mo interface in the catalyst promoted the formation of *CH 3 intermediate. Besides, the continuous coupling of the catalyst with CO 2 under the applied voltage improved the selectivity of acetate. [72] In addition, Liao believed that isolated active sites were significant in the formation of acetic acid. Traditional inorganic copper-based catalysts had low selectivity to acetic acid because their surfaces were not easy to form isolated active sites. When the active sites on the catalyst surfaces were too  close to each other, the C-C coupling pathway to generate ethylene and ethanol was prone to occur. They designed a 2D copper phthalocyanine covalent organic framework (PcCu-TFPN) as an electrocatalyst with a high Faradaic efficiency of 90.3% for acetic acid at a low overpotential of −0.8 V (vs RHE). Because the isolated copper-titanium cyanine active site with high electron density facilitated the C-C coupling of CO 2 − radicals with *CH 3 formation, while avoiding the production of ethylene and ethanol. [25] Different from the above mechanism of acetic acid generation, Gonglach et al. generated acetic acid by stabilizing oxalic acidtype intermediates. They designed a manganese corrole (Mncorrole) catalyst ( Figure 7D) and immobilized it on a carbon paper electrode to obtain a catalyst (Mn-Cor-CP). In a moderately acidic aqueous medium, acetic acid was the main product, and its Faradaic efficiency reached 63% ( Figure 7E). Because the active metal heterocycle in Mn-corrole stabilized the carboxyl group bound to the metal. What is more, the Lewis acid in the center of Mn(III) was more inclined to combine with the Lewis base at the oxygen site on the carboxyl group to promote C-C dimerization to form oxalic acid type intermediate, which was further reduced to acetic acid. [73] The only catalyst for the preparation of acetic acid in the electroreduction of CO 2 is copper-based catalyst, but by-products such as ethylene and ethanol are produced during the reaction process. Therefore, more attention should be paid to avoiding the occurrence of side reactions to improve the selectivity of acetic acid in the research of other catalysts.

Twelve-Electron Process CO 2 Electroreduction to Prepare Ethylene and Ethanol
Ethylene and ethanol are twelve-electron products of CO 2 electrocatalysis, which will be introduced in turn.

Figure 7.
A) Schematic diagram of the mechanism of CO 2 electrocatalysis to produce formic acid, acetic acid, and methanol. Reproduced with permission. [71] Copyright 2017, Royal Society of Chemistry. B) Schematic diagram of electrocatalytic CO 2 reaction on Mo 8 @Cu/TNA. C) Faradaic efficiency of different products at different potential on Mo 8 @Cu/TNA catalyst (green bars are acetic acid). Reproduced with permission. [72] Copyright 2021, Elsevier. D) Chemical structure of Mn-corrole; E) Stability test of Mn-corrole and Faradaic efficiency of acetic acid. Reproduced with permission. [73] Copyright 2020, Wiley-VCH.
According to the global newswire, the market interest rate of ethylene is expected to reach 159 billion dollars by 2027, and ethylene production mainly comes from the petroleum cracking industry at present. [74] However, the source of ethylene which is the basis of polymer industry will be seriously threatened with the continuous depletion of petroleum resources, the production of ethylene by CO 2 RR has attracted much attention. During the reduction of CO 2 to ethylene, C 2 H 4 can be formed by coupling two *CH or two *CO, which can also form ethanol (Figure 8A). Another approach is to dimerize *CO with one electron to form *C 2 O 2 key intermediate, followed by protonation to form *CO-COH, which further forms C 2 H 4 , CH 3 CHO, and C 2 H 5 OH ( Figure 8B). [61] Among the catalysts, copper-based materials are highly efficient catalysts for converting CO 2 to C 2 products, and forming process mainly occurs on Cu(100), Cu (111), and Cu(110) surfaces.
