Heterogeneous nanoporous organic frameworks-based catalysts for electrochemical CO₂ reduction reaction

Bulk MOFs

The isolated metal nodes of MOFs can serve as active sites, providing excellent monoatomic dispersity with 100% atomic efficiency. Bismuth (Bi) is considered to be a promising electrocatalytic material that can convert CO₂ into HCOOH^[52]. Therefore, Bi-based MOF (Bi-MOF) catalysts for CO₂RR can be expected to possess high catalytic activity and product selectivity at low overpotential^[53]. A stable Bi-MOF (Bi-BTC-D) was previously prepared by a hydrothermal process using trimesic acid (H₃BTC) and bismuth nitrate pentahydrate [Bi(NO₃)₃·5H₂O] with a high FE of 95.5% for HCOOH at -0.86 V_RHE^[54]. For this sample, the FE remained at ~90.0% even after 12 h of continued electrolysis, showing good stability. Moreover, the morphology of Bi-MOFs could be regulated to expose more active sites, thereby improving production efficiency of HCOOH. Inspired by this study, CAU-17 Bi-MOFs with hexagonal prism (CAU-17-hp) and nanofiber (CAU-17-fiber) morphologies have been synthesized by a wet-chemical process^[55]. In particular, the sample CAU-17-fiber with a carbonization temperature of 400 °C (CAU-17-fiber_400) with a hierarchical structure developed by H₃BTC-mediated morphology reconstruction had a high FE of 96.4% for HCOOH at -0.90 V_RHE.

The coordination environment of metal sites plays a significant role in CO₂ activation and stabilization of important intermediates in the electrocatalytic process. A porous zeolitic MOF (Bi-ZMOF) was prepared by introducing carboxylic acid groups (yrazole-3,5-dicarboxylic acid, PDC) as ligands, which could convert CO₂ to HCOOH with FE of 91% at -1.1 V_RHE^[56]. Theoretical calculations showed that the coordination of Bi with the pyrazole-N atom in the MOF facilitated electron mobility from the ligand to Bi sites, which promoted the activation of CO₂, thereby reducing the potential for the generation of OCHO* and ultimately facilitating conversion of CO₂ to HCOOH; Meanwhile, the weak bonding (hydrogen bonding) of COOH to the carboxyl group results in weak adsorption of COOH* on the MOF, ultimately inhibiting CO production. It has been reported that biomass-derived materials can capture Pb²⁺ in wastewater to synthesize Pb-based MOFs, which can not only solve lead (Pb) poisoning but also electrochemically convert CO₂ to HCOOH. The ellagic acid (EA) obtained from pomegranate bark could effectively capture Pb²⁺ in aqueous media to synthesize stable Pb-based MOFs (EA-Pb), which could efficiently convert CO₂ to HCOOH with FE of 95.37% at -1.08 V_RHE^[57]. At the same time, the removal rate of Pb²⁺ even at lower concentrations by EA was close to 99%. In the structure of EA-Pb, PbO₆, as a metal node, has a tunable coordination mode (coordination number of 4-9) with the EA linker, which was essential for the reduction of CO₂ to form the key intermediate HCOO*. Since the formation of CO and HCOOH involved the same RDS (*COO→HCOO*), the reaction pathway and product selectivity were revealed by theoretical calculations. Gibbs free energy change values for *HCOOH formation (-0.96 eV) and *HCOOH desorption (-0.59 eV) were more negative than *CO formation from HCOO* decomposition (-0.73 eV) and *CO desorption (-0.48 eV), indicating that the reduction of CO₂ to HCOOH was thermodynamically more favorable. In parallel, the energy barrier of HER (1.99 eV) was much higher than that of CO₂ reduction (1.37 eV), proving that the catalyst inhibited hydrogen evolution, thus showing the advantage of greater HCOOH selectivity.

The generation of any CO₂ reduction product requires a multi-electron transfer process, and the poor conductivity and electron-donating ability have been the primary limiting factors for the exploration of efficient MOFs. Therefore, the integration of electron-rich units, electron mobility, and active components into MOFs may be an efficient method to enhance CO₂RR performance. For example, (i) electron-rich polyoxometalates (POMs) can provide a large number of electrons; and (ii) metal-tetrakis(4-carboxyphenyl) porphyrins (M-TCPPs, where M refers to a metal species) exhibit an intrinsic macrocyclic conjugated π-electron system, which can be highly conducive to electron migration. A range of POM-based MOFs (M-PMOFs) have been synthesized by a one-step process using a POM and M-TCPP^[58]. The direct connection between POM and M-TCPP generated a directional electron transport channel from POM to M-TCPP under the application of an electric field. This helped promote the multi-electron migration during CO₂ reduction and efficiently realize the conversion of CO₂ to CO. The FE_CO of the cobalt (Co)-PMOF reached 99% at -0.8 V_RHE, with excellent stability of > 36 h of recycling. This study showed that the intramolecular multi-electron transfer in the catalyst can also improve the electrocatalytic CO₂RR performance. Theoretical calculations showed that the formation energy barrier of *COOH was significantly reduced after the assembly of POM and Co-TCPP (0.34 eV), compared with that of POM (0.96 eV), indicating that Co in Co-TCPP was the active site of the reaction rather than POM and there was an effective synergy between Co-TCPP and POM. The Co center in Co-TCPP could be reduced from Co(II) to Co(I) by accepting electrons captured by POM, and then Co(I) interacted with CO₂ to generate Co(II) *COOH, which combined with protons and electrons to form Co(II) *CO, and finally to form CO.

In addition, Zhu et al. synthesized a copper (Cu)-based MOF structure (Cu-MFU-4l) with copper triazole (Cu(I)N₃) as an active site to realize the process of CO₂-to-CH₄^[59]. The FE of CH₄ for this catalyst reached 92% at -1.2 V_RHE. It has been reported that the selectivity of product was greatly affected even by slight differences in coordination environment of metal ions. Liu et al. compared two kinds of coordination structures based on catechol (C₆H₆O₂)-derived ligands with copper oxide (CuO₄ and CuO₅) nodes to study the impact of metal d-orbital energy levels on the selectivity of electrocatalytic CO₂RR products^[60]. Compared to the square-planar CuO₄ site, the square-pyramidal CuO₅ site had a higher Cu 3d orbital energy level through coordination with its axial oxygen (O) atom. This enabled the CuO₅ site to form a stronger π-back bond with the intermediate CO, thus facilitating its adsorption and the subsequent multi-step hydrogenation process to form CH₄.

The high selectivity of C₂ products can be achieved by controlling the valence state of metal nodes in the MOF. Deng et al. used methanol as a reducing agent to reduce Cu²⁺ in Cu(II)-Trimesic acid (BTC) framework (Cu(II)-BTC) to Cu⁺ which played a key role in the selectivity of C₂ products, realizing the reconstruction of Cu⁺ and BTC and the generation of free carboxyl groups^[61]. The electroreduction of CO₂ to C₂H₄ was promoted by the change of the valence state of metal nodes and the generation of free carboxyl groups, and the FE was increased from 22% to 57% at -1.6 V_RHE. The electrocatalytic performance did not decay after 38 h, indicating good stability. In addition, controlling the crystal face of MOFs with different solvents can also be used to adjust the selectivity of CO₂ reduction. Lu et al. prepared Cu(I) 5-mercapto-1-methyltetrazole frameworks (Cu-MMTs) with (100) and (001) crystal faces using water and isopropanol as solvents, respectively^[62]. Cu-MMTs with (001) crystal faces were beneficial to the formation of multi-carbon products with 73.75% FE. The difference was that the Cu-MMT with (100) crystal faces promoted the formation of single carbon with 63.98% FE. Furthermore, the size of MOFs also affects the selectivity of CO₂ reduction. Four different sizes of Cu-MOFs (53, 109, 307, 1,335 nm) were synthesized^[63]. As a result, small-sized MOFs with polycrystalline structures converted CO₂ to ethylene and multi-carbon products with FEs of 55.4% and 81.8%, respectively, which was superior to large-sized single-crystal nanostructures.

Zinc (Zn)-based MOFs (Zn-MOFs) can also be used as electrocatalysts to convert CO₂ to CO^[64,65]. For this type of MOF, due to complete occupation of Zn(II) 3d orbitals, active sites mainly originate from the coordination of the ligands with the Zn(II) centers; i.e., the CO₂RR performance can be adjusted by regulating the ligands. Al-Attas et al. employed 1, 2, 4-triazolate and 2-methylimidazolate as ligands to synthesize two different Zn-MOFs, Calgary framework 20 (CALF20) and zeolitic imidazolate framework-8 (ZIF-8), respectively^[66]. The study indicated that CALF20 had a higher CO₂ adsorption capacity (more than 4.5 times that of ZIF-8). Moreover, the sp² C active sites in the 1, 2, 4-triazolate of CALF20 were far more in number than in the 2-methylimidazolate of ZIF-8. With increased CO₂ adsorption performance and multiple binding sites, CALF20 exhibited better performance with a high CO FE of 94.5% at -0.97 V_RHE. The theoretical results showed that sp² C in the azole ligand coordinated with Zn(II) site could be used as catalytic active center for the electrochemical reduction of CO₂. Compared with the diazole group in ZIF-8, the faster charge transfer efficiency of the triazole ligand in CALF20 promoted the accumulation of more electrons at adjacent active sites, thereby lowering the adsorption free energy of *COOH and improving the conversion efficiency of *COOH to CO.

Two-dimensional MOFs

Recently, two-dimensional (2D) MOF nanosheets have been widely used in electrocatalytic CO₂RR due to their outstanding advantages: (1) Compared to bulk MOFs, MOF nanosheets can expose a higher proportion of metal atom active sites on the surface for CO₂RR catalysis. It is particularly important that the exposed coordination-unsaturated metal atoms on the surface exhibit high catalytic activity^[67]; (2) The electronic structure of metal atoms in MOF nanosheets can be effectively controlled by selecting appropriate organic ligands^[68]. For electrocatalysis, the electronic structure of catalyst influences the binding strength of the intermediate products on the catalyst surface, and then determines the final activity and product selectivity; (3) Compared to bulk MOFs, MOF nanosheets have better conductivity^[69]; (4) The highly porous, ultrathin characteristics of MOF nanosheets are conducive to the transport of reactants and products during the catalysis reaction^[70,71]; and (5) Due to the diversity of MOF structures and the progress in material synthesis technologies, a large number of MOF nanosheets can be prepared easily to study their potential electrocatalytic properties^[72].

