Electrocatalytic Reduction of CO2 to Value‐Added Chemicals via C–C/N Coupling

The efficient conversion and utilization of CO2 is imminent due to the increase of global CO2 emissions year by year, so the strategy of electrocatalytic CO2 reduction to high value‐added chemicals comes into being. The CO2 reduction reaction (CO2RR) can yield diversified products, including C1 (like CO, CH4, etc.), C2 (like C2H4, C2H5OH, etc.), and other multicarbon compounds. Thereinto, selective transformation to C2 compounds is relatively significative by reason of high industrial value but being greatly challenging, particularly involving complex CC coupling. Herein, the reaction mechanisms of several C2 products (including C2H4, C2H5OH, and H2C2O4) and corresponding latest research progress are summarized and discussed. Besides, N atoms, from N2, nitrates, and nitrites, are integrated into CO2RR to yield nitrogen‐containing organics (like urea, amides, and amines), broadening the application field of CO2RR. Therefore, the part on the electrocatalytic coupling of CO2 and nitrogen sources (including N2, NO3−, NO2−) by CN bonds is also delved in this review, including recent advances, reaction mechanisms, and related key intermediates.


Introduction
The latest report released by the International Energy Agency (IEA) points out that global CO 2 emissions hit a record high of 36.2 billion tons in 2021. The resulting environmental problems, such as melting arctic ice, rising sea levels, and acidification of seawater, have seriously affected survival of human beings. [1] Therefore, it necessitates capturing, storing the emitted CO 2 , and further converting it into high value-added chemicals to achieve carbon balance. However, CO 2 is so stable at room temperature that it is hard to be activated and further reduced to "useful" chemicals. Typically, it takes an energy of 220-300 kJ mol À1 to break the C═O bond. [2] At present, there are three methods of converting CO 2 : thermal catalysis, photocatalysis, and electrocatalysis. [3] Thereinto, thermocatalytic CO 2 reduction usually requires high temperature and high pressure, while photocatalytic CO 2 reduction commands photosensitizer or even electron sacrificer, and the efficiency is relatively low. By contrast, electrocatalytic reaction has attracted more attention of researchers, [4] especially CO 2 electroreduction due to mild reaction conditions, high reaction rate, and convenient separation of products. [5] Electrocatalytic CO 2 reduction reaction (CO 2 RR) is a process of reducing CO 2 molecules on the cathode using electric energy. In this process, CO 2 can be converted into some chemicals and fuels through the multielectron transfer mechanism, including C 1 products (like CO, CH 4 , CH 3 OH, HCOOH/HCOO À , etc.), [6] multicarbon (C 2þ ) alkenes, and oxygenated compounds (like C 2 H 4 , C 3 H 6 , C 2 H 5 OH, etc.). [7] Among them, C 2þ products currently play a very important role in the supply of energy and certain chemicals. [8] For example, ethylene, a basic chemical raw material, can be used in synthetic fiber, rubber, plastic, and other important chemicals. Ethanol is an important solvent, liquid fuel, and raw material of many chemical products, which is widely used in medical and health care, food industry, and other fields. Also, N-propanol has a high octane number (18) and bulk energy density (27 MJ L À1 ). However, to obtain long carbon chain products by electrocatalytic CO 2 reduction, it requires relatively high overpotential, accompanying the low energy conversion efficiencies. Therefore, CO 2 RR is not suitable for the production of C 3þ products by consideration of energy and economy. For C 2 products, C─C coupling is a very vital step in the CO 2 reduction process, which is a rate-determining one. It is necessary to accurately design the catalyst conducive to C─C coupling, requiring a summary of the existing work and the underlying mechanism of coupling.
Moreover, the chemicals obtained by electrosynthesis can be further broadened by integrating atoms other than C, H, and O into CO 2 RR. For example, C─N coupling can produce some organic nitrogen compounds, including acetamide, urea, amino acids, etc., which have an extensive range of applications in medical chemistry, agriculture, and other fields. [9] Industrially, however, these organic nitrogen compounds are produced by thermocatalytic conversion under high temperature and pressure. As early as 1963, Rapson and Bird prepared glycine by electroreduction of oxalic acid and nitric acid, and the interaction between the electrochemically generated glyoxylate and hydroxylamine resulted in the C─N coupling of glycine. [10] Plus, in the last few years, researchers have made a throng of attempts to integrate N atoms to CO 2 RR, [11] and it turns out that this strategy for electrosynthesis of nitrogenous organic compounds is feasible. Similar to C─C coupling, C─N coupling is also a central step in the electrosynthesis of organic nitrogen compounds, which deserves to ulteriorly study its mechanism.
As mentioned above, electroreduction of CO 2 to high valueadded chemicals by C─C/N coupling is a promising strategy. To the best of our knowledge, there is no review article that integrates these two parts, which is what we will review here ( Figure 1). A brief introduction about various electrolytic cells, which are very important for the evaluation of catalytic performance, is given. Subsequently, the recent research advances of C─C coupling to C 2 compounds (mainly including ethylene, ethanol, oxalic acid), C─N coupling to organic nitrogen compounds (including urea, amides, amines), and the corresponding reaction mechanisms are described. Last but not least, we make a summary and outlook for C─C and C─N coupling in the process of CO 2 electroreduction. We hope that this review will shed light on the synthesis of high value-added chemicals from CO 2 RR, especially in the burgeoning field involving C─N coupling. It is believed that this will broaden the application field of CO 2 conversion and utilization to a certain extent.

C─C Coupling to C 2 Compounds
The CO 2 RR occurring at the cathodes of electrolytic cells results in different products based on the number of transferred electrons and protons. As shown in Table 1, the electrochemical reaction equations of electrochemical CO 2 reduction to various products and the corresponding reduction potentials are listed. It can be seen from the table that the products with higher added value often commands more than two electron transfers, thus requiring higher energy to hydrogenate the adsorbed *CO intermediate. In 1985, Hori's group [12] studied the reduction products of CO 2 on various metal electrodes for the first time. They found that HCOOH was mainly generated on the surface of Cd, In, Sn, and Pb electrodes, CO was mainly generated on Ag and Au electrodes, and a considerable amount of deep reduction product methane was obtained on the surface of Cu electrode. Subsequently, they found the presence of ethylene in the reduction products on the Cu electrode. [13] So far, Cu has been the only metal that acts as the active site of the C 2 products. The various products are due to the distinct formation energy barrier and adsorption-desorption capacity of different intermediates. [14] For instance, the HCOO À is readily produced by *OCHO formation and *HCOOH desorption, whereas CO is obtained by *COOH formation and *CO desorption. [14] The production of C 2 products, involving the coupling of multiple protons and electrons, is more complex. It is worth noting that *CO is a pivotal intermediate in the formation of C 2 products from CO 2 through *CO dimerization and succeeding hydrogenation processes. [15] Additionally, the C─C coupling reaction is a very critical ratedetermining step, which requires the surface of the catalyst to firmly bind the *CO intermediate to establish sufficient *CO coverage for further *CO dimerization. [16] Cu-based catalyst is sufficient to provide the surface coverage of *CO desired by C─C coupling on account of its relatively suitable adsorption energy for *CO, which provides a good platform for the study of C─C coupling preparation of C 2 products. Although the exact reaction mechanism of C 2 product formation is still ill-defined, and the reaction path may change with different reaction conditions, there are four basic steps widely recognized in modeling this heterogeneous catalytic process: 1) CO 2 molecules are chemically adsorbed onto the surface of the electrocatalyst to form partially charged *CO 2 δÀ species via interacting with metal atoms on the surface. It is worth noting that CO 2 can not only be absorbed by linear adsorption, but also be further activated by electrode surface defects and solvent interaction. [17] 2) The C─O bonds are cleaved by transferring electrons, forming reduced intermediates.
3) The formation of C─C bonds gives rise to a variety of C 2 species. This step is challenging because it requires sufficient *CO coverage for further C─C coupling toward C 2 products via *CO dimerization, CHO─CO coupling, CO─COH coupling, or CO─C coupling. Besides, this coupling step and subsequent hydrogenation process involving varieties of reaction conditions (such as pH, over potential, and cation) are crucial to branch into C 1 /C 2 productions. In a word, the formation of C─C bond is limited by many factors in realistic electrocatalytic environment. Complex pathways have an effect on the selectivity of C 2 compounds. 4) Rearrangement of the product followed by desorption from the active site and diffusion into the bulk electrolyte solution. Undoubtedly, there are still some daunting challenges in the CO 2 RR process: C═O activation, precise control of C─C coupling, inhibition of competitive hydrogen evolution (HER), and separation of mixed products. There have been many reviews to summarize various catalysts to various products (including C 1 , C 2 , C 3þ compounds) up to now. [3c,15,18] Here, we only make a summary of the C─C coupling involved in the reduction of CO 2 to C 2 compounds (C 2 H 4 , C 2 H 5 OH, H 2 C 2 O 4 ), focusing on the mechanistic part.