By studying the reaction of CO 2 on the surfaces of Cu(100), Cu (111), and Cu(110) single crystals, Huang et al. found that the onset potential for the formation of C 2 H 4 was always 300-400 mV negative than the onset potential of CO. Besides, C 2 H 4 could only be formed after a large amount of CO gas, the high surface coverage of *CO intermediates was the key to the selective formation Figure 8. The mechanism of CO 2 to prepare ethylene: A) Two *CH or CO coupling to produce ethylene; B) *CO dimerization to form C 2 product. Reproduced with permission. [61] Copyright 2017, Elsevier. C) CO generates ethylene on Cu(111) surface. D) CO generates ethylene on Cu(100) surface. Reproduced with permission. [77] Copyright 2012, American Chemical Society.
of C 2 H 4 . [75] Similarly, Hori et al. investigated the electroreduction of CO 2 with various copper single crystal electrodes. The result showed that ethylene was the main product on the surface of the Cu electrode based on the (100) step surface. Moreover, the introduction of (111) or (110) steps on the (100) face further promoted the formation of ethylene. [76] Schouten et al. studied the electrochemical reduction of CO on two single-crystal copper electrodes, and the results indicated two pathways for the formation of ethylene. One pathway was the common intermediate with the formation of methane, which occurred on the (111) and (100) surfaces. The other pathway was the selective formation of C 2 H 4 from CO at relatively low overpotentials, preferentially occurred on the (100) plane ( Figure 8C,D). [77] Ethanol, as an essential solvent, liquid fuel, and raw material for the production of many chemical products, is also expected to be obtained by electrocatalytic CO 2 . In the process of CO 2 electroreduction to ethanol, two *CO molecules are coupled through C atoms, and the dimer *C 2 O 2 − is formed by electron transfer under the action of electrode polarization. One of the C atoms and a O atom are each combined with two Cu atoms, and the negative charges are mainly distributed on the CO group that is not bonded to the surface. Then, the negatively charged O atom undergoes hydrogenation reaction to form *HOCCO intermediate. After that, the intermediate is further hydrodehydrated to form the *CCO intermediate, which then experiences a hydrogenation reaction between the carbonyl O and C atoms to form the *CH 2 CHO intermediate. There are two cases for the subsequent reaction: 1) If the hydrogenation reaction occurs between C-O bonds, C 2 H 4 will be formed; 2) If the hydrogenation reaction occurs between C-C double bonds, *CH 3 CHO intermediates are Figure 9. Mechanism of CO 2 production of ethanol. Reproduced with permission. [78] Copyright 2013, Wiley-VCH. generated, and ethanol will be produced after two hydrogenation reactions (Figure 9). [78] The type and morphology of the catalysts affect the selectivity of the product. The researchers introduced electron-deficient structures or Cu + sites by doping silver in the catalysts, and changing the morphology of the catalysts to improve the selectivity of ethanol. Lv et al. prepared a silver-doped copperbased catalyst (Cu 3 Ag 1 ) with electrodeposition of Cu followed by electro-replacement with Ag(I), which had an electron-deficient structure. Compared with the reversible hydrogen electrode, the Cu 3 Ag 1 electrocatalyst enabled the Faradaic efficiency of CO 2 to ethanol conversion to reach 63%, and the ethanol partial current density was −25 mA cm −2 . Compared with the bare electrodeposited copper substrate catalyst, the selectivity was increased by 126 times and the activity increased 25 times. The improved performance was attributed to electron-deficient Cu sites created by the transfer of interphase electrons from Cu to Ag, which favored the adsorption of key intermediates in the alcohol reaction pathway, such as CH 3 CHO* and CH 3 CH 2 O*. Therefore, this electron-deficient catalyst realized a more favorable C 2 H 5 OH reaction pathway with an alcohol/ethylene ratio of 38:1. [38] Different from Lv, Iyengar et al. studied the catalytic behavior of copper nanocubes (Cu cubs ) with different sizes under the tandem catalytic framework of Ag nanospheres. The research showed that ethanol was selectively produced at the step edge and ethylene was produced at the step site, the smaller Cu cubs had a higher selectivity to ethanol due to the larger edge-to-face ratio. The result pointed to the importance of developing Cu nanostructures with only edge