Two-dimensional conductive MOFs (c-MOFs), as a member of the 2D materials family, are developed by combining transition metal nodes with conjugated organic ligands^[73,74]. The framework of these c-MOFs and the corresponding physical properties are largely derived from the incorporated ligands. However, the synthesis of 2D c-MOFs is limited by available ligands. Macrocyclic compounds are regarded as infinite π-conjugated systems, exhibiting unique electronic behavior regardless of the presence or absence of substituents. Therefore, conjugated macrocyclic structures with well-defined shapes and non-foldable and fully π-conjugated backbones can be used to construct 2D MOFs^[75,76].

Two different ligands, 1,4-benzenedicarboxylic acid (BDC) and 2,6-naphthalenedicarboxylic acid (NDC), were selected to prepare two 2D-MOFs (Cu-UBDC and Cu-UNDC) regulating electronic structure of Cu active centers^[77]. It was found that the ratio of Cu⁺/Cu²⁺ was 1.12 for Cu-UNDC, far exceeding that of Cu-UBDC (0.85), indicating the change of electron densities of metal sites, thereby promoting electron migration. Consequently, FE of C₂H₅OH and C₂H₄ for Cu-UNDC was 24.3% at -1.0 V_RHE, far exceeding that of Cu-UBDC (13.2%). In addition, for Cu-UBDC, Cu₂O as a photocatalyst could be generated under the condition of CO₂ reduction. Therefore, more efficient charge separation efficiency and more electrons can be used for CO₂ reduction under illumination, resulting in an improved FE of 26.2% for the C₂ product of Cu-UBDC, which was better than that of Cu-UNDC (21.8%).

Metalloporphyrin organic ligands exhibit tetrapyrrole macrocyclic conjugated structures. When they are used to construct MOFs, different topological structures can be formed by changing the number of nodal metal clusters or connections, which greatly improves the structural diversity of the porphyrin-metal framework materials. The metal centers of metalloporphyrins are easily reduced to a low-oxidation state during CO₂ reduction, and the macrocyclic ligand skeleton helps maintain chemical stability^[78]. Moreover, the structure of metalloporphyrins can be tuned and functionalized using redox-active molecular catalytic linkers^[79]. Metalloporphyrin-based 2D MOFs have been widely studied as CO₂RR electrocatalysts^[80]. Han et al. synthesized well-defined ultrathin 2D nanosheets (STPyP-Co) by axial coordination assembly of tetra(4-pyridyl) porphyrin Co(II)^[81]. Originating from the ultrathin characteristics of 2D MOFs, the nanosheets exhibited excellent conductivity and unsaturated five-coordinated single sites on their surface. The axial pyridine coordination led to the increase of the d_z² energy level in the Co 3d orbit, which enhanced the Lewis basicity of the surface Co center, thus allowing more electrons to be transferred from Co to CO₂, which was beneficial to the activation of CO₂ and reduced the free energy required for the CO₂*→COOH process. As a result, STPyP-Co could convert CO₂ into CO with a high FE of 96% at -0.62 V_RHE. This sample had good stability with FE of above 90% throughout the reaction process.

It has been reported that the Co-N₄ site in the porphyrin ring can be used as an active site, with strong *COOH binding energy and moderate *CO adsorption energy in the process of electrocatalytic CO₂RR. Therefore, a 2D-MOF based on Co-porphyrin is an excellent electrocatalyst for achieving the transformation of CO₂ to CO^[82]. Chen et al. prepared 2D Co-based porous porphyrin nanolayer (Co-PPOLs) catalyst with Co-N₄ sites through a bottom-up self-assembly strategy^[83] [Figure 4A]. The ultrathin nanolayer showed an average thickness of 3.8 nm [Figure 4B], which was conducive to the exposure of more active sites and faster electron transfer, so it could promote the adsorption, activation and transformation of reactants/intermediates, thereby improving the activity, selectivity and stability of electrocatalytic CO₂RR. Therefore, the highest FE_CO was 94.2% at -0.9 V vs. RHE [Figure 4C]. Moreover, the electrocatalytic CO₂ reduction performance of Co-PPOLs under industrial currents was also studied. At a full battery voltage of 2.7 V (higher than the industrial grade current of 200 mA), the CO FE of CO-PPOLs could still reach 92% for 20 h in MEA cells [Figure 4D and E]. In addition, the electrocatalytic process with the assistance of light was conducive to the stabilization of CO intermediates bound to the catalyst, and promoted the formation of high-electron product CH₄. Based on the above conditions, MOFs based on cobalt 5, 10, 15, 20-tetri (4-carboxyphenyl) porphyrin (CoTCPP) (Co-MOF-525) achieved an FE_CH4 of 14% at -0.49 V vs. NHE^[84][Figure 4F and G]. Under light conditions, the generation of [Co(III) TCPP]⁺ species could stably bind to CO, and then be electroreduced to CH₄.

Figure 4. (A) The synthesis of ultrathin 2D Co-PPOLs. (B) AFM image of Co-PPOLs. (C) FE of CO and H₂ at different potentials. (D) FE_CO of Co-PPOLs in MEA under different potentials. (E) Long-term stability of Co-PPOLs in MEA^[83]. Copyright 2023 Wiley-VCH. (F) Schematic illustration of the photo-assisted CO₂RR on Co-MOF-525. (G) The FE of CH₄ and CO under dark and light^[84]. Copyright 2023 Wiley-VCH.

A Cu-MOF nanosheet, Cu₂(Cu-TCPP), with porphyrinic Cu and Cu paddle wheel in two different chemical environments, was synthesized under cathodization conditions^[85]. Due to cathodized reconstruction, the Cu-MOF nanosheets exhibited an FE_HCOOH of 68.4% at -1.55 VAg/Ag⁺, superior to those of CuO, Cu₂O, Cu, and Cu-TCPP. Similarly, a 2D porphyrin organometallic framework nanozyme (PPF-100) with Cu-paddle-wheels, Cu₂(COO)₄ structure could realize highly selective CO₂RR and significantly inhibit the HER^[86]. The FEs of CO and HCOOH were 72.4% and 24.1% at -3.0 V vs. Ag/Ag⁺ with no generation of H₂, attributed to the hydrophobicity of PPF-100, which was conducive to inhibiting the HER, thereby improving the electrocatalytic performance of CO₂ reduction. Furthermore, Wu et al. prepared ultrathin 2D nickel (Ni)-based ZIF nanosheets [Ni(Im)₂] by exfoliating bulk MOFs to nanosheets of varying thicknesses as electrocatalysts to convert CO₂ to CO^[87]. Due to the existence of more Ni sites and rapid mass and electron transfer by ultrathin nanosheet structure, the CO FE for the 2D Ni(Im)₂ nanosheets with thickness of 5 nm reached ~78.8% at -0.85 V_RHE, much higher than that of bulk Ni(Im)₂ ZIF (33.7%).

Metallophthalocyanines (MePc) have also been used for the preparation of 2D-MOFs with thermal/chemical stability and structural tunability at the molecular level^[88,89]. 2D NiPc-NiO₄ nanosheet MOF was previously constructed using Ni phthalocyanine (NiPc)-2,3,9,10,16,17,23,24-octanol (NiPc-OH) and Ni(II) ions^[90]. The NiPc linked with nickel-catecholate [Ni(C₆H₄O₂)₂] served as an active site in this MOF; the high d-π orbital overlap between the two increased the conductivity of the nanosheets to ~4.8 × 10^-5 S m^-1. The adsorption energy of CO₂ on NiPc was 0.23 eV, much higher than that on NiO₄ (0.02 eV), indicating that CO₂ molecules were more easily bound to NiPc through non-covalent interactions. Furthermore, the electron-rich environment in NiPc was more favorable for the reduction of CO₂. In addition, the COOH* formation energy barrier (1.93 eV) on NiPc was lower than that of H* (1.98 eV), whereas on NiO₄, it (2.53 eV) was higher than that of H* (1.58 eV), indicating that CO₂RR occurred preferentially on NiPc, While HER was dominant on NiO₄. Owing to rapid electron transport capacity, these nanosheets exhibited a FE_CO of 98.4% and partial current density of 34.5 mA cm^-2. This study showed that the catalytic ability of Ni active sites can be improved by inserting phthalocyanine (H₂Pc) molecules into 2D c-MOF materials. The well-maintained NiPc active sites after CO₂RR allowed the electrocatalyst to show superior stability with a high FE_CO of 86%.

To further improve the electrical conductivity of H₂Pc-based MOFs, NiPc-Ni(NH)₄ nanosheets have been synthesized by using 2,3,9,10,16,17,23,24-octaaminophthalocyaninato nickel(II) [NiPc-(NH₂)₈] and Ni(II) ionsm^[91]. Due to the NiPc links with Ni(NH)₄ nodes, the electronic conductivity of NiPc-Ni (NH)₄ can reach up to 2.39 × 10^-4 S m^-1, improving the electron transfer capacity. Controlled experiments and density functional theory (DFT) calculations indicated that, unlike NiPc-NiO₄ nanosheets, a Ni-N₄ unit in the H₂Pc ring acted as catalytic active center, while the Ni(NH)₄ node accelerated transport of protons/electrons to the active site, thereby promoting the process of CO₂-to-CO. The FE_CO for this catalyst reached 96.4% at -0.7 V_RHE with a current density of 24.8 mA cm^-2 at -1.1 V_RHE, which were remarkably good characteristics for H-type electrolytic cells among the reported 2D MOFs.