CO 2 to Ethylene
Cu is considered as the efficient single-metal catalyst for the generation of C 2 compounds, which benefits from its optimal the binding energy of *CO in order to further hydrogenation process and regulates product. Two major pathways toward ethylene and ethanol are shown in in Figure 2. The key intermediates (OC**CO and *COCHO) are generated from the dimerization of *CO as well as the coupling between *CO and *CHO. It is responsible for the formation of ethylene and ethanol via hydrogenolysis and hydrogenation, respectively. In CO dimerization process, Koper et al. show that CO coupling step mediated by electron transfer is a rate-determining step. [19] Further research shows that benefited from stabilization by charged water layer on Cu (100) and Cu(111), the formation of CO dimer is more exergonic. And C 2 products are superior at low overpotentials based on a lower theoretical activation barrier of dimerization, consistent with all experimental observations on Cu (100). [20] However, when applied potential is too low to meet the need of À0.6 V(RHE), the rate of ethylene formation will decrease. Because ab initio molecular metadynamics simulations (AIMμD) free-energy calculations demonstrate that H* has an edge over *CO to be adsorbed on surface sites at pH 7. [21] Similarly, Liu's work elucidates that C 2 products activity and selectivity are tuned by CO coverage at high overpotentials, according to energetics estimated from an explicit solvent model. [22] Thus, the appropriate CO coverage is more conducive to obtain C 2 products selectively. And then, it is the first trial to apply the constant electrode potential (CEP) model that presents *COCHO(3) is the key for selectivity toward ethylene and ethanol. [23] Given the sensitivity of reaction mechanism to reaction conditions and catalytic surfaces, there is an important CO─COH to branch into C 2 products at neutral pH. [24] Interestingly, the combination of surface *C (originated from the reduction of CO via COH*) and CO to CCO* is favorable in thermodynamics at the interface. It is rational to consider the energetic difference between barriers for CCO* and CH* formation as efficient descriptors to direct the product selectivity at pH 7. [25] More efforts for Cu materials, such as crystalline faceting, surface reconstruction, alloying strategy, and so on, have been made. These strategies can reduce the overpotential of electroreduction CO 2 and improve the selectivity to a certain extent.
Hori et al. found that the (100) and (110) crystal planes of Cu can selectively catalyze the reduction of CO 2 to C 2þ products, in which the (100) crystal planes tend to form C 2 H 4 , while the (110) facets facilitate the formation of oxy-containing hydrocarbons, including C 2 H 5 OH and CH 3 COO À . [26] Although Cu(100) facet favors C 2 H 4 formation, it is liable to reconstruction into the (111) crystal planes during the electrocatalytic process. [27] Sargent et al. proposed a synthesis strategy designed to expose and preserve the Cu (100) surface for addressing this problem (Figure 3a-c). [27] They mainly use intermediates (such as *CO 2 , *COOH, and *CO) adsorption to adjust the Cu exposure surface, yielding a Cu catalyst with a high proportion of (100). The role of these intermediates can be likened to that of capping agents. Density functional theory (DFT) calculations showed that the adsorbent intermediates under the condition of CO 2 reduction with CO 2 were conducive to the reconstruction of Cu (100) on the surface, while the formation of Cu (111) was more favorable under hydrogen evolution reaction (HER) without CO 2 . Ultimately, the result revealed that the Faradaic efficiency (FE) of %70% for the C 2 H 4 products at high current density (580 mA cm À2 ) was obtained over the Cu─CO 2 RR sample. Moreover, a large number of previous experiments and DFT calculation results demonstrated that the grain size and morphology of Cu also have a crucial influence on the product selectivity during CO 2 reduction. [28][29][30] Varieties of Cu cubes with nanoscale edge were developed by colloidal chemistry-based method and the study revealed a size dependence of the selectivity originated from edge sites. As shown in Figure 3d,e, the cubes of 44 nm benefited from an optimal balance between plane and edge sites presented the highest FE for C 2 H 4 (41%), which was superior to 24 and 63 nm cubes tested. [28] Likewise, Strasser et al. prepared smaller Cu nanoparticles in the size range of 2-15 nm. They found that the explosion of undercoordinated atoms (coordination number [CN] < 8) on below 2 nm Cu NPs induced strong interaction with intermediates, which was unfavorable for the production of C 2 H 4 . However, the weaker binding of CO and H on 5-15 nm Cu NPs was suitable to the formation of C 2 H 4 . [29] Furthermore, ultrafine (%2 nm) Cu NPs, stabilized by alkyne moieties from pyrenyl-graphdiyne (Pyr-GDY), are efficient for the formation of C 2 products via modulating binding strength between CO* and catalyst. [30] What is more, surface reconstruction on copper catalysts offers an efficient strategy for CO 2 conversion performance recently. [31][32][33] Sargent et al. [31] reported a copper-chloride-derived catalyst for tuning selectivity of C 2 H 4 . A CuCl film was grown on an electropolished Cu foil substrate and converted to a Cu(I) oxide surface by the method of wet-oxidation ( Figure 4a). Reconstructed Cu showed high FE(C 2 H 4 ) (>50%) in potential range À1.9 to À2.4 V versus Ag/AgCl, while FE(CH 4 ) was suppressed down to 5% instead of FE(CH 4 ) > 40% on electropolished Cu ( Figure 4b). As shown in Figure 4c, the FE comparison of other products suggested that surface reconstruction process played a key role to shift the pathway from CH 4 to C 2 H 4 , which may transfer the facet from Cu (111) in electropolished substrate to Cu (100). Recently, grazing incidence X-ray absorption spectroscopy (GIXAS) and X-ray diffraction (GIXRD) were applied to uncover the structural evolution of polycrystalline Cu electrodes and structure-function relationship in electrocatalysis process (Figure 4d,e). [32] The results revealed that polycrystalline Cu maintained the metallic state during the CO 2 RR and preferred reconstructing toward (100) facets only under CO 2 atmosphere. With increasingly negative applied potential, the surface reconstruction was enhanced while the production of C 2 H 4 was increased consequently. Besides, benefited from high mobility of S atoms, Xiong et al. [33b] took advantage of polycrystalline Cu nanoparticles with rich high-index facets donated from Cu 2Àx S. Special active sites emerging during surface reconstruction were effective for the C─C coupling and exhibited an excellent performance for the formation of C 2 H 4 . Apart from single Cu sites, alloying Cu with other metal also has been demonstrated to be helpful strategies to break scaling relationship for tuning the interactions between catalysts and intermediates. [34,35] Previous researches reported that a strong bonding energy between Cu and key intermediates was favored for the formation of hydrocarbon because of adjusted adsorption energetics of alloys. For example, Qiao group [34] summarized how the O affinities and H affinities of the metal affect the selectivity of products ( Figure 5a). Cu─Ag/Au nanoframes [36] were prepared to surmount the kinetic barriers of CO 2 conversion. It was satisfying to realize a high FE(C 2 H 4 ) of 69 AE 5% by means of decoupling the functions of CO formation (promoted by Ag/Au) and C─C coupling for C 2 H 4 product (promoted by Cu). The operando IR spectroscopy and DFT calculations were responsible for favored CO dimerization pathway based on close coupling between Cu 3d and C/O 2p orbitals at more negative bias (Figure 5b-d). Similarly, Buonsanti et al. [37] synthesized Ag-Cu nanodimers to realize tandem catalysis induced by the presence of Ag and Cu as segregated domains for the improvement of *CO dimerization and achieved remarkable selectivity of C 2 H 4 (Figure 5e-f ).
Additionally, there are other effective strategies, like surface functional modification and the regulation of electrolytes. Modulate adsorption energy of intermediates by modifying the surface of catalysts can suppress the proton transfer to a certain degree. [38] And the formation of C 2 H 4 depended on pH value and halide species on copper catalysts. [39] The majority of study implied that the morphology of copper catalysts and ion-metal interaction are pH-dependent in terms of halide species enhancing C 2 H 4 products. [40] Peng et al. identify the carbon and hydroxide binding strengths as descriptors to unveil pH dependence in dicarbon oxygenate/hydrocarbon selectivity, which is related to dehydroxylation step to remove O atoms from C 2 HC pathway. [41] 2.2. CO 2 to Ethanol As shown in Figure 2, the reaction mechanisms that produce ethylene and ethanol are somewhat different, so researchers may tend to focus on a little different strategies for Cu-based electrodes, such as atomic coordination regulation, bimetallic synergy, and interface regulation.
It is believed that *CO is the first step of C─C coupling as key intermediate, which CO adsorption state is related to the valence state and configuration of Cu. Recently, Wang group [42] explored the true active sites during CO 2 RR of CuO clusters supported on the N-doped carbon nanosheets by complementary some operando spectroscopy (like XAS, X-ray photoelectron spectroscopy (XPS), and FTIR), X-ray absorption near edge structures (XANES) and DFT calculations, as shown in Figure 6a. It proved a structure transmission from CuO clusters to Cu n -CuN 3 site under potential application and Cu/N 0.14 C showed high activity Figure 3. a) Energy profiles of CO dimerization on three facets of copper. b) Surface energy changes in connection of the surface coverage of intermediates. c) Adsorption energies of intermediates on three facets of copper. Reproduced with permission. [27] Copyright 2019, Springer Nature. d) Models of spherical Cu NPs with 2.2 and 6.9 nm diameters. Reproduced with permission. [28] Copyright 2016, John Wiley and Sons. e) Population (relative ratio) of surface atoms with a specific CN in connection of particle diameter. Reproduced with permission. [29] Copyright 2014, American Chemical Society.  [44] Attenuated total reflection Fourier-transform infrared (ATR-FTIR) ( Figure 6b) and DFT calculation revealed that the valence state of Cu well maintained between 0 and þ1 under low applied potential, which decrease ΔG PDS of C─C coupling toward alcohol. Furthermore, Jung et al. [45] synthesized porous Cu/Cu 2 O aerogel through a simple wet chemistry method under Ar atmosphere, which enhanced the adsorption of CO on Cu 0 and Cu þ between densely Cu 0 -Cu þ interfaces and exhibited a ethanol FE of 41.2% (Figure 6c,d).
Similarly, the strategy of achieving bimetallic synergies by alloying also has been used to improve the selectivity of ethanol. Hydrodeoxygenation process can be tuned by introducing another active site on the Cu surface that promotes the path way to ethanol. CuZn alloys were prepared through cosputtering strategy by Oh et al. [46] With Zn contents increasing, CO accumulated and ethanol production was enhanced. As a result, Cu 9 Zn 1 achieved a stable ethanol FE of %25% and an ethanol full-cell energy efficiency (EE) of %11%. In the latest Xiong's work, [47] they synthesized amorphous CuTi alloy with coordinatively unsaturated Cu sites, which achieved the electrocatalytic CO 2 reduction reaction to produce alcohol with a FE of 48.82% at À0.8 V versus RHE. It can be seen from Figure 7a-c that Ti has lower electron affinity than Cu that the electron could transfer to Cu from Ti to facilitate the coupling of *CO intermediates. Besides, silver-modified Cu 2 O electrocatalyst was prepared through a newly one-pot seed-medium method by Qiao's group, [48] which exhibited an ethanol FE of 40.8% and partial EtOH current density with 326.4 mA cm À2 . In situ the attenuated total reflection-infrared reflection absorption spectroscopy  [31] Copyright 2018, John Wiley and Sons. d) Schematic illustration of study process observed by GIXAS and GIXRD. e) Ratio of the area of Cu (200) to Cu (111) Bragg peaks at a probe depth of 2.6 nm in CO 2 -and Ar-purged electrolytes as a function of the applied potentials. Reproduced with permission. [32] Copyright 2021, American Chemical Society. (ATR-IRAS) spectroscopy (Figure 7d,e) revealed that the coordinated number and oxide state of surface Cu sites were optimized as a result of redispersion of Ag in Cu. Therefore, the binding strength of *CO was tuned and C─C coupling was enhanced to produce ethanol. Furthermore, CO 2 RR is an interfacial dependent reaction, so putting some effort into improving the catalyst interface will be an effective strategy. Wang's group developed an interface structure strategy to fabricate Cu@N-doped graphene interface by template-directed chemical vapor deposition (CVD) and following nitrogen doping. [49] The accumulation of *CO intermediate induced by the interaction of N─G layer and CO 2 molecule with the N─G coating layer enhances formation of key intermediate *HCCHOH. Meanwhile, the average valence state of Cu is between 0 and þ2, which not only promotes C─C coupling to producing ethanol, but HER is suppressed by the strong adsorption between H and Cu. Therefore, Cu@N-doped graphene exhibits the FE of ethylene was 15.2% at À1.0 V (vs RHE). In addition, metal-organic frameworks (MOFs) are usually used to anchor transition metal species and produced interface structure. A latest work designed a MOF-derived CO 2 RR electrocatalyst by one-pot pyrolysis of Cu-based HKUST-1 to synthesis carbon-coated CuO x @C. Operando Raman spectra and DFT calculations revealed that C─C coupling was enhanced by the stable Cu þ species with the carbon coating under high current density. CuO x @C shows ethanol FE of 46% while the partial current density of ethanol reached impressively 166 mA cm À2 . [50] Heterostructure between different metals are also discussed. Hu et al. [51] synthetized the morphology tunable electrocatalyst AuAgCu with heterointerface structure by a three-step seedmediated growth strategy. Transmission electron microscopy (TEM) and high resolution transmission electron microscopy (HRTEM) proved that heterostructure of AuAgCu was prepared successfully. The obtained Au 1 Ag 1 Cu 5 exhibited a FE of  [34] Copyright 2018, Elsevier. b) CO 2 RR performance of Cu 3 -Ag 3 Au tested in 0.1 M KHCO 3 aqueous solution. c) Partial density of states (PDOS) of the proposed structure with absorbed CO. d) Operando attenuated total reflectance infrared spectra (operando ATR-IR spectra) recorded while ramping down the potential from À0.4 to À1.2 V versus RHE on Cu 3 -Ag 3 Au NFs. Reproduced with permission. [36] Copyright 2021, John Wiley and Sons. e) Schematic representation of the proposed mechanism of C 2 H 4 promotion in the Ag-Cu NDs. f ) FE for C 2 H 4 obtained on different Ag/Cu nanocrystals at À1.1 V versus RHE. Reproduced with permission. [37] Copyright 2019, American Chemical Society. 37.5% toward alcohol owing to electronic effects and nanoheterostructure.
To sum up, there are complicated reaction steps and numerous intermediates so that it is challengeable to obtain satisfactory reactivity and selectivity of C 2 products for Cu-based catalysts. Currently, the *CO dimerization generated from the surface of copper catalysts has been demonstrated as the prerequisite step for C 2 products by means of theoretical calculation. [52] Figure 6. a) Proposed scheme for the reversible formation of the catalytically active Cu n -CuN 3 cluster. Reproduced with permission. [42] Copyright 2022, Springer Nature. b) Potential-dependent ATR-FTIR spectra of Cu 2 S 1Àx HN andCu 2 S HN with CO 2 bubbling. Reproduced with permission. [44] Copyright 2022, Springer John Wiley and Sons. Characterization of Cu 0 -Cu þ interfaces by Cu and Cu 2 O domains. c) XRD patterns and d) magnified XRD peaks of aerogel catalysts with varying amounts of NaBH 4 . Reproduced with permission. [45] Copyright 2021, Springer John Wiley and Sons. The purple, orange, red, and gray spheres represent Ti, Cu, O and C atoms, respectively. c) In situ Raman spectra for reaction intermediates on a CuTi@Cu catalyst at the potential of 0.8 V. Reproduced with permission. [47] Copyright 2021, John Wiley and Sons. In situ ATR-IRAS obtained during chronopotentiometry in a potential window 0.2 to À1. Subsequently, there are multipath cross-coupling possibilities on hydrodeoxygenation. Other opinions imply the coupling of adsorbed *CHO or CHO─CO coupling, CO─COH coupling, and CO─C coupling which are supported by in situ characterization methods and theoretical calculations. [24,25,40a] In order to construct realistic electrochemistry environment as fully as possible, theoretical calculations have applied multiple methods and models such as computational hydrogen electrode, constant electrode potential, and implicit/explicit solvation model from the perspectives of thermochemistry, kinetic barriers, or fully integrating the effects of solvation. But some works heavily rely on the selected methods and assumptions resulting in difficultly meeting the demands toward fine study of reaction mechanisms in multiphase electrochemistry. Constant-potential hybrid-solvation dynamic model offers a promising example to take into account dynamic evolution of catalyst structure under working conditions. [53] With the development of technology, it is benefited to introduce machine learning to achieve time-saving and profit-increasing. Therefore, it can be expected that a combination of varieties of feasible strategies, in situ experiments used to detect reactive intermediate species, and rational theoretical calculations will take the mechanism of complex reactions to the next level.