sites exposed to further improve the selectivity of ethanol. [79] Woo-Bin et al. improved ethanol selectivity by introducing Cu + sites into carbon-based materials. They used chemical vapor deposition to grow graphene and performed a post-oxidation process to prepare an electrocatalyst (Cu + /hf-Cu) with the coexis-tence of Cu + sites and high crystal planes. The introduced Cu + promoted the stabilization of *CO due to its strong *CO bonding, and the high-facets provided a favorable C-C coupling center for the formation of ethanol, which Faradaic efficiency of ethanol reached 53% at −0.8 mV (vs RHE). [80] The electrochemical method provides an effective approach for CO 2 conversion. However, the preparation of high value-added C 2 products such as ethylene and ethanol still has problems of low selectivity and activity, which seriously hinder the development of industrial applications. Cu-based catalysts can realize the C-C coupling step because they have appropriate adsorption capacity for key intermediates in CO 2 RR. Ethylene and ethanol generation paths are relatively complicated, designing efficient Cu-based catalysts which can separate them precisely, and deeply understanding their formation mechanism on other catalysts are very important for preparing the product.

Fourteen-Electron process CO 2 Electroreduction to Prepare Ethane
Ethane is the only 14-electron C 2 product and the selectivity is negligible compared to ethylene, ethanol, and other products in CO 2 electroreduction. However, ethane with high energy density can replace fossil fuels by utilizing existing infrastructure for storage, distribution, and consumption. [81] Therefore, the reduction of CO 2 to ethane by electrocatalysis is of great importance.
Ma et al. synthesized Cu(OH) 2 and CuO nanoarrays on copper foil. Then, CuO nanoarrays were electroreduced into Cu nanowire arrays with different lengths, which were used in CO 2 electroreduction. They proposed that ethane was generated through the CH 3 CH 2 O intermediate and there was ethanol on longer copper nanowires (Figure 10B). Because longer copper nanowires formed a higher local pH value ( Figure 10A). The Figure 10. A) Schematic diagram of electrolyte diffusion into copper nanowire arrays. B) Reaction pathway of CO 2 electroreduction on copper nanoarrays. Reproduced with permission. [82] Copyright 2016, Wiley-VCH. C) Two-step preparation of ID-Cu by vacuum-sealed ampoule calcination-electrochemical reduction. D) Selectivity determination of ethoxylated intermediates between ethane and ethanol. Reproduced with permission. [83] Copyright 2020, Wiley-VCH. E) Faradaic efficiency of ethane at different potentials for different samples. F) Reaction on derivatized copper catalysts with different thicknesses. Reproduced with permission. [84] Copyright 2017, American Chemical Society. enhancement of local pH altered the *CO coupling mechanism, thereby reducing the selectivity of ethylene. [82] Based on this ethane generation mechanism, Vasileff et al. prepared an iodidederived copper catalyst (ID-Cu) by calcining foamed copper and iodine in a vacuum-sealed ampoule and electrochemical reduction ( Figure 10C). The Faradaic efficiency for CO 2 to ethane was higher than that of oxide-derived copper catalysts (OD-Cu). Because the ethoxy intermediate (*OCH 2 CH 3 ) with an oxygen bond had better stability on the copper lattice containing a trace amount of iodine. Besides, the ethoxy intermediate effectively inhibited the formation of ethylene and ethanol ( Figure 10D). [83] In addition to being formed from ethoxy intermediates, ethane can also be generated by dimerization of *CH 3 intermediates. Handoko et al. prepared three oxide-derived copper catalysts (samples A, B, C) with different thicknesses for CO 2 electrore-duction. The experiment exhibited that on the thicker derivatized copper catalyst (Sample C), the Faradaic efficiency of ethane reached a maximum of 10.5% at a potential of −0.78 V (vs RHE) ( Figure 10E). Their study showed that ethanol was the only product in experiments where the -OCH 2 CH 3 intermediates could be detected, ruling out that ethane was generated from ethoxy groups. They considered that ethane could be produced by dimerization of the *CH 3 intermediate ( Figure 10F), which was formed from -COH. [84] The exploration of different ethane formation mechanisms helps to further study the interaction between different intermediates and catalysts. Then we develop low-cost and high ethane selectivity catalysts.