The conversion of CO₂ to C₂H₄ has great significance for sustainable energy production. However, this process can be quite difficult due to high energy barriers for the hydrogenation of *CO and C-C coupling steps^[92]. For such cases, electrocatalysts with dual active sites are expected to be a better choice. By assembling metallo-ligand (2,3,9,10,16,17,23,24-octahydroxyphthalo-cyaninato) Cu(II) (PcCu-(OH)₈) with CuO₄ nodes, a PcCu-Cu-O 2D MOF was prepared for conversion of CO₂ to C₂H₄^[93]. It exhibited a high FE of 50(1)% and current density of 7.3 mA cm^-2 at -1.2 V_RHE with the synergistic effects of CuPc and CuO₄; i.e., the CO formed on the CuO₄ site could rapidly diffuse on CuPc site and dimerize with *CHO to form *OCCHO, thereby reducing C-C dimerization energy barrier. In summary, this study provided a strategy for designing and utilizing electrocatalysts with dual active sites to reduce CO₂ into C₂₊ compounds.

Ligands with heteroatoms, such as tricycloquinazoline (TQ) and hexaazatrinaphthylene (HATNA), have also been reported for the 2D c-MOF preparation. Liu et al. selected an electron-deficient TQ moiety in the ligand incorporated into the 2D MOF framework and coordinated it with Cu²⁺ and Ni²⁺ to prepare M₃(2,3,7,8,12,13-hexahydroxytricycloquinazoline) [M₃(HHTQ)]^[94]. The Cu₃(HHTQ)₂ exhibited a cellular 2D network that passes through π-π interaction, which was further stacked along the c-axis; the crystal edge with a ligand terminal was observed, indicating good crystallinity and conductivity. In CO₂-saturated electrolytes, the current density of Cu₃ (HHTQ)₂ can reach 45 mA cm^-2 at -1.2 V_RHE. Meanwhile, the Cu₃(HHTQ)₂ showed FE of 53.6% for CH₃OH under low overpotential conditions (-0.4 V_RHE). Similarly, a HATNA-Cu-MOF by assembling HATNA as a ligand with electron-deficient conjugated molecule with Cu catecholate node could efficiently convert CO₂ to CH₄, with FE of 78% and current density of 8.2 mA cm^-2 at -1.5 V_RHE^[95].

Single-atom MOFs

Atomically dispersed metal sites (ADMSs) in metal nodes of the coordinatively unsaturated sites are isolated metal atoms with well-defined local coordination structures, whose coordination bonds, bond lengths, and coordination numbers can be tuned to change the electrocatalytic CO₂ selectivity/activity. The maximum atomic utilization and the unsaturated coordination environment reduce the excess metal cost and improve the catalyst performance. The design of monatomic catalysts with good electrical conductivity, discrete active centers, and good mass transfer performance for efficient electrocatalysis can be a challenging process. ADMSs can be easily introduced into (i) metal nodes and organic ligands due to their highly ordered arrangements; and (ii) the pores of MOFs. This property leads to advantages in several reactions, making ADMSs an important component of heterogeneous catalysis research.

Chen et al. prepared an electrocatalyst (2Bn-Cu@UiO-67) by embedding N-heterocyclic carbene (NHC)-linked Cu single-atom sites (SAs) (2Bn-Cu) into porous MOF shell (UiO-67) for CH₄ electrosynthesis^[96]. It had FE of ~81% for CH₄ at -1.5 V_RHE, with a corresponding current density of 420 mA cm^-2, due to the following properties: (i) Cu-SAs as the catalytic center of CO₂RR were anchored on NHC by metal-carbon bonds. The electron donor effect of the NHC increased the charge density around metal sites which was beneficial to the adsorption of CHO* intermediate, thus promoting the CH₄ formation; and (ii) The MOF framework had a strong ability to capture CO₂ molecules and provided multiple channels to facilitate the diffusion of CO₂ to 2Bn-Cu. These two aspects synergistically promoted efficient electrocatalytic CO₂RR.

SA coordination with diverse heteroatoms exhibits advantages for the selectivity of various hydrocarbons^[97]. Zhang et al. synthesized Cu-based c-MOF (Cu-DBC) consisting of a conjugated graphene-like ligand (dibenzo-[g, p]chrysene-2,3,6,7,10,11,14,15-octaol, 8OH-DBC) and Cu-O₄ active sites to study the effects of coordination environment on CO₂RR^[97]. It was found that highly conjugated organic ligands imparted distinctive redox properties and electrical conductivity to Cu-DBC, while the abundant and uniformly distributed Cu-O₄ sites contributed to highly selective transformation of CO₂ to CH₄. The Cu-DBC catalyst had FE of ~80% for CH₄ with partial current density of 162.4 mA cm^-2 at -0.9 V_RHE. The relationship between the selectivity of CO₂RR and coordination environment of a separate Cu site was studied by DFT. The results showed that Cu-O₄ site in Cu-DBC had a lower energy barrier than the N-coordinated Cu site in 2BN-Cu@UiO-67, showing better CO₂RR performance.

SAs are more likely to reduce CO₂ into C₁ products, such as HCOOH, CO, CH₄, etc., while the formation of C₂₊ products by C-C coupling often requires double active sites^[98]. Based on the above considerations, a MOF with Ir-porphyrin and Cu-SAs double active sites (Cu-SAs@ Ir-PCN-222-PA) was formed by confining Cu SAs in a nano-framework of Ir-porphyrin-based MOF (Ir-PCN-222), realized the tandem catalytic process of CO₂→CO→C₂H₄^[99] [Figure 5A]. As a result, Cu-SAs@Ir-PCN-222-PA could promoted the transformation of CO₂ to C₂H₄ with high FE of 70.9% and current density of 20.4 mA·cm^-2 at -1.0 V_RHE [Figure 5B]. However, no C₂H₄ but CO products were produced at Ir-PCN-222 and Cu-SAs@PCN-222-PA, demonstrating that Cu-SAs and Ir-porphyrin worked together to promote the reduction of CO₂ to C₂H₄. The reaction mechanism was investigated by in-situ diffuse reflectance infrared Fourier transform spectroscopy (DRIFT) and DFT calculations. The signals of Cu-*CO, Ir-*CO where *CO was the intermediate in the process of CO₂-to-C₂H₄, and free CO and Cu-*COCHO were found [Figure 5C]. In addition, CO molecules first appeared in Ir-PCN-222 and then transferred on Cu-SAs with a large diffusion coefficient (2.3 × 10^-7 m² s^-1), and then underwent C-C coupling reaction with *CO^[100], which was consistent with DFT calculation. Specifically, CO adsorption energy on Ir-porphyrins (-0.70 eV) was significantly lower than that of Cu-SAs (-0.82 eV) [Figure 5D], proving that CO was formed on Ir-porphyrins sites. Then, CO molecules desorbed from the Ir-porphyrin site and reacted with Cu-*CO to form *COCHO due to the low barrier energy of *COCHO formation, followed by multiple dehydrogenation and hydrogenation steps to form C₂H₄.

Figure 5. (A) Synergistic tandem electrocatalytic CO₂ to C₂H₄ for Cu-SAs@Ir-PCN-222-PA. (B) FE at -1.0 V_RHE. (C) In situ FTIR spectrum for the electrocatalytic CO₂RR on Cu-SAs@Ir-PCN-222-PA. (D) CO formation on Cu-SAs and Ir-porphyrin of Cu-SAs@IrPCN-222-PA^[99]. Copyright 2024 the American Chemical Society.

The electronic structure adjustment of sp² C sites in imidazole organic ligands influences the performance of electrocatalytic CO₂RR. M_xZn_y/ZIF-8 (M: transition metal) was prepared by substituting a portion of Zn²⁺ in ZIF-8 into Ni²⁺, Fe²⁺, and Cu²⁺ through a doping strategy^[101]. As a result, the optimum Cu_0.5Zn_0.5/ZIF-8 could convert CO₂ to CO with the highest FE_CO of 88.5% and partial current density of 11.57 mA cm^-2 at -1.0 V vs. RHE, which was much higher than that of pristine ZIF-8 (FE_CO: 43.7%). The enhancement of performance was attributed to the enrichment of electrons on the sp² C in the imidazole organic ligands because of the introduction of additional electronic states by Cu doping, promoting the rapid transfer of electrons. Furthermore, the change in the electronic structure of sp² C made Cu_0.5Zn_0.5/ZIF-8 more strongly bound to COOH*, promoting the transformation process of CO₂-to-CO. Furthermore, the electronic properties of SAs, such as binding energy and reaction coordinates, are considered the key factors determining their catalytic performance^[102]. Xue et al. created a model to study these properties by confining a series of 3d transition metal SAs into the microporous cavity of UiO-66-NH₂ for electrocatalytic CO₂RR^[103].

Metal-MOF composites

MOF films can be deposited on highly reactive catalytic materials, such as gold nanostructured microelectrodes (AuNMEs), silver (Ag), and other solid electrodes, to obtain high catalytic efficiency^[105]. Ag, as an electrocatalyst, can reduce CO₂ to CO^[106]; however, this process can be impeded by competitive HER and selectivity of electrochemical reductive products^[107]. Non-catalytic Zr₆-oxo-based MOF (UiO-66) thin films with different thicknesses and chemical compositions have been assembled on Ag electrodes^[108]. The existence of MOF membranes can systematically regulate the chemical environment of the active site on a molecular level, resulting in a significant improvement of electrocatalytic performance. Three key mechanisms have been used to verify such an improvement in electrocatalytic activity. Firstly, porous MOF membranes weakened the mass transfer of CO₂ and H⁺ to Ag electrocatalysts. Compared to the reactants in the bulk solution, the local concentration of reactants near the catalyst changed significantly, thereby changing the catalytic path. Secondly, Zr₆-oxo nodes of MOF contained brønsted acidic groups close to the catalytic activity surface. These groups accelerated electrocatalysis by stabilizing the activated *COOH intermediates. Thirdly, the decoration of MOFs with a positively charged ligand, such as (3-carboxypropyl) trimethylammonium (TMA), can control the proton transfer at the catalytic activity site, thereby improving the electrocatalytic selectivity. The combination of these three mechanisms increased the selectivity of CO for Ag from 43% to 89% at -0.8 V_RHE.