CO 2 to Oxalic Acid
Actually, oxalic acid generated from a two-electron transfer reaction exhibits significant instruction to deeply understand mechanisms for the formation of C─C bonds. CO 2 has a high solubility in nonaqueous aprotic electrolytes instead of a lower level of energy conversion efficiency toward CO 2 RR in aqueous electrolytes. What is more, the presence of H 2 O inevitably leads to the hydrogenation reaction toward formate. Thus, it will be an attractive alternative that using the solvents with proton-poor as electrolyte. [54][55][56] Hori et at. [56] demonstrated that CO 2 anion radicals (CO 2 ·À ) considered as an intermediate species were in connection with CO covered Pt electrode for the formation of oxalic acid through self-coupling of CO 2 ·À in tetraethylammonium perchlorate acetonitrile-water mixtures. A combination of chronoamperometry and infrared reflectance spectroscopy confirmed the formation of oxalic acid at Pb leads cathode, while a band at 2037 cm À1 pointed to *CO intermediate or CO did not appear at À 2.5 V versus Ag/AgCl during 300 s (Figure 8a). [54] It implied the metals of sp group were capable to be responsible for dimerization of CO 2 in TEAP-PrC solution like Pb leads. Weber et al. [55] investigated the reductive process of converting CO 2 to oxalic acid on an atomic bismuth model catalyst under aprotic conditions. IR photodissociation spectroscopy revealed [Bi(CO 2 ) n ] À cluster ions on catalyst and complex involved bidentate (η 2 ) interactions between CO 2 and Bi may be a key intermediate for C─C coupling (Figure 8b). These works paved the way for further studies on producing oxalic acid via CO 2 RR. Currently, many efforts were made on screening a suitable electrolyte system mainly focused on nonaqueous aprotic electrolytes [54][55][56][57] and involved CO 2 ·À intermediate or *COOH intermediate for synthesis pathway. [57a] Interestingly, a work on a Cr─Ga thin film realizes the formation of oxalic acid in an aqueous electrolyte and propose a new mechanism which avoids energy-intensive CO 2 À intermediate. [58] There is no doubt that a different pathway to produce oxalic acid from aqueous CO 2 needs to be further explored. Consequently, it is promising to compare and analyze the similarities and differences over existing C─C coupling mechanisms as well as enrich the understanding of C─C coupling to oxygenated Figure 8. a) Spectra of the reduction of CO 2 -saturated in 0.2 M TEAP-PrC on a Pb electrode recorded during chronoamperometry measurements at À2.5 V versus Ag/AgCl. Reproduced with permission. [54] Copyright 2010, Elsevier. b) Comparison of experimental and calculated IR spectra. Experimental spectrum (black). Calculated spectrum (red/blue). Reproduced with permission. [55] Copyright 2016, John Wiley and Sons.
www.advancedsciencenews.com www.advenergysustres.com compounds (such as CH 3 CH 2 OH, CH 3 COOH, and C 3 H 7 OH) in turn. In a word, there is a long way to figure out mechanisms of converting CO 2 to oxalic acid.