Among the reported catalysts, copper is capable of catalyzing the production of a large amount of hydrocarbons at a high reaction rate in a CO 2 saturated aqueous solution and could Figure 11. A) Schematic diagram of membrane electrode. B) The ionomer enhances the diffusion of CO 2 in the catalytic layer. C) The gas diffusion layer reduces the interaction between CO 2 and OH − . D) Thinner catalyst layer increases the CO concentration on the catalyst surface; E) thinner membrane; and F) higher water uptake reduces the water flux at the cathode. Reproduced with permission. [89] Copyright 2021, Springer Nature.
realize C-C coupling at room temperature and pressure. [82] Different electrolysis conditions also affect the product selectivity. Therefore, controlling the target products remains a major scientific challenge. In addition, to develop other high-performance and low-cost catalysts, it is necessary to deeply understand the formation mechanism of C 2 products and explore the relationship between the formation mechanism and electrolysis conditions to obtain the target products in a controllable manner.

Introduction of CO 2 Membrane Electrode
H-type electrolytic cell is the most commonly used device in CO 2 electroreduction. The gas CO 2 is dissolved in the electrolyte, transferred to the surface of the catalyst. Then, CO 2 molecular reacts on the surface of catalysts, generating valuable chemicals such as CO, formic acid, ethylene, ethanol, etc. [85] Although high Faraday efficiency can be achieved in the H-type electrolytic cell, the current density in the electrocatalytic process of CO 2 is relatively low due to the low solubility of CO 2 in the electrolyte and the serious hydrogen evolution reaction. [86] In order to solve these problems, the membrane electrode is used in the electrolytic cell to make CO 2 directly reach the catalyst surface through the gas diffusion layer to increase the CO 2 concentration. [57] The membrane electrode is the core component of the battery and the place to realize the conversion of chemical energy to electrical energy. [87] The membrane electrode is mainly composed of an ion exchange membrane, a catalyst layer, and a gas diffusion layer. The catalytic layer provides a three-phase mass transfer interface and a reaction zone. The ions pass through the ion exchange membrane to the cathode catalyst layer for a reduction reaction. The generated gas and liquid diffuse through the gas layer out. The interface between the microporous layer of the gas diffusion layer and the catalytic layer plays an important role in the overall mass transfer process. [88] The gas diffusion layers and membranes control the mass transport and reaction, directly regulating the CO 2 reduction process. Thinner catalytic layers increase CO 2 concentration ( Figure 11D). Ionomers can enhance the diffusion of CO 2 in the catalyst layers ( Figure 11B). In addition, the gas diffusion layer could reduce the interaction between CO 2 and OH − ( Figure 11C) and achieve CO 2 electroreduction under alkaline condition. Thinner ion exchange membranes and higher water absorption decrease water flux to the cathode ( Figure 11E,F). Besides, the ion exchange membrane conductivity affects the permeation selectivity. [89]

Catalyst Layer
The catalyst layer is porous structure, in which the solid part of the catalyst is composed of catalyst nanoparticles. [90] The catalytic layer is the core part of the membrane electrode and it is a micro-nano-porous structure electrode composed of perfluorosulfonic acid resin ionomer adhesion catalyst particles. [91] Effective electrocatalytic reactive sites require stable gas-liquid-solid three-phase interfaces. [88] In order to facilitate the transfer of reactants, electrons, and protons in the reaction process and the mass transfer resistance in the catalyst layer should be small enough. Commonly, the catalyst layer is modified by gradient and ordered structure design to increase the mass transfer rate of the reaction process. [92,93] Ye et al. prepared a catalyst layer with platinum content and pore gradient distribution. In this catalyst layer, the pore size of the platinum-depleted layer was larger than that of the platinumrich layer (Figure 12A). This structure enhanced the platinum  [94] Copyright 2017, Elsevier. C) Performance comparison of membrane electrodes with