Considering that the solubility of CO₂ in acetonitrile was higher than that in water, coating the Bi electrode with UiO-66 film decorated by nitrile functional groups (Bi-UiO-66-CN) as a CO₂ enrichment layer can increase the local concentration to 0.82 M, which was 27 times the concentration of CO₂ in the bulk phase, and further improved the electrocatalytic rate and product selectivity^[109]. By optimizing the thickness of the film, the maximum HCOOH FE could reach 93% at -0.75 V_RHE for Bi-UiO-66-CN. Especially, the partial current density of HCOOH could reach 166 mA cm^-2 at -0.9 V vs. RHE, more than seven times that of a bare Bi electrode. The improved activity was attributed to the change of the Bi surface microenvironment because of the anchor of UiO-66-CN membranes. At first, coating a UiO-66-CN membrane could enhance the CO₂ adsorption capacity of the catalytic site, thus accelerating the catalytic reaction. Furthermore, the interfacial structure between Bi and UiO-66-CN promoted CO₂ activation and intermediate stabilization, thereby promoting the selective formation of HCOOH.

The high SSA and periodic framework chemical functionality of MOFs provide the possibility of embedding metal active sites within their framework at a high density to include the characteristics of nanocrystals (NCs) for CO₂RR^[110]. Kung et al. coated MOF films (NU-1000) with Cu nanoparticles (NPs) to prepare an electrocatalyst (Cu-SIM NU-1000) for CO₂RR using the following process^[111]: Initially, Cu(II) was uniformly coated onto the NU-1000 film through solvothermal deposition in MOFs (SIM). Subsequently, Cu(II) was partially reduced to Cu(0) through an electrochemical process. Cu embedded in the MOF film did not change its crystallinity and morphology, but the SSA reduced compared to that of pure NU-1000. As a result, Cu-SIM NU-1000 primarily produced HCOOH as main product with FE of 31% at potential of -0.82 V_RHE. The size of Cu NPs in this sample was limited to < 10 nm because of the MOF film channel size, so the inhibition of HER was less obvious.

The heterogeneous interface constructed by the intimate contact between MOFs and Cu nanowires (NWs) can adjust the adsorption energy of specific intermediates, thus achieving precise regulation of product selectivity. Additionally, porous structure of MOFs facilitates the local enrichment of reactants, the exposure of active sites, and promotes the interfacial diffusion of reactants between Cu-NWs and MOFs, thereby improving catalytic performance. In-situ encapsulation of ZIF-8 on the Cu-NW surface to synthesize the composite Cu@ZIF-8 NWs with core-shell structure resulted in an FE_{CH4 + C2H4} of 57.5% at -0.7 V vs. RHE^[112] [Figure 6A and B]. However, it only produced CO for ZIF-8. For pure Cu-NWs, the mixed total FE of products CO, CH₄, C₂H₄ and HCOOH, C₂H₅OH was low, but for H₂, it was more than 50% at -0.5~-0.9 V. The difference in catalytic activity and product distribution between the composite structure and the single material suggested that the existence of the interface between ZIF-8 and Cu-NWs enabled the regulation of product selectivity and the effective inhibition of HER. In addition, wrapping of ZIF-8 on Cu regulated the adsorption energy of the reaction intermediate, which was also verified by DFT calculation. The combination of *CO and H to produce *CHO was regarded as a crucial stage in the process of CO₂ to hydrocarbon products. The energy barrier of *CHO for Cu@ZIF-8 NWs was 0.62 eV, significantly lower than that of pure Cu (0.79 eV), proving that the formation and stabilization of *CHO was facilitated at the interface between Cu and ZIF-8, promoting the process of CO₂→CH₄+C₂H₄ [Figure 6C].

Figure 6. (A) Schematic diagram of Cu@ZIF-8 NWs for electrocatalytic CO₂RR. (B) FE of CO, CH₄, and C₂H₄. (C) Free energy of CO₂RR on Cu NWs and Cu@ZIF-8 NWs^[112]. Copyright 2024 Elsevier. (D) The possible mechanism of Bi/UiO-66 for CO₂RR^[113]. Copyright 2023 Elsevier. (E) The synthesis of Ag@UiO-66-SH^[115]. Copyright 2023 the American Chemical Society.

It was reported the capture form of CO₂ molecules on MOFs was also one of the factors affecting the efficiency of CO₂ reduction^[113]. Bi NPs, as an electrocatalyst, were loaded on UiO-66 (Bi/UiO-66) for the transformation of CO₂ to HCOOH. The FE of HCOOH could reach 65%-85% at -0.4~-0.7 V_RHE for both Bi and Bi/UiO-66. Notwithstanding, the current density of Bi/UiO-66 was -265 mA cm^-2 at -0.7 V_RHE, 4.6 times higher than that of pure Bi, which was attributed to the presence of CO₂ adsorbed on the UiO-66 in the form of carbonate to promote CO₂ conversion. The mechanism of CO₂RR based on Bi/UiO-66 was also proposed. The CO₂ enriched on Bi/UiO-66 reacted with OH^- (KOH electrolyte) to form HCO₃^- which combined with UiO-66 falling from the electrode to form [Zr₂(OH)₂(CO₃)₄]^2-[114], representing the CO₂ capture process accompanied by the structural transformation of UiO-66. Meanwhile, HCOOH produced on Bi ionizes to release H⁺. The H⁺ then reacted with the carbonate ion to release CO₂, which could be used directly in the CO₂ conversion process [Figure 6D].

Ag NP catalysts are prone to sintering or agglomerating during the reaction, leading to a decrease in their activity. Confining NPs in the MOF is a good means of preventing aggregation while providing a path for product diffusion in its pores. To increase the pore size of MOFs, Aparna et al. replaced the BDC linker in UiO-66 with 2-mercaptobenzoic acid (2-MBA) to create defective MOFs (UiO-66-SH) [Figure 6E]^[115]. Due to the strong interaction between the thiol group and Ag, it helped anchor Ag NPs on MOFs to prepare Ag@UiO-66-SH, which could promote the transformation of CO₂ to CO with 74% FE and high partial current density of 19.5 mA cm^-2 at -1.1 V vs. RHE. To strongly inhibit the HER, Guntern et al. embedded Ag NCs into Al-PMOF ([Al₂(OH)₂-(TCPP)]) [tetrakis(4-carboxyphenyl) porphyrin (TCPP)] to prepare Ag@Al-PMOF, ensuring close contact between the Ag NCs and MOF^[110]. The CO FE of the Ag@Al-PMOF composites was more than twice that of the plain Ag NCs, and the HER was inhibited significantly. Due to the tunability of the synthesis process, the close contact between the Ag NCs and MOF interface acted as a crucial factor in facilitating charge transfer between them and enhancing the catalytic performance.

Metal-oxide-MOFs composites

Metal oxide NPs can act as effective electrocatalysts for CO₂RR^[116,117]. However, their low SSA limits the adsorption of reactant molecules. Moreover, they tend to agglomerate during the reaction, which limits the CO₂ conversion. By embedding metal oxide NPs into MOFs, the porous structure is maintained and the interaction between the NPs and MOFs is enhanced, significantly boosting the strong electrocatalytic performance for CO₂ reduction^[118,119].

Cu₂O, as an electrocatalyst, can convert CO₂ into C₂H₄ with low FE and poor stability. CO₂ capture and adsorption are key steps in the heterogeneous catalytic process; therefore, it is a feasible strategy to combine Cu₂O with MOFs with high SSA to form an integrated catalyst, which can improve the CO₂ adsorption capacity of Cu₂O to realize the conversion of CO₂→hydrocarbons with high selectivity and stability. Tan et al. combined Cu-MOF with Cu₂O to prepare a hybrid electrocatalyst for CO₂RR^[120]. The composite Cu₂O@Cu-MOF exhibited high SSA of 1,467 cm² g^-1 and CO₂ adsorption capacity of 83.6 cm³ g^-1, about ten times higher than that of Cu₂O (8.3 cm³ g^-1). The catalyst showed an excellent performance with an FE_CH4+C2H4 of 79.4%, with the FE for CH₄ being 63.2%. The unsaturated coordination active centers in Cu₂O@Cu-MOF led to a large CO₂ adsorption capacity and the embedded Cu₂O contributed to efficient charge transfer. To further increase the selectivity for C₂H₄, Cu₂O/Cu-CuTCPP heterostructures (Cu₂O/CPFs) with three Cu-active centers (Cu₂O, Cu-N₄, Cu-O₄) were constructed^[121]. The FE of C₂H₄ for Cu₂O/CPFs was 61.8% at -1.3 V_RHE, significantly higher than that of pure MOF (32.6%) and Cu₂O (18.1%). The current density of C₂H₄ was -7.96 mA cm^-2, far exceeding that of CPFs and Cu₂O, respectively. The presence of the built-in electric field between Cu₂O and CPFs promoted the rapid transfer of electrons from CPFs to Cu₂O and was beneficial to the stabilization of Cu-O₄ sites in MOFs. Meanwhile, Cu₂O and Cu-O₄ sites were connected by Cu²⁺-O-Cu⁺-O bond, promoting adsorption, consumption of reactant molecules, as well as C-C coupling of *CHO and *CO, and finally C₂H₄ products could be generated at Cu₂O and Cu-N₄ sites.

Owing to slow C-C coupling reaction kinetics and complex intermediates, C₂H₅OH selectivity over Cu-based catalysts is usually relatively low. To further improve the C₂H₅OH selectivity of the Cu-based electrocatalyst, the construction of asymmetric refined structures can enhance the charge polarization effect, which acts as a critical factor in promoting the proton-coupled electron transfer (PCET) process of CO₂ to produce C₂H₅OH^[122]. Zhang et al. coated Cu₂O NPs with a NiCu-MOF grown on a Cu foam to synthesize Cu₂O@MOF/CF with an asymmetric fine structure^[123]. Cu₂O@MOF/CF could serve as a cathode for CO₂ reduction, providing high FE of 44.3% for C₂H₅OH at -0.615 V vs. RHE. The internal electric field induced by the existence of multiple self-polarization units in Cu₂O@MOF/CF led to the asymmetric distribution of electrons and promoted C-C coupling process in CO₂RR. The free energy of *OCCOH formation (1.26 eV) at the Cu site of Cu₂O in Cu₂O@MOF/CF was significantly lower than that of pure Cu₂O/CF (2.13 eV) and Cu₂O@CuMOF/CF (1.39 eV), indicating that the dimerization of *OCCOH tended to proceed at the Cu site of Cu₂O and the introduction of MOF contributed to the C₂H₅OH formation.