C─N Coupling to Organic Nitrogen Compounds
In 2019, Jiao group found that acetamide products could be obtained by coelectrolysis of CO and NH 3 to form C─N bonds on copper electrodes. And the acetamide derived from nucleophilic attack of NH 3 to *C═C═O, a key intermediate of CO 2 / CO reduction. [11a] Following this, Wang et al. achieved C─N coupling from CO 2 and N 2 on Pd 1 Cu 1 /TiO 2 catalyst to form urea. [11b] They proposed that the possible mechanisms for C─N bond formation were a thermodynamic spontaneous reaction between *N═N* and CO. According to the above we can find that the whole process involves not only electrochemical processes but nonelectrochemical processes. Especially, it is fulfilling and desirable to integrate the N atom into the electrocatalytic CO 2 reduction. Thereinto, the N-containing feedstocks can be N 2 , NH 3 , nitrates/nitrites, and so on. Considering the increasing attention of researchers on urea, amide, and amine organic compounds in recent years, we will focus on two parts: the electrocatalytic coupling of CO 2 and nitrogen sources (including N 2 , NO 3 À , NO 2 À ) to urea and amide/amine organic compounds.

CO 2 and N Source to Urea
Urea is the nitrogen fertilizer with the highest nitrogen content, and the usage is also a lot because of preservation easily and little damage to the soil. Whereas conventional industrial urea synthesis is dependent on the reaction of NH 3 and CO 2 at high temperature and pressure (150-200°C, 150-250 bar), [59] while the acquisition of ammonia is also through the Haber-Bosch process at high temperature and pressure (400-500°C, 200-300 bar). [60] As can be seen from aforementioned information, the industrial synthesis of urea often requires harsh reaction conditions. The electrochemical coupling of CO 2 and N 2 driven by renewable electricity, by contrast, can achieve the synthesis of urea under ambient temperature, which is of great interest. In addition, without the reduction step from N 2 to NH 3 , the C─N coupling between N 2 and carbon intermediates can directly form urea product, effectively reducing the reaction energy barrier. This electrochemical method has the advantages of simple operation, high reaction efficiency, and little secondary pollution. [61] Researchers found that nitrate and nitrite also can be electrochemically coupled with CO 2 to form urea, apart from N 2 . As shown in Table 2, we summarized the study on the electrosynthesis of urea by CO 2 and various N reactants (like N 2 , NO 3 À , and NO 2 À ) over different catalysts. Comparative analysis of these performance and proposed reaction mechanisms will benefit the understanding of C─N coupling process, which is great significance for further optimizing the reaction system in the future. www.advancedsciencenews.com www.advenergysustres.com