different cathode structures under ambient pressure. D) In situ impedance curves of membrane electrodes with different cathode structures. Reproduced with permission. [95] Copyright 2010, Elsevier. utilization rate and increased the mass transfer rate in the catalytic layer. The performance of the membrane electrode using this catalytic layer was improved by 11% compared to the traditional membrane electrode ( Figure 12B). [94] In addition, Su et al. prepared a novel bilayer catalyst with different Pt contents. The higher Pt content in the inner layer could enrich Pt. The lower Pt content in the outer layer was used to maintain a suitable catalytic layer thickness. The performance of the new double-layer catalyst was significantly improved, compared with the traditional bilayer and the single-catalyst. What is more, the current density of the new double-layer catalyst was 35.9% higher than that of the single-layer catalyst ( Figure 12C), which was due to the gradient of platinum content produced a more efficient electrochemically active layer ( Figure 12D). [95] Ordered structure can distribute catalysts, catalyst supports, proton conductors, and other substances in order to increase the mass transfer channel. The mass transfer channels are able to reduce mass transfer resistance by expanding the gas-solid-liquid three-phase interface area. [93] In the cathode catalytic layer, the diffusion and mass transfer of gas are affected by the thickness and the porosity of the catalytic layer. When the catalytic layer is relatively loose and has porous structure inside, it is convenient for the mass transfer process of gas and liquid. [96] Increasing the average pore size, enhancing the graphitization level of the catalyst, and decreasing its hydrophilicity can improve the transport of liquid on the surface of the catalyst. [91] At present, the catalytic layer based on heterogeneous catalysis has the following three disadvantages: 1) The nanoparticles that are not covered by the ionomers or located in the micropores of the support cannot complete the catalytic reaction, which limits the three-phase interface. 2) Ionomer agglomeration results in gas mass transfer resistance. 3) Atoms in the core region of the nanoparticle cannot be used effectively. [97]

Ion Exchange Membrane
The ion exchange membrane is a polymeric material that conducts ions between electrodes. General polymer-based solid electrolyte membranes are often prepared from some polyphenylene and polyolefin polymer materials. The hydrophobic microdomains in the polymer significantly hinder the ions in the membrane. Therefore, polymers have problems with tortuous ion transport paths, increased short circuits, and dead zones. In recent years, the development of organic molecular sieve materials, such as metal organic frameworks and covalent organic frameworks, have provided new ideas for solving this problem. [98] The interfacial area between the ion exchange membrane and the catalyst layer affects the effective active area of the metal catalyst. Thus, researchers have done a lot of work to expand the interfacial area between them. Cho et al. prepared a surfacemodified ion-exchange membrane by plasma etching. The membrane electrode using the ion-exchange membrane etched for 10 min showed a 19% improvement in performance in comparison to the untreated membrane. However, the performance of the membrane electrode of the ion-exchanged etched for 20 min improved by 8%, indicating that the performance of the membrane electrode using the etched membrane was affected by the different film thickness and surface morphology due to the different etching time (Figure 13). [99] Cuynet et al. obtained nano-and micro-scale patterned Nafion films by laser irradiation. They connected ion-exchange films and thin catalytic layers with a small amount of platinum catalyst by plasma magnetron sputtering. The laser irradiation precisely controlled the surface structure of the ion exchange membrane and improved the utilization rate of the catalyst. [100] In addition to the surface morphology, the pore structure of the ion exchange membrane has an important influence on the interface area with the catalyst layer. Using a porous membrane for the membrane electrode extends the interface area between the ion exchange membrane and the catalyst layer. In addition, when the