The unusual flaky structure and electronic characteristics of 2D MOFs lead to several advantages in the construction of novel high-performance composites^[118,119]. Liu et al. employed bismuth oxide (Bi₂O₃) NWs as active sites to be uniformly deposited on a metal-organic layer (MOL) to prepare a Bi₂O₃/MOL composite for CO₂RR^[118]. This composite had FE of > 85% for HCOOH at potential of -0.87~-1.17 V_RHE, far more than that of Zr-based MOF (Bi₂O₃/UiO). The structural and chemical stability of the MOL and NWs allowed the composite to maintain its FE at -0.97 V for more than 21 h. Ultrafine CuO NPs were uniformly deposited on a 2D copper 1,4-dicarboxybenzene (1,4-BDC) MOF by hydrothermal method to prepare CuO/Cu-MOF composites^[119]. The Cu-MOF contained a large number of channels with pore structures larger than the molecular size of CO₂. It led to a CO₂ adsorption capacity of the CuO/Cu-MOF composite at 1.0 atm to be ~5.0 mg g_cat^-1, significantly higher than that of the commercially available CuO NPs. The C₂H₄ FE for this composite reached 50.0% at potential of -1.1 V_RHE, compared to CuO (25.5%) and Cu-MOF (37.6%). The improved electrocatalytic performance of the composite came from the improvement of adsorption and activation of CO₂ molecules at the interface between Cu-MOF and CuO.

Functional-molecule-MOFs composites

The electron-rich units are embedded into the MOF to construct MOF-based composites with multiple electron transfer channels, which change the charge distribution around the active sites of MOF and further improve the performance of CO₂RR. Metallocene (MCp₂) is an electron-rich molecule with 18 delocalized electrons due to the presence of two cyclopentadienes (Cp) and a transition metal, making it a good electron donor. Meanwhile, MOFs with metalloporphyrin (M-TCPP), such as MOF-545-Co, may be a suitable electron acceptor. The metallocene molecule (CoCp₂) modification into the MOF-545-Co channel (CoCp₂@MOF-545-Co) overlapped with the large ring conjugated π-electron system of the Co-TCPP group to increase its electron density, which greatly facilitated the electron transfer from the Co-active center to CO₂, and improved the catalytic efficiency and selectivity [Figure 7A]^[124]. Although the CO₂ adsorption capacity of the composite decreased due to the intercalation of CoCp₂, the adsorption energy of CO₂ for CoCp₂@MOF-545-Co (-0.92 eV) was nearly four times higher than that of MOF-545-Co (-0.24 eV) because of the strong interaction between CoP₂ and Co-TCPP via C-Co bond (C in CoCp₂ and Co in Co-TCPP) [Figure 7B and C]. As a result, FE of CO for CoCp₂@MOF-545-Co composite was 97% at -0.7 V vs. RHE, much higher than that of MOF-545-Co (54.6%) [Figure 7D]. Moreover, polypyrrole (PPy) is a highly conducting polymer that can be used as an ideal additive for hybridization with MOFs to achieve efficient CO₂ conversion. Xin et al. inserted PPy molecules into the channels of MOF-545-Co to synthesize PPy@MOF-545-Co, improving the electron transport capability due to the strong π-π interaction [Figure 7E]^[125]. Therefore, FE of CO for PPy@MOF-545-Co was 98% at -0.8 V vs. RHE, twice that of MOF-545-Co [Figure 7F].

Figure 7. (A) Comparison of MCp₂@MOF and MOF for CO₂RR. (B) CO₂ adsorption at 298 K. (C) The free energy of CO₂RR to CO for Co-TCPP and CoCp₂/Co-TCPP. (D) FE for CO and H₂^[124]. Copyright 2020 Elsevier. (E) Schematic presentation of PPy in the channel of MOF-545-Co. (F) FE_CO at different voltages^[125]. Copyright 2021 the American Chemical Society.

For Zn-MOFs (such as ZIF-8), the catalytic active site for CO₂ reduction is located on the ligand coordinated with Zn; therefore, the electrocatalytic activity of such catalysts can be improved by adjusting the ligands. As a nitrogen-containing heterocyclic compound, 1,10-phenanthroline has strong electron donating ability. Doping it into the ligand of ZIF-8 enables the formation of stable coordination bonds with metal ions in MOFs, thereby enhancing the stability and functionality of the material^[126]. The sp² C atom in methylimidazole ligand of ZIF-8 was the catalytic active site for CO₂RR, which was still the best catalytic active site after doping with phenanthroline molecules. Due to the doping of the electron donor unit, charge density of catalytic active site was significantly enhanced (3.702 to 4.884), so that the doped ZIF-8 (ZIF-A-LD) exhibited excellent CO FE (90.57%) at -1.1 V vs. RHE. The theoretical calculation further made known that doping of electron-donating molecules reduced the formation energy of intermediate *COOH on the catalytic active site and promoted the activation of CO₂ molecules, thus improving its CO₂ electrocatalytic activity.

Heteroatom-doped carbon materials

Porous carbon materials have been widely employed in energy storage equipment due to their large SSA, stability, wide source of raw materials, and environment-friendly aspects^[131]. Porous carbons prepared from MOFs as precursors have the advantages of controllable porous structure and high SSA. MOFs can also be used to prepare heteroatom-doped porous carbon, and the existence of heteroatoms can improve the electron density and interface wettability of the porous carbon material, which is conducive to the supply of electrolyte ions, thereby improving the electrochemical performance of the porous carbon.

As a low-cost material, N-doped carbon (NC) materials have a high SSA and adjustable electrochemical activity and conductivity, along with a strong inhibition of HER, thereby attracting much attention as electrocatalysts for CO₂RR. Guo et al. prepared ZIF-8/MWCNTs by growing ZIF-8 on multi-walled carbon nanotubes (MWCNTs), and pyrolyzed them to prepare NC materials (ZIF-CNT-FA-p) for selective electrocatalysis of CO₂ to CO against the HER^[132]. This composite had a FE of 100% and total current density of 7.7 mA cm^-2 at -0.86 V vs. RHE. The superior selectivity could originate from the enhancement of electron transport and CO₂ mass transfer exhibited by the MWCNTs.

The existence of the N heteroatom can change electronic environment of the C atom, promote CO₂ adsorption, and stabilize intermediate. It has also been found that the number and types of the doping N atoms determine the number of active centers, and ultimately affect the catalytic activity of the material^[133]. Mesoporous NC (MNC-D) materials with adjustable structures and different N dopants have been constructed for CO₂RR through the pyrolysis of ZIF-8, followed by treatment with N, N-dimethylformamide (DMF)^[134]. MNC could efficiently promote the conversion process from CO₂ to CO with FE of ~92% and current density of -6.8 mA cm^-2 at -0.58 V_RHE. Further studies showed that the pyridinic-N and defects produced by DMF treatment acted as active centers of the MNC catalyst, which was beneficial to adsorption and activation of CO₂ molecules, further promoting electrocatalysis. In general, co-doping of pyridinic and graphitic N (active N) can synergistically improve the electrochemical CO₂RR performance. Porous NC materials with a large number of active N sites have been prepared via calcinating O-rich MOF (Zn-MOF-74) and melamine (C₃H₆N₆)^[135]. The active N content in the composite was adjusted by regulating temperature and time. This catalyst well inhibited the HER, and had high FE_CO of 98.4% at -0.55 Vvs. RHE for the sample prepared at 1,000 °C (NPC-1000). The lower number of pyridinic-N had a significant effect on reducing the absolute overpotential^[128]. NC materials, rich in pyridinic-N and quaternary-N species, have been prepared by pyrolysis of ZIF-8 at 700~900 °C in N₂ atmosphere^[136]. The CO FE for the samples pyrolyzed at 900 °C reached ~78%. The pyridinic-N and quaternary-N in the carbon matrix lowered energy barriers of the COOH* for the CO generation, leading to its high selectivity.

Metal/metal oxide interfaces

MOF-derived metal/metal oxide interfaces have a unique porous micro-/nanostructure and a high SSA. Their special morphology is conducive to uniform distribution of reactant molecules, adsorption, and activation, resulting in high selectivity and long cycle life for CO₂RR^[129,137]. A series of Cu/Cu₂O hybrids with porous octahedral structures and varying ratios of Cu⁰ and Cu⁺ were prepared by the pyrolysis of Cu-MOF octahedrons in N₂ atmosphere at different temperatures and heating rates^[138]. The optimal catalyst (Cu-MOF_20/300) with higher SSA (129.11 m² g^-1) and more Cu⁺ sites could reduce CO₂ to CO and H₂, where the highest FE of CO was 43.8% at -0.76 V vs. RHE for H-type cells. It was worth noting that compared with H-type fuel cells, the activity of the catalyst in the flow MEA reactor was significantly improved due to the smaller gas diffusion resistance. The current density was 34.97 mA cm^-2 at -0.64 V_RHE, ~3 times higher than that of in H-type cells. Moreover, the energy loss was also greatly reduced due to the shorter distance between the electrodes in the MEA reactor.