N 2 as the N Source
Nitrogen is up to 78% in the atmosphere, but nitrogen is inert and difficult to activate. As mentioned above, urea from N 2 generally requires very demanding reaction conditions such as high temperature and pressure. Researchers have applied external forces to assist synthesis of urea, such as plasma [62] and electromagnetic fields. [63] The energy consumption, however, cannot be negligible. Kayan et al. [64] first studied the simultaneous electrocatalytic reduction of CO 2 and N 2 (30 bar CO 2 þ 30 bar N 2 ) on polyaniline (PANi)-and polypyrrole (PPy)-coated platinum electrodes. Whereafter, as shown in Figure 9a,b, Wang group achieved relatively high urea formation rate (0.12 mmol g À1 h À1 ) and FE (8.92%) at À0.4 V (vs RHE) by constructing PdCu alloy nanoparticles on TiO 2 .
[11b] Yuan et al. proposed to construct heterogeneous catalysts for simultaneous adsorption and activation of raw gas. [65] There are electrophilic and nucleophilic sites in the constructed Bi/BiVO 4 heterojunction structure (Figure 9c,d), which can adsorb the electron-rich N in N 2 and the electron-poor C atom in CO 2 . The urea yield was 5.91 mmol g À1 h À1 , and the FE was 12.55% at À0.4 V (vs RHE). Similarly, they demonstrated that perovskite BiFeO 3 /BiVO 4 showed higher FE of 17.18%. [66] Moreover, they designed unique frustrated Lewis pairs (FLP) in Ni 3 (BO 3 ) 2 catalysts, where the surface hydroxyl and adjacent Ni site are Lewis base and acid sites. The two sites can facilitate the adsorption and activation of CO 2 and N 2 , which can be attributed to the interaction of bonding and antibonding orbitals of feedgas with the vacant orbital of Lewis acid and the nonbonding orbital of Lewis base, respectively. They also proposed that the FLP idea could be applied to the InOOH catalyst. [67] There are not only electrochemical processes involved in the C─N coupling of CO 2 and N 2 to produce urea, but also nonelectrochemical processes. Therefore, in order to achieve high yield of urea, it is necessary to further study the specific reaction mechanism, which requires experimental and theoretical research. Kayan et al. proposed that NH 4 þ from N 2 reduction reacted with CO 2 to ammonia carbamate, but there was no solid evidence for this. [64] Based on operando synchrotron-radiation Fourier transform infrared spectroscopy (SR-FTIR), Wang group observed a peak at about 1449 cm À1 slowly appeared with increasing potential, which can be attributed to C─N stretching vibration, indicating the production of urea to a certain extent (Figure 10a). [11b] They also found the N 2 adsorption with a side-on configuration was conducive to reduce the CO 2 to CO at adjacent metal sites. The molecular orbitals of CO matched with those of *N═N* to form *NCON*, a key reaction intermediate. As shown in   Figure 10b, the competition reaction to *NNH species required very high energy barrier, presenting that the ammonia served as by-products was suppressed. Subsequent hydrogenation on *NCON* follows two alternative pathways of distal hydrogenation and is spontaneous ( Figure 15, path I). The same mechanism was supported in several works by Yuan et al. [65,67] Unfortunately, *NCON* intermediates have not been captured experimentally. Therefore, further studies are needed to validate the proposed C─N coupling step.

Nitrate/Nitrite as the N Source
In addition, researchers have also considered other nitrogen sources as a source of nitrogen in urea, for example, there are often a lot of nitrite or nitrite in the industrial wastewater and domestic sewage. [68] Therefore, they tried to achieve C─N coupling by co-reduction of CO 2 and nitrate/nitrite. In 1995, Shibata et al. first obtained urea with Cu-loaded gas diffusion electrode (GDE) at 1 bar CO 2 and 0.02 M NO 3 À /NO 2 À (À0.75 V vs SHE). [69] Meanwhile, they found that CO from CO 2 reduction and ammonia derived from nitrite were the key species for urea production. The team then further investigated the urea production from CO 2 and NO 3 À /NO 2 À with various electrodes, including metal, [70,71] metal boride, [72] and metallophthalocyanine (MPc) [73] catalysts, and the specific results are shown in Table 2. However, the ability of these catalysts to reduce NO 3 À to NO 2 À and NH 4 þ is poor, hindering the electrochemical synthesis of urea. Therefore, it is necessary to reduce carbon dioxide and nitrate ions simultaneously in this coupling reaction. Years later, the line of research has attracted more attention from researchers. Te-doped Pd nanocrystals (Te-Pd NCs, Figure 11a) were applied for production of urea by coupling CO 2 RR with electrochemical reduction of NO 2 À . [74] As shown in Figure 11b, 12.2% FE and 88.7% N atom efficiency (NE) at À1.1 V (vs RHE) were obtained over Te-Pd NCs, higher than pure Pd NCs. There was electron transfer from Te to Pd, facilitating CO 2 adsorption and ammonia production ( Figure 15, path III). Subsequently, AuPd nanoalloy also showed outstanding catalytic performance ( Figure 15, path IV). [75] B-FeNi-DASC achieved a high urea yield rate of 20.2 mmol h À1 g À1 and FE of 17.8% as a consequence of efficient coordinated adsorption and activation of multiple reactants by the bonded Fe─Ni pairs in C─N coupling process( Figure 15, path V). [76] Moreover, defect engineering is one of the most effective strategies for catalyst design. Unsurprisingly, researchers have applied this strategy to urea synthesis. Metal oxides and hydroxides with oxygen vacancies showed prominent catalytic performance in the electrosynthesis of urea, including Cu-doped TiO 2 ( Figure 15, path III), [77] ZnO ( Figure 15, path VI), [78] CeO 2 ( Figure 15, path VII), [79] and InOOH ( Figure 15, path VIII). [80] Thereinto, the urea FE of ZnO-V achieves 23.26% (0.79 V vs RHE), about 3 times as high as that of ZnO (8.10%) (Figure 11c,d). [78] Additionally, the production rate of urea over ZnO-V is high to 16.56 mmol h À1 (Figure 11e). They believe that the oxygen vacancies accelerate CO and NH 3 precursors. This year, Wang's group also demonstrated that the oxygen vacancyenriched CeO 2 was an efficient electrocatalyst for producing urea. [79] Besides, high urea yield of 533.1 μg h À1 mg À1 and high FE of 53.4% on In(OH) 3 nanocubes (denoted as In(OH) 3 -S) were obtained at À0.6 V (vs RHE) by Yu et al. (Figure 11f-g). [11c] The (100) facets are more conducive to achieve direct C─N coupling of *NO 2 and *CO 2 intermediates to obtain urea. Compared with the system of N 2 as the N source, two N atoms are not bond with one C atom simultaneously in co-reduction of CO 2 and nitrate/nitrite for urea, which is a crucial consensus. [9c] There are two main categories from the perspective of intermediated species from CO 2 RR and NO 3 RR/NO 2 RR during the first C─N coupling. Feng et al. (Figure 12a) suggested that CO 2 RR to *CO, NO 2 RR to *NH 2 , *CO, and *NH 2 are intermediate species for *CONH 2 formation by some control experiments. [74] Zheng's report showed the same possible mechanism. [77] Meng et al. (Figure 12b) offered a slightly different result on the urea formation pathway in combination with the observation by online differential electrochemical mass spectrometry (DEMS) and in situ ATR-FTIR. [78] Similarly, NO fixed on the bridge sites of the bonded Fe─Ni pairs can be reduced into *NH and then couples with *CO. [76] Interestingly, on the other reported Vo-deficient electrocatalyst, conducive nitrogen oxides intermediated species  participated in the urea synthesis seem to be accepted both experimental results and theoretical calculations (Figure 12c-e). As Wang group demonstrated, oxygen vacancies can stabilize *NO intermediate, which favors subsequent C─N coupling over protonation. [79] Yu et al. found *NO 2 as a thermodynamically spontaneous product from NO 3 À coupling with *CO 2 to form *CO 2 NO 2 on the In atoms adjacent to the V o (Figure 12fg). [80] The mechanism of electrochemical urea synthesis from coupling of carbon dioxide and nitrate/nitrite ions is still disputable due to the complexity of multiple surface reactions. Currently, eight thermodynamically feasible C─N coupling mechanisms between the key intermediates from CO 2 RR (*CO 2 , *CO, *COOH) and the ones from NO 2 /NO 3 RR (*NO 2 , *NO, *NH 2 OH, *NH, *NH 2 ) are shown in Figure 15.