moisture of the membrane pore wall is high enough, the ion conduction rate can be improved so that the membrane electrode shows excellent performance. [101] The cation exchange membrane not only inhibits the transport of HCO 3 − and CO 3 2− , but also leads to a high proton concentration for the hydrogen evolution reaction. [102] Therefore, the anion exchange membrane is more favorable for the membrane electrode, compared with the cation exchange membrane. Some high-conductivity anion exchange membranes have many tiny gaps with electrodes, which greatly affects the mass and electron transfer during the use of the device, thereby reducing the electrolysis efficiency. Improving the interfacial contact state between the ion exchange membrane and the catalyst could reduce the interface ohmic impedance and provide a guarantee for the high performance of the energy conversion process. [103]

Gas Diffusion Layer
The gas diffusion layer is the outermost structure of the membrane electrode, which is use to connect to the flow channel and the catalytic layer on the electrode plate. It plays an important role in distributing water, gas, electricity, and heat. Therefore, once the gas diffusion layer is deformed, it will affect various mass transfer processes and further affect the performance of the battery. [104] The thickness and porosity of the gas diffusion layer exert an influence on the flow uniformity, diffusion flux, and ohmic resistance. Flow uniformity and ohmic resistance will increase with increasing thickness and porosity, which have opposite effects on diffusion flux. [105] Moreover, the distribution of porosity seriously affects the distribution of gas and current density in the catalyst layer. Thereby, it is necessary to design the porosity gradient. Huang et al. designed a 3D two-phase nonisothermal gas diffusion layer model to study the interaction between heat and moisture transport. Their study showed that the distribution of the linear gradient of porosity in the gas diffusion layer increased the water flux. The limiting current density was improved from 10 696 to 13 136 A m −2 when the porosity of the parallel channel was 0.7. [106] The gas diffusion layer is usually composed of a base layer and a microporous layer. Generally, carbon fiber paper, carbon fiber cloth, and graphite are used as the base material layer. It has an important influence on the distribution and selectivity of electrochemical CO 2 reduction products. The increased current density leads to a higher local pH at the cathode, which not only effectively suppresses the competing hydrogen evolution reaction but also promotes the C-C coupling pathway. During the operation of the system, the gas diffusion layer gradually loses its hydrophobicity. After a long period, the catholyte will penetrate the back of the gas diffusion layer resulting in blockage of electrode pores, thereby decreasing the conversion rate and stability. When operating in KOH or KHCO 3 cathode electrolysis liquid conditions, the massive formation of carbonate can also lead to the attenuation of the gas diffusion layer. [107] The future development of gas diffusion layers should focus on optimizing transport across scales and components. At the same time, consideration should be given to the control of structure and wettability. For the catalytic layer, the activity of the novel catalytic layer is large enough on the rotating disk electrode, but still requires considerable improvement at the membrane electrode assembly and stack level. [108] The membrane electrode with gas diffusion electrode could solve the problem of low solubility of CO 2 in the solvent and improve the mass transfer efficiency. What is more, the separation of catholyte and anolyte is beneficial to enhance the efficiency of product separation and avoid product re-oxidation. However, the membrane electrode will have the problems of concentration polarization and membrane dehydration when the current density is high. Further exploration of the reaction device, membrane and catalyst materials is needed. For example, 1) add new polymers to the exchange membrane to optimize the electrode conductivity. 2) Develop new highly active nanomaterials as catalyst carriers to boost the production of target products. 3) Study different electrolytes to suppress the hydrogen evolution reaction better.