Transition metal oxides as electrocatalysts may have low product selectivity and stability due to the deactivation of the cathode; the catalytic efficiency can be further improved by exposing more active sites through the construction of bimetallic oxides. Porous bimetallic oxides with desired composition and structure by MOF pyrolysis can be synthesized to facilitate the electrochemical reaction^[129,139]. Payra et al.^[129] prepared cerium dioxide (CeO₂) and cerium-doped titanium oxide (Ce_1-xTi_xO₂) spherical particles by thermal decomposition of Ce-UiO-66^[140,141] and Ce_1-xTi_x-UiO-66-NH₂ MOFs for CO₂RR^[142]. CeO₂ served as an electrocatalyst to efficiently reduce CO₂ to CH₃COOH, whereas Ce_1-xTi_xO₂ exhibited excellent inhibitory effect on HER. MOF-derived porous indium (In)-Cu bimetallic oxides with adjustable Cu/In ratios have been employed as efficient electrocatalysts (InCuO-x, x denotes the Cu/In molar ratio) to reduce CO₂ to CO^[139]. InCuO-0.92 had a CO₂ adsorption capacity of 59.4 mg g^-1 at 298 K, 5.7 times greater than that of In₂O₃. The smaller Nyquist plots indicated efficient electron migration between In₂O₃ and Cu₂O during the conversion of CO₂, consistent with the electrocatalytic performance of the composite^[143]. Therefore, a good electrocatalytic performance could be obtained by the synergistic effects of In and Cu oxides, which promoted CO₂ adsorption capacity, mass transport, and charge transfer. The HER could also be inhibited by adjusting Cu/In ratios in MOFs. The InCuO-0.92 composite had 92.1% of FE for CO and total current density of 11.2 mA cm^-2 at -0.8 V_RHE. Additionally, InCuO-0.92 showed no cathode deactivation and significant decrease in activity and selectivity during the 24 h continuous test, indicating better stability.

Metal/metal-oxide/carbon nanocomposite interfaces

The carbonization of MOFs provides a direct method for fabricating functional and porous metal/metal-oxide/carbon hybrid materials. Through high-temperature treatment, the MOFs can be directly calcinated into porous carbon materials, and the metal centers can be in situ incorporated into the carbon matrix^[144,145]. Thus, the structure provides an excellent electronic connection between the metal/metal oxide interface and carbon skeleton.

Metal NPs deposited on carbon materials

Due to their significant electronic conductivity and high SSA, porous carbons can be used as substrates to load metal NPs^[146], which can prevent NP aggregation, and enhance interface contact to provide better electrical conductivity. Therefore, increasing the density of active sites and interactions between metal NPs and carbon nanomaterials can further improve the catalytic performance^[147]. MOFs are considered a promising precursor to synthesize metal NPs that can be distributed on the carbon substrate. For example, Bi NPs (~4.4 nm) have been grown in situ and uniformly distributed on functionalized MWCNTs to prepare Bi NP@MWCNT catalysts for CO₂RR^[146][Figure 8A and B]. For this composite, the optimal FE and current density for HCOOH reached 95.2% and 10.7 mA cm^-2, respectively, at -1.5 V vs. SCE (saturated calomel reference electrode) [Figure 8C and D]. It was speculated that the MWCNTs in this composite significantly increased the charge density on Bi NPs, promoting the rapid transfer of electrons from Bi to CO₂, thus facilitating the selective conversion of CO₂ to HCOOH^[148]. Moreover, Ni NPs coated with carbon loaded onto NC (Ni-NC_ATPA@C) could effectively reduce CO₂ to CO with FE of ~94% at overpotential of 0.59 V^[149]. At -1.1 V_RHE, the CO current density and TOF of the catalyst reached 22.7 mA cm^-2 and 697 h^-1, respectively. The direct contact between Ni NPs and aqueous electrolytes could be avoided due to the encapsulation of Ni NPs by carbon, which could significantly inhibit HER. The experimental results combined with the DFT calculations revealed that the carbon layer coated with Ni and N stabilized *COOH without affecting the ability to easily desorb *CO, which acted as a critical factor in promoting electrocatalytic performance.

Figure 8. (A) The preparation process of Bi NP@MWCNT; (B) TEM image, (C) FE and (D) Current density of HCOOH for Bi NP@MWCNTs at different potentials^[146]. Copyright 2020 the American Chemical Society. (E) The preparation process of Ni_1/150NCs@NC; (F) CO current density and (G) Stability. (H) The calculated free energy of CO₂RR^[151]. Copyright 2023 Springer Nature.

In previous studies, it was found that smaller metal NPs are more likely to produce *COOH with lower energy. In addition, small metal clusters exhibit thermodynamic stability to avoid aggregation, thereby leading to high catalytic efficiency^[150]. Wang et al. employed a Ni-Zn bimetallic MOF as a precursor to prepare Ni nanoclusters with a high dispersion of ~2 nm on NC material^[151]. The particle size and content of Ni catalyst can be effectively regulated by adjusting Ni:Zn ratios in MOF precursors [Figure 8E]. The Ni nanoclusters (diameter = 1.9 nm) loaded on NC material from a MOF with a Ni:Zn ratio of 1:150 created a catalyst (Ni_1/150NCs@NC) with FE of 98.7% for CO and partial current density of 40.4 mA cm^-2 at -0.88 V [Figure 8F], attributing to the enrichment of CO₂ originating from hybridization of *COOH and Ni and enhanced electron mobility from Ni to NC. The catalyst also showed long-term stability (~40 h) without decay of activity and selectivity and Ni aggregation [Figure 8G]. The DFT calculations revealed that Ni particles supported on NC facilitated the formation of *COOH, which promoted the process of CO₂-to-CO [Figure 8H].

The reduction of CO₂ to syngas (CO and H₂) in a specific ratio is an important industrial feedstock for the synthesis of some chemicals (e.g., methanol, acetic acid, dimethyl ether, etc.)^[152]. HER and CO₂RR are two competitive processes, so regulating the adsorption energy of catalysts for H⁺ and CO₂ is the key to producing CO/H₂ with an appropriate ratio^[153]. It was reported that by optimizing the electronic structure of metal alloy composites such as Cu-Co alloys, the adsorption energy of H⁺ and CO₂ could be adjusted, and then a specific ratio of CO and H₂ could be produced. Song et al. introduced Cu²⁺ into the synthesis process of ZIF-67 to synthesize bimetallic CoCu-MOFs, which were carbonized at high temperatures to synthesize CoCu@C composite catalysts^[154]. The FE and current density of CO for the optimal sample Co₁Cu₃@C (34%, 11.67 mA cm^-2) at -0.7 V_RHE were significantly higher than those of Co@C (28%, 3.5 mA cm^-2), originating from the rapid transfer of electrons from Co to Cu in the Co-Cu alloys. It was worth noting that the ratio of CO:H₂ generated for Co₁Cu₃@C was about 1:1.7~1:4, which met the industrial demand. In the CO₂RR process, a CO₂ molecule combined with one electron to synthesize *CO₂^-, which captured H⁺ to form *CO₂H. It was observed that there was a strong interaction between the carbon atoms in *CO₂H and metal sites on the catalyst surface by DFT calculation, which facilitated the conversion of *CO₂H to *COOH, thus promoting the conversion of CO₂ to CO.

The Bi-based catalyst shows better selectivity and current density in HCOOH^[155], but due to the presence of competing HER, the FE_HCOOH at low overpotential is low, and can only exceed 90% at a narrow potential window (< 750 mV)^[156]. Combining different metals with Bi to form alloys by doping strategy can adjust the electron cloud distribution of the catalyst, thereby affecting its affinity for reactants and the stability of reaction intermediates, and further improving the generation efficiency of HCOOH. Ma et al. substituted partial Bi sites in Bi-MOF with Sb to prepare Sb-doped Bi-MOF (Sb_x/Bi-MOF)^[157], which was carbonized to prepare Sb_x/Bi@C. Sb_2.5/Bi@C with Sb doping of 2.5% showed excellent FE_HCOOH of 94.8% at -1.4 V_Ag/AgCl, far exceeding that of Bi@C. Furthermore, the FE_HCOOH of Sb_2.5/Bi@C remained above 92% in the range from -1.3 to -1.6 V. The doping of a small amount of Sb could effectively promoted the generation of *OCHO and inhibited the formation of *H, thus significantly enhancing the selective formation of HCOOH in Bi materials. Furthermore, a series of Ce-doped Bi@C nanorod (NR) electrocatalysts were prepared by pyrolysis of Ce-doped Bi-MOFs^[158]. Firstly, due to the doping of Ce, the number of active sites and adsorption of CO₂ were increased; secondly, the increase of electron density around Bi sites promoted rapid electron transport and reduced the formation energy barrier of *OCHO. Therefore, Ce_0.05Bi_0.95@C NRs achieved high FE_HCOOH of 96.1% at -1.5 V_RHE, accompanied by high stability for 36 h. It was important that FE_HCOOH could exceed 90% at an ultra-wide potential window (1,000 mV).

Carbon-supported metal composite catalysts are known to be quite active in producing C₂ and C₃ products due to the high CO₂ pressure that can be generated at the three-phase boundary^[159]. Zhao et al. prepared an oxide-derived (OD) Cu/C catalyst by carbonization of HKUST-1 precursors^[159]. The FE of the obtained composite for CO₂ conversion to alcohols was in the range of 45.2%-71.2% at potential of -0.1~-0.7 V_RHE. For OD Cu/C-1000 catalyst obtained by carbonizing HKUST-1 at 1,000 °C, the CH₃OH and C₂H₅OH production rates were 5.1~12.4 and 3.7~13.4 mg L^-1 h^-1, respectively. Equally notable was that the initial potential for generation of C₂H₅OH was ~-0.1 V_RHE, corresponding to overpotential of ~190 mV. The increased performance of OD Cu/C originated from interaction between the uniform distribution of Cu and porous carbon matrices.

Metal oxides deposited on carbon

Carbon nanotubes (CNTs) are a type of carbon material with graphene layers that exhibit semiconducting/metallic properties and tubular morphology. Theoretical studies have shown that the curling of the graphene layer results in an inward-to-outward shift of π-electron density, resulting in an electron-rich state on the outer surface. The catalyst dispersed into the CNT cavities exhibits a confinement effect, which affects the mass transfer and distribution of the reactants, thus changing the catalyst redox properties. The ultra-small cobalt (II, III) oxide (Co₃O₄) NPs with a high dispersion were confined within CNTs to prepare a composite electrocatalyst (Co/CNTs) by the pyrolysis of ZIF-67 to effectively reduce CO₂ to CO^[160]. Owing to the excellent nanostructure of the Co/CNTs, Co₃O₄ NPs served as the maximally exposed primary active sites, greatly improving the electrocatalytic efficiency. At the cathode potential of -0.7 V_RHE, the Co/CNTs obtained 90% FE_CO and current density of 20.6 mA cm^-2. In addition, the CNTs prevented the aggregation of the Co₃O₄ NPs in the electrolysis process, thus enabling the catalyst to remain stable for a long period (40 h). Similarly, Deng et al. used Bi-MOFs as sacrificial precursors to prepare a composite catalyst (Bi₂O₃@C-800) for the reduction of CO₂^[161]. The composite exhibited a stable FE of 93% for HCOOH at an initial potential of -0.28 V_RHE and a high partial current density of 208 mA cm^-2 at -1.1 V_RHE in a flow-cell configuration. The carbon matrix improved the activity and selectivity, while oxides were favorable to the promotion of reaction kinetics and selectivity.