CO 2 with N Source to Amides and Amines
It is well known that amides and amines are very crucial organic compounds in pharmaceutical chemistry, agriculture industry, and so on. [81] Currently, amides/amines, synthesized through the reaction of carbon monoxide/alcohols and ammonia under high-temperature and high-pressure conditions, are www.advancedsciencenews.com www.advenergysustres.com energy-and resource-intensive process. [82] It is a great challenge to explore the solution of an alternative ambient and environmentally friendly approaches. Benefited from some advanced electrocatalysts which realize C─N coupling under room temperature and pressure conditions as shown in earlier section, electrocatalytic co-reduction of CO 2 and N-integrated to value-added chemicals offers possible alternatives to the thermochemical synthesis routes of amides and amines. We sum up reports on expanded product scope of C─N coupling process beyond urea in Table 3. Formamide, acetamide, methylamine, ethylamine, and other N-methylamine products are included. We focus on the possible mechanisms as well as differences and similarities of the key step between urea and amides/amines aiming to inspire new explorations.

Electrosynthesis of Amides
The formation of urea via C─N coupling shows disparate mechanisms which depends on N-source. Consequently, suitable N-integrated may be synthesized various chemicals based on C─N moiety. Jiao et al. [11a] creatively proposed C─N formation pathway from CO and NH 3 coreduction (Figure 13a). Full-solvent quantum mechanical calculations demonstrated that under Figure 12. a) Scheme of the urea synthesis from CO 2 RR and NO 2 RR on Te-Pd NCs. Reproduced with permission. [74] Copyright 2020, American Chemical Society. b) In situ ART-FTIR spectra of ZnO-V under CO 2 , NaNO 2 , and both. Reproduced with permission. [78] Figure 13. a) Scheme of the acetamide synthesis from CO and NH 3 on copper catalyst. b) The mechanism for CO reduction on Cu. Reproduced with permission. [11a] Copyright 2019, Springer Nature. In situ surface enhanced infrared absorption spectroscopy in the attenuated total reflection mode (ATR-SEIR) using formic acid with different N-sources: c) 14 NO 2 À and d) 15 NO 2 À . e) Quasi-in situ EPR trapping of carbon radical. Reproduced with permission. [84] Copyright 2022, American Chemical Society. f ) The performance of formamide electrosynthesis as different current densities over Pt and g) the performance over Pt and Pt-Ti at 100 mA cm À2 . Reproduced with permission. [85] Copyright 2022, Springer Nature.
www.advancedsciencenews.com www.advenergysustres.com The selectivity of acetamide achieves nearly 40% at industrial reaction rates. Moreover, they successfully tried to expand product scope to N-methylacetamide, N-ethylacetamide, N,N-dimethylacetamide, acetic monoethanolamide, and aceturic acid (Figure 13b). Subsequently, a sustainable technology was developed to couple CO 2 and liquid phase NH 3 at a gas-liquid-solid boundary over commercial Cu catalyst which is generally considered as a great choice for binding C-based intermediate species appropriately (Figure 15, path IX and XI). [83] With partial current densities ranging from 0.1 to 1.2 mA cm À2 , the peak of FEs for formamide and acetamide at approximately 1%. In addition, an original operando infrared spectroscopic method realized to estimate the reaction pathway. The conclusion is in accord with nucleophilic attach mechanism. As is known to all, it is relatively efficient for formic acid and methanol from CO 2 electroreduction at an industrial current density. Thus, there are practical roots to upgrade downstream products of CO 2 electroreduction into amides, which is helpful to establish efficient tandem process and recycle carbon footprint. Recently, Zhang group [84] reported an electrochemical process on co-reduction of formic acid and nitrite in order to produce formamide over copper catalyst, which offered the selectivity of formamide to 90% and FE of 29.7% at À0.4 V (vs RHE). *CHO (1720 cm À1 ) and *NH 2 (1155 cm À1 ) were confirmed to be the key intermediates by in situ ATR-FTIR ( Figure 15, path X). Quasi-in situ electron paramagnetic resonance (EPR) detected the sextet peaks indexed to carbon radicals implying the formation of *CHO (Figure 13c-e). At the same time, they exhibited an economic advantage approach to transform methanol and ammonia into formamide via Pt electrocatalyst, which can provide a FE of 40.39% at 100 mA cm À2 and operate stably for 46 h in the flow cell (Figure 13f-g). [85] It was proved that an aldehyde-like intermediate (*CH 2 O) could combine with *NH 2 intermediate to C─N coupling via feedstock replacement experiment. In situ ATR-FTIR and online DEMS were carried out for a real-time monitor to further understand the catalytic process. In general, a series of meaningful studies have demonstrated the successful strategy of C─N coupling for amides while production efficiency still needs to be promoted toward future application in terms of paralleling competitive reactions and technological process.

Electrosynthesis of Amines
Amine, one of value-added product from CO 2 and N-integrated by electrosynthesis, is of great importance for C─N coupling product scope expansion and mechanistic studies. Limited by formidable N≡N bond energy (942 kJ mol À1 ), the researches mainly focus on sustainable alkylamine synthesis converting CO 2 and nitrate/nitrite to amines. Wang group first reported that a cobalt β-tetraaminophthalocyanine molecular catalyst supported on carbon nanotubes triggered CO 2 and nitrate into methylamine in aqueous media under ambient conditions as well as put forward pathbreaking reaction pathway of the eight steps (Figure 14a and 15, path XII). [86] The overall FE of methylamine was achieved for 13% at À0.92 V (vs RHE) and maintained for at least 16 h (Figure 14b,c), giving a total turnover number (TON) of 5600. It was elucidated that there were 14 electrons and 15 protons involved to form methylamine, where the key C─N bond-forming step is the nucleophilic attack between formaldehyde (HCHO*) from CO 2 RR and hydroxylamine (NH 2 OH) from NO 3 RR (Figure 14d). Systematic control experiments played an important role to identify intermediates. They further expanded the scope of the reaction to yield N-methylamines from CO 2 and abundant nitro compounds, which was based on a descriptor for the nucleophilicity of N-source summarized by their studies; nitrogen substrates including aliphatic, aromatic, primary, and secondary amines, hydrazine, and hydroxylamine were included (Figure 14e). [87] There was dependence of amines selectivity on the concentration and nucleophilicity of N-source via a variety of experiments. Moreover, it was also demonstrated that high nucleophilicity of N-source was crucial to steer the reaction pathway toward chemical condensation from HCHO* and hydroxylamine intermediate, formed from four-electron reduction of the nitro compound, instead of the electrochemical reduction of HCHO* to CH 3 OH (Figure 14f ). A CuO-derived Cu electrocatalyst successfully yielded ethylamine from CO 2 and NO 3 À -integrated by spontaneous condensation between acetaldehyde and hydroxylamine intermediates which was consistent to the methylamine work ( Figure 15, path XIII). [88] It is promising to apply "nucleophilic attack strategy" to produce many valuable chemicals while balance competitive reactions with chemical condensation. In addition, the revelation of interaction between active sites and reactants is necessary and helpful to rationally develop efficient catalysts in the future.
There is the same O═C─N moiety of key intermediates in the pathways of urea and amides (In fact, urea belongs to a special amide.). Benefited from inherent structure of N 2 , two N atoms are bond with one C atom simultaneously in path I and path II, which prevents from the second C─N coupling when NO 3 À / NO 2 À as N source. And the next step involved in two competitive reactions in path III-XI, the second C─N coupling or hydrogenation for O═C─N moiety, will determine the final products toward urea or amides. As for amines, nucleophilic attack between formaldehyde (HCHO*) from CO 2 RR and hydroxylamine (NH 2 OH), playing a significant role in C─N coupling, removes H 2 O resulting in C═N─O moiety and then forms the products via hydrogenation. It is apparent that more considerate works need to be explored to expand and strength the comprehension of the mechanism in C─N coupling and modulate the selectivity.
To sum up, because C─N coupling is based on C-intermediate bonded with N-intermediate simultaneously, dual-site catalysts (generated from alloying, oxygen vacancy-rich oxides, heterostructure) provide a promising paradigm of activating CO 2 and N-integrated, respectively, as well as a suitable platform to complete C─N bond. It should be noted that appropriate interaction between active sites and C/N-intermediates is favor of high selectivity over C─N compounds. Otherwise, over wake/strong adsorption leads reactions to the direction of high C─N coupling barrier energy and totally self-reduced products.