Conclusion and Outlook
Green and easy-to scale CO 2 electroreduction has become a strategy of much interest in closing the carbon cycle, while also effectively addressing the problem of resource shortages. Its operating conditions are mild, the additional voltage required in the conversion process can be obtained from clean energy sources and the various chemical products with high added value can be obtained by changing catalyst types and additional voltage. Therefore, it could be widely used in producing industrial chemical fuels. High selectivity of CO 2 electroreduction products is not only of great value for future industrial production but also of great significance for the scientific understanding of these complex multi-step reactions. Thus, it is significant to conclude the research progress of CO 2 electroreduction to guide future studies. In this work, we have briefly described the four catalyst enhancement mechanisms: synergistic effect, strain effect, ligand effect, and defect effect. Then, on the basis of the number of electrons transferred in the CO 2 electrocatalysis process, the mechanism and application of CO 2 electroreduction to prepare oxalic acid, acetic acid, ethylene, ethanol, and ethane were listed in turn. These reactions could effectively solve the problem of resource shortage and reduce the consumption of fossil fuels, showing great research significance. Finally, a CO 2 electroreduction generating device-membrane electrode was briefly introduced from the three aspects of ion exchange membrane, catalyst layer, and gas diffusion layer. The advantages and disadvantages of these three components of membrane electrode were expounded. CO 2 electroreduction is of great significance for alleviating climate and energy problems. Various reported works have made a lot of effort to develop catalysts. However, most catalysts' selec-tivity and activity are still insufficient for practical applications. Among the currently discovered metal catalysts, copper could achieve C-C coupling. The development of new catalysts with larger specific surface area, higher catalytic activity, and better selectivity is still the key difficulty to improve the efficiency of CO 2 electroreduction. For example, single-atom catalysts could exhibit excellent catalytic performance and selectivity by exploiting their maximum atom utilization and tuning metal active centers with unique electronic properties. In addition, well-designed metal nodes and organic ligands have specific pore structures and active site. Therefore, metal organic frameworks (MOFs) composed of metal nodes connected by organic bridging ligands can enhance the electrochemical reduction of CO 2 In addition, CO 2 electroreduction also has the following problems: 1) High overpotential will increase electrolysis energy consumption and reduce energy conversion efficiency. 2) The low solubility of CO 2 in water reduces its utilization rate. 3) The hydrogen evolution reaction seriously affects the yield. 4) The selectivity of a single product selectivity is low due to the numerous reaction paths. 5) Catalysts can be poisoned at a negative potential, resulting in poor stability. 6) Low current density resulting in higher costs, and so on.
In the subsequent research on catalysts, theoretical calculations should be combined with artificial intelligence to predict the stability, activity, and selectivity of catalysts through the energy and structural descriptors, guiding the design of catalysts. Besides, developing high-throughput synthesis techniques can also rapid and reproducible screen the prepared catalysts. In order to better understand the relationship between material structure and catalytic activity, in situ characterization techniques, such as X-ray photoelectron spectroscopy, infrared spectrum, and so on, should be vigorously developed to detect structural changes in catalysts. Adjusting the chemical and physical properties of the catalysts through various methods can improve the catalytic activity of the material and achieve high stability. In addition, coating or composite the electrode surface with the material to extend the durability of the interaction between the electrode and the catalyst, thereby achieving excellent catalytic performance. For the sake of realizing the industrialization of CO 2 electroreduction, the catalyst materials should not only display good catalytic performance but also meet the requirement of the catalyst cost. The electrochemical performance has a significant impact on the economic feasibility of CO 2 RR. The current density is one of the important influencing factors. The low current density during the reaction process could cause the problem of excessive electrode area. The increased electrode area requires larger equipment, thereby increasing energy consumption and costs. The current density could be improved by attaining more active site of the catalyst. At the same time, the current density could also be raised by enhancing the conductivity of the catalyst and using gas diffusion electrodes to promote CO 2 diffusion. For the influence of the electrolytic cell on the reaction, we need to use flow electrolytic cells to solve the limited current density due to the corresponding low mass transfer efficiency in H-type electrolytic cells. Additionally, more in-depth and comprehensive research is needed on the C-C coupling steps. Understanding the mechanism is crucial for improving the selectivity of Cu catalysts toward individual C 2+ products. In the future, CO 2 electrocatalysis technology will be of great significance in closing the carbon cycle and creating economic benefits, which is expected to solve resource shortage and environmental problems. We believed that the discussion in this work could provide some practical suggestions for the application of CO 2 electroreduction.