Different structural catalysts with tailored porosities can be produced through MOF pyrolysis, and treatment using different etching reagents provides efficient ways to generate specific active sites^[130]. Sn/SnO₂/C composites with exposed Sn (101) faces have been synthesized by the carbonization of a Sn-based MOF under an argon (Ar) atmosphere and different temperatures as CO₂RR electrocatalysts^[130][Figure 9A]. The Sn (101)/SnO₂/C-500 (pyrolysis at 500 °C) catalyst had 93.3% FE and current density of 8.2 mA cm^-2 for HCOOH at -0.8 V_RHE with no performance decay for 13 h [Figure 9B and C]. With the increase of calcination time (30 to 120 min), the peak intensity of Sn (101) in Sn (101)/SnO₂/C-500 gradually rose, while the FE of HCOOH improved from 45.4 to 93.3% at -0.8 V_RHE. This suggested that the Sn (101) could be used as an active site to improve the selectivity of HCOOH. In order to show the superiority of Sn (101) for CO₂RR, Sn (200) was used as a comparative sample and further verified by DFT. The activation of CO₂ to *OCHO occurred spontaneously on both Sn (101) and Sn (200). However, the binding energy for *OCHO on Sn (101) was -0.28 eV, significantly stronger than that on Sn (200) (-0.08 eV), so it was more conducive to the transformation of *OCHO to *HCOOH on Sn (101). Simultaneously, the adsorption energy of H* on Sn (101) (0.23 eV) was much higher than that of *OCHO, proving that Sn (101) could suppress HER well [Figure 9D].

Figure 9. (A) The Sn (101)/SnO₂/C-500 preparation process for CO₂RR. (B) FE and (C) long-term stability of Sn (101)/SnO₂/C-500. (D) The free energy of HCOO^-, CO formation^[130]. Copyright 2021 the American Chemical Society.

Single-atom catalysts

In recent years, transition-metal SACs (TM-SACs) have been extensively studied owing to their superior electrocatalytic performance, low cost, and stability. TM-SACs can act as adsorption and activation sites for H₂O, O₂, CO₂, etc., and facilitate redox of adsorbed molecules^[162]. The choice of support is particularly important for the formation of TM-SACs because individual atoms tend to agglomerate in the preparation process. In addition, the electron and mass transfer during electrocatalytic reaction are also affected by the substrate. With the advantages of adjustable structure and various available modification methods, MOF precursors/templates can be potentially employed in the synthesis of TM-SACs with high loading capacity, stability, and electrocatalytic activity^[163].

Well-defined metal-nitrogen (M-N_x) sites confined into a carbon matrix (M-N-C) have been successfully obtained by the thermal decomposition of MOFs^[164]. A Fe-doped MOF has been used to synthesize a carbon nano-skeleton with a hierarchical pore structure (micropores to large mesopores) and atomically dispersed metalloporphyrin-like Fe-N₄ active centers^[165]. The presence of mesopores in the nano-framework improved the amount and accessibility of monoatomic active sites and facilitated mass and charge transport. The FE for this composite reached 86.9% at -0.47 V_RHE for CO. Isolated M-N_x sites could serve as active centers for CO₂RR. However, the content of the isolated M-N_x active sites in the carbon matrix was generally limited, and they can be maximally exposed using various feasible strategies to achieve efficient catalysis. The functionalization of ZIF-8 NPs was first achieved by the selective restriction of ammonium ferric citrate (C₆H₁₁FeNO₇, AFC) on its surface using a post-synthesis modification strategy (PSMS)^[166]. Subsequently, isolated Fe-N sites were confined into the carbon matrix after the MOF thermal decomposition (C-AFC©ZIF-8). The maximally exposed Fe-N active sites showed excellent selectivity and mass transfer ability for conversion of CO₂ to CO with FE of 91.6% at -0.63 V_RHE, superior to most reported noble metal catalysts. The study of CO₂RR mechanism and the site specificity of TM-SACs could instruct the design of M-N-C electrocatalysts to obtain higher product selectivity of CO₂RR. The Fe-N-C catalyst prepared by employing ZIF-8 as a template had FE of > 93.5% for CO at 90 mV of overpotential^[167]. The location of Fe-N₄ may influence the activity of CO₂ reduction, which was verified by DFT calculation [Figure 10A]. The conversion of CO₂ to *COOH on Fe-N₄ groups at the bulk (bulk-Fe-N₄) and edge (edge-Fe-N₄) sites of graphite layers without defects occurred spontaneously, which was beneficial to the CO₂ reduction at low overpotential. However, the adsorption energy of the intermediate *CO on bulk-Fe-N₄ (-1.71 eV) and edge-Fe-N₄(-1.55 eV) was particularly high, which was not conducive to the desorption of CO, thus causing the poisoning of the Fe center. However, for Fe-N₄ in the graphite layer with defects (Fe-N₄-pore), the low energy barrier (0.30 eV) of the CO₂-to-*COOH process could ensure the realization of CO₂RR at low overpotential, while the small barrier (0.37 eV) of *CO-to-CO conversion could effectively promote the CO desorption [Figure 10B]. Therefore, it can be considered that Fe atoms in Fe-N₄-pore were the active sites of CO₂RR. The more appropriate binding strength of Fe-N₄-pore for the intermediates *COOH and *CO promoted the conversion of CO₂ to CO.

Figure 10. (A) Fe-N₄ moiety in the defective porous graphitic layer (Fe-N₄ pore). (B) Gibbs free energy on various sites^[167]. Copyright 2019 the American Chemical Society. (C) The preparation process of C-Zn_xNi_y ZIF-8. (D) FE and (E) current density of CO at different potentials^[168]. Copyright 2018 Royal Society of Chemistry. (F) The synthesis of M₁-N-C by pyrolysis of MTV-MOFs. (G) FE of CO for M₁-N-C in pure CO₂. DFT calculations of (H) reaction paths and (I) free energy^[172]. Copyright 2020 Wiley-VCH.

Although M-N-C as an effective electrocatalyst can promote the process of CO₂-to-CO, the current of the competitive HER grows sharply with the increase of the overpotential, resulting in a rapid decline in the FE of CO, and it is difficult to obtain a high partial current density of CO. Therefore, obtaining high current density and FE for CO₂RR simultaneously is an important challenge for M-N-C. Based on the above considerations, Yan et al. embedded unsaturated Ni-N units into porous carbon by pyrolyzing ZIF-8 doped with different Ni contents (Zn_xNi_y ZIF-8) [Figure 10C]^[168]. The FE of CO for all the C-Zn_xNi_y ZIF-8 could be maintained 92.0%~98.0% at a range of -0.53~-1.03 V vs. RHE, and the current density of CO increased with potential, reaching 71.5 ± 2.9 mA cm^-2 at -1.03 V_RHE [Figure 10D and E]. The theoretical calculations clarified that unsaturated Ni-N active sites possessed lower *COOH free energy and higher *H adsorption energy, thus exhibiting high CO₂RR activity and inhibiting competitive HER. The catalyst breaks the "seesaw" effect limitation of selectivity and activity on the M-N-C material, and realizes efficient catalysis of CO₂RR. In addition, Ni-N₄ sites have two forms: embedded at the edge (Ni-N₂₊₂) and in the bulk (Ni-N₄) of the carbon matrix^[169,170]. However, the catalytic active sites for CO₂RR between these two types are difficult to distinguish. Over these considerations, Pan et al. used ZIF-based Ni-N-C catalyst to achieve highly selective conversion of CO₂ to CO with FE of 96% at overpotential of 570 mV^[171]. The DFT calculations revealed that the Ni-N₂₊₂-C₈ with active C atoms possessing dangling bonds were considered to be the active sites due to lower activation energy in the dissociation of COOH* intermediates. In contrast, Ni-N₄-C₁₀, which was tightly embedded into graphite layers, may not be active. Moreover, Ni-N₂₊₂-C₈ could efficiently inhibit the competitive HER as well.

Influence of coordination metal

Owing to the diversity in MOF structures and compositions, M-N-C catalysts containing different metal elements could be constructed using the tunability of the ligand central metal. By only changing the types of monatomic metals, the physicochemical properties of MOF-derived monatomic materials can be strictly controlled, thus creating a prerequisite for identifying and comparing the catalytic performance^[172].

M-N-C with different M-N₄ groups (M = Mn, Fe, Co, Ni and Cu) was prepared to explore the influence of metal centers on electrocatalytic performance^[173]. It was found that the current density of CO₂ reduction to CO was significantly different among the five catalysts, in which Co-N-C showed the best intrinsic activity (at -0.6 V_RHE), followed by Fe-N-C, Ni-N-C, Mn-N-C and Cu-N-C. Due to similar structure of these five model catalysts, the binding affinity of metal sites for the key intermediates might be the reason for the difference in the electrocatalytic CO₂RR activity. The binding ability of Mn-N₄ and Fe-N₄ to CO was too strong, so the desorption of CO* determined its activity. However, Ni-N₄ was weakly bound to CO, so the activation of CO₂ to form CO₂*^- was the main process affecting their activity. The more suitable energy barrier of Co-N₄ for CO₂*^-and CO* was the reason for its best activity. Mn-O₄-O could be formed due to the strong adsorption of Mn-N₄ on O, and Mn-O₄-O had a high energy barrier for CO₂ adsorption, which explained the poor activity of Mn-N-C for CO₂RR. For Cu-N-C, the central Cu ion was reduced to Cu (0), resulting in low activity. In terms of selectivity, the FEs of Fe-N-C, Ni-N-C and Mn-N-C were > 80%, whereas Co-N-C exhibited lower selectivity (48%), which was attributed to the different binding energies between CO₂*^- and H*.