Electrolytic Reactors
There are usually relatively complex reactions during the electrocatalytic reduction, and the reaction apparatus is vital and www.advancedsciencenews.com www.advenergysustres.com significant for the accurate evaluation of catalyst performance. An ideal electrolytic reactor should have a satisfactory function of facilitating electron and ion transport, promoting the controlled diffusion of gas, and preventing electrolyte mixing between the anode and cathode. [2a,89,90] At present, these electrocatalytic reaction devices are widely used, as shown in Figure 16, which are three-electrode diaphragm electrolytic cells (H-type electrolytic cells), three-electrode three-chamber flow cells, membrane electrode assembly (MEA) cells, and electrolytic cells with solid-state electrolyte. The following is a brief introduction to these typical electrolyzers.
The H-type electrolytic cell is a relatively simple and very mature device, as shown in Figure 16a, which can be separated from the anode and cathode by a proton exchange membrane (such as Nafion membrane) to prevent the products generated by the anode and cathode reactions from interfering with each other. [11b] Moreover, the cell is usually relatively closed to reduce the influence of the surrounding environment, and then reduce the error in the evaluation of catalytic performance.
However, the limitations of H-type electrolytic cells are gradually exposed with the deepening of research. The reactant gases are introduced to catalysts by the means of being bubbled. The limited solubility of feed gas (such as CO 2 , N 2 ) in the electrolyte, coupled with limited material transport between the working electrode and the counter electrode due to the poor diffusivity, makes the current density too low to meet the needs of practical applications or industrial-relevant studies (current densities > 200 mA cm À2 ). [91] Inspired by the devices of fuel cells, researchers further developed flow cells based on three-electrode three-chamber GDEs (Figure 16b). The three chambers are gaseous cathode chamber for feed gas flow, liquid cathode, and anode chambers for electrolyte flow. Different from the H-type electrolytic cells, the electrocatalytic reduction reaction in the flow cell occurs at the gas-liquid-solid three-phase interface. In a typical flow cell, feed gases can be directly delivered to the cathode and react at the interface through a GDE, which evites the mass-transport limitation in the H-type electrolytic cells. And the flow cells, a high c) Product distribution and stability at À0.94 V versus RHE. d) The reaction pathway of the eight steps. Reproduced with permission. [86] Copyright 2021, Springer Nature. e) A function of the nucleophilicity of the N-source according to the Mayr scale. f ) The reaction scheme with sample 1 H NMR spectrum. Reproduced with permission. [87] Copyright 2021, American Chemical Society.
www.advancedsciencenews.com www.advenergysustres.com activity benefited from lower cell resistance and cell voltages, have been demonstrated to achieve high current densities. On this basis, researchers found that the hydrophobicity of GDE tends to deteriorate with the extension of electrolysis time, and finally destroy the three-phase catalytic interface, thus affect the catalytic activity. To solve the problem of catalytic stability, researchers remove the cathode liquid chambers of original flow cells and derive the MEA cell (Figure 16c). [89,92] Thereinto, water vapor in feed gas as a proton source participates in the reduction reaction. In MEA electrolytic cell, the serpentine gas channels can increase the contact between reaction gas and catalyst interface, obtaining high reaction rates and energy efficiency. [93] Finally, this electrolyzer not only improves the catalytic stability of the electroreduction reaction, but also facilitates the separation of gas products. The device is popular in the study of electrocatalytic reduction reactions to date. Unfortunately, a thin gas film, always generated from the product gas on the surface of electrodes, keeps the electrolyte away from and affects the reduction Figure 15. Proposed mechanism of urea, amides, and amines from the co-reduction between CO 2 /HCOOH and N 2 /NH 3 /NO 3 À /NO 2 À .
www.advancedsciencenews.com www.advenergysustres.com reaction further. [93b] In addition, cation-exchange membranebased MEAs tend to enhance HER in a highly acidic reaction environment, [92] which triggers the exploration of bipolar (BPM) or anion-exchange membranes (AEM). [94] Apart from the common electrolyzers mentioned above, researchers have also designed some new-fashioned electrolyzers to achieve higher current densities and solve crossover problems. For instance, electrolyzers with solid polymers (SPE) or oxides (SOE) as electrolytes have been developed. As shown in Figure 16d, Wang et al. [95] have developed a flow cell with a cathode filled with a conducting polymer electrolyte. Finally, they achieved the preparation of pure liquid formic acid at a lower overpotential, which greatly improved the economy of the preparation of formic acid from CO 2 RR. But low CO 2 conversion efficiency is not satisfying. In addition, various components of the flow cell still lack to be optimized, including fabrication of GDE, membranes, flow field, and so on. The development optimization of electrocatalytic reduction reaction apparatus is an essential step in the industrial application of electrocatalytic reaction. It is believed that electrocatalytic reactions will eventually be industrialized in the near future by integrating knowledge from many disciplines (e.g., engineering, chemical engineering, chemistry, etc.).
In general, MEA cell may be a better choice in further study faced on practical applications. It can offer improved mass transport and appropriated coverage of gaseous reactants as well as play a vital role in exhibiting great electrocatalytic performance. For instance, using GDE increases the local CO 2 concentration. Especially, it is favorable to efficiently suppress HER and accelerate *CO dimerization at a rational region of pH (pH > 7), resulting in lower overpotential, higher current density [96] (even >1 A cm À2 ), and enhanced C 2 compounds selectivity. [22,40a,97] Therefore, MEA cell may offer a promising example for producing high-value C 2 compounds by C─C coupling toward practical applications. Although the majority of C─N coupling works have been supported by H-type electrolytic cells so far, there are profitably attempts to introduce GDE in electrolysis system and the system has far outstripped the performance in H-type electrolytic cells. [11a] With the booming development on C─N coupling, more attentions are expected to be paid on exploring reactors.

Summary and Outlook
The strength of adsorption between intermediates and active sites is crucial for tuning the selectivity of products. Thus, all efforts on C─C coupling are valuable references both in experiment methods and in theoretical calculations. For C─C coupling, there exists four mechanisms (*CO dimerization, CHO─CO coupling, CO─COH coupling, and CO─C coupling) based on a key intermediate-*CO. It is noted that the other precursor in C─C coupling always is from the reduction of *CO (such as *COH, *CHO, *C). The strength of adsorption between *CO and active sites determines many possible C─C coupling pathways and branches into many products. [98] Similarly, for the co-reduction of CO 2 and N-integrated, the strength of adsorption between C/N-intermediates and catalysts also plays a crucial role in C─N coupling process. For Figure 16. a) A H-type electrolyzer. b) A three-electrode three-chamber flow cell. Reproduced with permission. [11b] Copyright 2020, Springer Nature. c) A MEA electrolyzer. Reproduced with permission. [89] Copyright 2018, American Chemical Society. d) A reduction electrolyzer with a solid electrolyte (anion and cation exchange membrane). Reproduced with permission. [95] Copyright 2019, Springer Nature.
www.advancedsciencenews.com www.advenergysustres.com instance, inadequate interactions between CO and AuPd nanoalloy will induce a high energy barrier in coupling *CO with *NH 2 OH. [75] Besides, urea is generated from the second C─N coupling between *CONH 2 and N-intermediate but the hydrogenation of *CONH 2 will produce formamide due to distinct strength of adsorption between the same intermediate *CONH 2 and different active sites. Therefore, the role of atomic carbon and nitrogen in directing electrocatalytic reduction of CO 2 to value-added N-containing organics is expected to explore further, which can refer to corresponding understanding from C─C coupling. This review is aimed to offer comprehensive overview regarding electrocatalytic reduction of CO 2 to value-added chemicals via C─C/N coupling on developing electrocatalysts and constructing efficient strategies. Multitudinous important attempts are published as mentioned above; nevertheless, there remains critically challengeable for further optimizing relevant processes.

C─C Coupling to C 2 Compounds
Except for the only single-metal catalyst (copper) that plays a key role in the generation of C 2 compounds, bimetallic catalysts without Cu as well as metal-free catalysts [99] also exhibit potential in recent reports, such as Pd-Au, [100] Ni-Ga, [101] and Ag-Co [102] materials. Based on CO* and C* as descriptors, many non-Cu-based alloys are favorable for C 2 selective performance by the means of theoretical screening. [103] Besides, Arrigo et al. report a novelty mechanism that the surface C atoms, generated from CO 2 reduction, will be dissolved into Fe phase via carbon-supported FeOOH. And then, subsequent hydrogenation happens in order to produce hydrocarbons at the expenses of stable carbide. [104] The latest works have proposed rationalizing insights into noncopper electrocatalysts based on new mechanisms. More works are looked forward to making great progress in developing noncopper electrocatalysts. Moreover, the acid electrolyte is an alternative to avoid pointless CO 2 losses and large voltage losses on count of reaction between CO 2 and alkaline electrolyte. It is a promising avenue to improve the output efficiency of CO 2 -to-C 2 compounds. Nevertheless, it is challenging for weakening unavoidable HER in a proton-rich environment.

C─N Coupling at a Preliminary Stage
Engineering the Surface Microenvironment of Catalysts: First, interface, a crucial part in heterogeneous catalysis, offers an important location to finish dissolution, adsorption, coupling, and desorption steps of electrocatalytic reduction reaction. Given some similarities, the strategies from C─C coupling will inspire relevant researches of C─N coupling. Engineering the surface microenvironment of catalysts, such as surface activation/functional modification (hydrophobicity, huge electrode surface, and plasma treatment) and the modulation of electrolyte (electrolyte composition and pH), [105] has been demonstrated as available approaches for converting CO 2 to C 2 products via reducing adsorption energy of key intermediates and suppressing the competitive reactions.
Engineering Targeted Active Sites: Realizing the isolated reduction of CO 2 and N-containing reactants is not difficult, but it is challenging to finish the coinstantaneous activation of reactants and achieve C─N bonding instead of self-reduction (C─C or N─N coupling) in electrocatalytic co-reduction with CO 2 and N-integrated. For instance, in Yu's works, [80] C─N coupling is triggered by the thermodynamically spontaneous reduction of NO 3 À to the *NO 2 , whose adsorption free energy is À2.65 eV. Consequently, the surface of catalyst will be fully covered by *NO 2 resulting in inhibiting CO 2 activation and hindering the generation of urea. Hence, the coinstantaneous activation of reactants is significant to trigger the generation of C─N moiety. The strategy of constructing electrophilic and nucleophilic sites as well as frustrated Lewis pairs exhibits a promising paradigm toward efficient activation. As mentioned in Section 3, developing further dual-site catalysts, including alloying, oxygen vacancyrich oxides, and heterostructure, has provided potential of activating CO 2 and N-containing reactants in isolation. In addition, it should be noted that appropriate interaction between active sites and C/N-intermediates is favor of high selectivity over C─N compounds. Otherwise, over wake/strong adsorption leads reactions to the direction of high C─N coupling barrier energy and totally self-reduced products. [75] For now, it has been realized that the comprehending of electrocatalytic reduction of CO 2 to value-added chemicals via C─C/N coupling is limited and far away from satisfying demand for the further study. An attempt to achieve the regulation of the electronic structures for precise catalysis via creative explorations from expanded perspectives, such as lattice, atomic orbital, and spin degrees of freedom, is promising. The corresponding researches on oxygen evolution reaction (OER), oxygen reduction reaction (ORR), and HER can offer references. [106] Introducing External Conditions into Co-Reduction process: Introducing temperature/light condition into reaction system is an innovative choice to affect thermodynamic potentials and kinetic overpotentials of co-reduction process. It can be realized by available solar energy which is more sustainable development and economic feasibility for following application research.

Strengthening the Comprehending of the Mechanism in Experimental and Theoretical Level
Electrocatalytic reduction of CO 2 to value-added chemicals exhibits a competitive advantage on account of adsorption of key intermediates in C─C/N coupling in terms of published studies. With regard to this prospective field, it is urgent to propel universal and constructive discovery via systematic experiment and theory for high-efficient process. In order to identify various intermediates, exploring more advanced high-resolution in situ/operando characterizations is necessary to provide insight into the evolution of whole reaction process. Online DEMS, in situ X-ray absorption spectroscopy, in situ ATR-FTIR, in situ Raman spectroscopy, and in situ EPR are included. Meanwhile, DFT calculations based on operando catalytic conditions are challenging for electrocatalysis. The combination of in situ/operando characteristics and theoretical calculations reveals actual reaction pathway.

Optimizing the Electrolysis System
Currently, the system of H-type electrolytic cells with an ionexchange membrane is mainstream research in laboratory. It is of great significance to design high-performance electrocatalytic device architecture for the sake of responding to the emergent appealing for cutting back carbon emission, and preparing for the requirement of commercialization, which is highly attractive for a sustainable energy economy. It is well known that mass transfer processes, to a great extent, govern the regulation of reactivity and selectivity for electrocatalytic reduction of CO 2 to valueadded chemicals. Designing catalyst architectures for CO 2 in situ capturing instead of CO 2 gas bottle offers a promising route for efficient and practical CO 2 reclamation. Inevitably, it is indispensable to solve the problem of corrosion to device due to nitrogen and sulfur compounds from raw gases. Next, a mixture of reduced products needs to be separated and purified via ancillary facilities.
In addition, it is noted that the stability of catalysts and electrolysis systems is of great importance in industrial level. Unfortunately, especially for C─N coupling process, it has stayed at the laboratory level and it has not yet drawn enough attention and effectively being resolved facing commercial scenario. We are keen on more efforts on understanding and solutions of catalyst degradation and the corrosion of devices. Therefore, as far as high current density and long-term stability are concerned, the rational design of electrocatalytic reactor configurations as well as robust electrocatalysts offers opportunities to make great progress in achieving industrial-scale production and drive the field forward.

Developing Integrated Electrolysis
The overall CO 2 electrolysis process consists of CO 2 RR and OER. Given solubility and competitive reaction in alkaline/acidic electrolyte, OER coupled with CO 2 RR under near-neutral pH conditions is at the expense of performance for the development of CO 2 RR devices to some degree. Thus, exploration emerging valuable oxidation reactions instead of the conventional anodic reaction is more attractive to increase the energy conversion efficiency and economic benefits for co-reduction process. The evaluation of alternatives to adapt the integrated electrolysis remains to be taken a more thorough investigation thermodynamics, dynamics, values of products, and separation of products.