The structural and functional diversity of multivariate MOFs (MTV-MOFs) provides the possibility to construct M-N-C with multiple metal centers to achieve electrocatalytic conversion under low CO₂ concentration. Based on this, a series of MOFs with metal species at the center have been constructed by introducing both porphyrin and metalloporphyrin ligands using a mixed ligand strategy^[172]. M-N-C (M = Fe, Co, Ni, and Cu) with the same chemical environment and substrate was prepared using MTV-MOFs with porphyrin groups as templates [Figure 10F]. The FE of Ni₁-N-C catalyst reached 96.8% for CO, far exceeding that of Fe₁-, Co₁- and Cu₁-N-C in pure CO₂ [Figure 10G]. The free energy calculation for the intermediate path from CO₂ to CO towards different catalysts revealed that free energy of COOH* formation along with CO desorption in the RDS of Ni₁-N-C was optimized [Figure 10H and I]. Ni₁-N-C inhibited the competitive HER most effectively, explaining the high FE for the CO₂RR. Furthermore, upon application for the more challenging reduction of low-concentration CO₂ (15%), Ni₁-N-C maintained 80% of FE for CO.

Effect of coordination environment

The metal active center in M-N-C interacts with its surrounding atoms to form a certain coordination environment, which can regulate electron density of the metal site, and then affect adsorption behavior of key reaction intermediates on the monatomic active center, thereby optimizing catalytic performance^[174,175]. At present, the change of coordination environment of metal active centers mainly includes the regulation of number, species of coordination atoms, and cooperation of adjacent metal sites.

Non-metallic POPs

Conjugated triazine frameworks (CTFs), as porous polymeric materials linked by strong triazine bonds, possess favorable porous structure, high stability, and intrinsic N-rich doping. In addition, the conjugated structure of CTFs is beneficial to electron transport and mass transfer, so they are considered as potential electrode materials for the electrocatalytic reduction of CO₂^[190]. Zhu et al. prepared triazine-based frameworks (TTF-1 and TTF-2) by ZnCl₂ ionothermal method at 600 °C using 2,6-dicyanopyridine and 1,3-dicyanobenzene as monomers, respectively^[190]. The abundant porous structure and large SSA of TTF-1 (1,234 m² g^-1) and TTF-2 (2,522 m² g^-1) endowed the electrode and electrolyte with a large contact surface, which was conducive to charge and mass transfer. However, the content of pyridinic N that served as active sites in TTF-1 was 13.8 atom%, much higher than that of TTF-2 (6.8 atom%). Therefore, CO FE of 82% was achieved at -0.68 V_RHE for TTF-1, far exceeding the highest CO FE of 22% at -0.43 V_RHE for TTF-2. In addition, TTF-1 could maintain about 75% FE_CO for 12 h, indicating its long-term stability.

Covalent organic frameworks (COFs) are a type of organic polymer with porous crystal structures and strong covalent bonds, and their properties such as CO₂ adsorption capacity and electronic conductivity can be achieved by adjusting the building units or linkages. It is noted that N sites (pyridine N, pyrrole N, etc.) are considered as active centers for electrocatalysis, while sp³ N sites show higher CO₂ adsorption capacity compared to sp² N sites^[191]. Therefore, the design of amine-linked crystalline COFs containing sp³ N can be used as electrocatalysts to achieve efficient CO₂ reduction. Based on the above considerations, Ke et al. introduced piperazine (Pz) linking motifs into the crystalline COF to construct TPTA-Pz-COF (TPTA: [1,1′:4′,1″-terphenyl]-3,3″,5,5″-tetracarbaldehyde) with aminal linkage and high density of sp³ N for efficient electrocatalytic reduction of CO₂ to C₂H₄^[192]. To demonstrate the superiority of the presence of sp³ N, an imine-based TPTA-PDA-COF with sp² N was prepared as a reference sample [Figure 12A]. TPTA-Pz-COF had a higher SSA of 1,470 m² g^-1 compared to TPTA-PDA-COF (1,110 m² g^-1) ensured a high electrochemical active surface area (ECSA) (12.7 μF cm^-2) [Figure 12B], while TPTA-Pz-COF showed strong CO₂ capture ability due to obvious adsorption hysteresis based on the presence of sp³ N, which can promote mass transfer in the CO₂ reduction process. As a result, TPTA-Pz-COF could convert CO₂ to C₂H₄ as a gas-phase product and C₂H₅OH and HCOOH as liquid-phase products with FE of 19.1%, 21.86%, and 3.94% at -1.0 V_RHE [Figure 12C]. However, TPTA-PDA-COF could not reduce CO₂ to C₂H₄ under the same conditions [Figure 12D], suggesting that sp³ N played a key role in the C₂H₄ production. It was also found that the adsorption energy of CO₂ on TPTA-Pz-COF was -0.263 eV, lower than that on TPTA-PDA-COF (-0.199 eV) [Figure 12E]. At the same time, the existence of high-density sp³ N sites in TPTA-Pz-COF was conducive to charge accumulation, which further facilitated the activation of CO₂ and intermediates. Moreover, good hydrophilicity of TPTA-Pz-COF promoted the formation of hydrogen bonds, which was beneficial to the stabilization of intermediates *COOH and *CO, and then the C-C coupling process.

Figure 12. (A) The construction of TPTA-Pz-COF and TPTA-PDA-COF based on different linkages. (B) N₂ adsorption isotherms of TPTA-Pz-COF. FE of H₂, CO, CH₄ and C₂H₄ for (C) TPTA-Pz-COF and (D) TPTA-PDA-COF. (E) CO₂ adsorption energies of TPTA-Pz-COF and TPTA-PDA-COF^[192]. Copyright 2024 Wiley-VCH. (F) The synthesis of N⁺-COF, NH-COF and N⁺-NH-COF. (G) CO FE. (H) The CO current density^[193]. Copyright 2023 Springer Nature.

The linkage and linker of COF are simultaneously modified by post-modification strategy to enhance the CO₂ binding capacity and conductivity, thus improving CO₂RR performance^[188]. Firstly, a crystalline COF (CoTAPP-PATA-COF) was prepared by utilizing 5,10,15,20-tetrakis(4-aminophenyl) porphinato]-cobalt (CoTAPP) and 4,4′,4″,4‴-(1,4-phenylenebis(azanetriyl)) tetrabenzaldehyde (PATA) as raw materials^[193]. Then, ionization of PATA was conducted by in-situ ammonium groups to form N⁺-COF, while NaBH₄ was used to reduce the C=N in COF to C-N to prepare N⁺-NH-COF [Figure 12F]. The conductivities of N⁺-NH-COF and N⁺-COF were 6.7 × 10^-9 and 8.1 × 10^-9 S m^-1, far exceeding that of NH-COF (3.0 × 10^-10 S m^-1), indicating that ionic modification was beneficial to electron transport and further improved the activity of CO₂ reduction. In addition, the minimum Tafel slope of N⁺-NH-COF indicated that the dual modification can significantly elevate CO₂RR kinetics. As a result, N⁺-NH-COF could efficiently convert CO₂ to CO with the highest FE of 97.32% at -0.8 V_RHE, which was higher than that of NH-COF (95.26%), N⁺-COF (92.07%) and CoTAPP-PATA-COF (81.75%) [Figure 12G]. The presence of C-N bond promoted the adsorption of CO₂, while the hydrophobicity of the catalyst can inhibit the competitive adsorption of water, thus improving the selectivity. Meanwhile, the best current density of CO was 28.01 mA cm^-2 for N⁺-NH-COF at -1.0 V, significantly higher than that of NH-COF (22.83 mA cm^-2), N⁺-COF (23.73 mA cm^-2), CoTAPP-PATA-COF (13.26 mA cm^-2) [Figure 12H]. The *COOH formation energy barrier on N⁺-NH-COF was lower than that of N⁺-COF, CoTAPP-PATA-COF and NH-COF, which promoted the *COOH→*CO process.

Metalized POPs

Metal complexation and PSMS are two common methods for preparing metalized POPs. The PSMS of POPs is generally achieved by introducing metal ions through cation exchange or coordination^[194]. Zhang et al. used a rigid N-containing ligand 2-(5-(3-(5-(pyridin-2-yl)-1H-1,2,4-triazol-3-yl) phenyl)-1H-1, 2, 4-triazol-3-yl) pyridine (H₂bptb) and Cu⁺ to synthesize two stable coordination polymers NNU-32 and NNU-33 (S) (where S = sulfate)^[195]. During CO₂RR, NNU-33 (S) can be converted to NNU-33 (H) (where H = hydroxyl radical) by the substitution of (SO₄^2-) by OH^- by anion exchange, further strengthening cuprophilic interactions within the catalyst structure. NNU-33 (H) efficiently catalyzed the process of CO₂-to-CH₄ with 82.17% FE at -0.9 V_RHE. In addition, the DFT calculations revealed that energy barrier in the *H₂COOH→*OCH₂ process could be lowered by the cuprophilic interactions, effectively promoting the CH₄ production. Moreover, it has been reported that the catalyst prepared by changing the ligand type can generate C₂ products during the CO₂RR^[196], and the selectivity of these products can also be adjusted by changing the distance between two adjacent metal atoms in the coordination polymer^[197]. Sakamoto et al. synthesized binuclear coordination polymers [Cu₂(μ-X)₂(PPh₃)₂(μ-BPY)]_n (CuX-BPY) (X = Cl, Br, I) with different distances between two adjacent Cu atoms by changing halogen bridging elements on which the catalytic properties of polymers depend^[197]. The catalytic efficiency of copper chloride CuCl-BPY with shorter Cu-Cu distance for C₂ product was > 8 times higher than that of copper iodide CuI-BPY. On the one hand, the binding force between Cu and Cl was weaker due to the smaller ionic radius of Cl compared with Br and I, so the bridged Cl was easily desorbed from Cu-μX-Cu, and each Cu atom was coordinated with two CO₂ molecules, facilitating C-C coupling reaction; On the contrary, the charge repulsion caused by larger radius of I increased the distance between two adjacent Cu atoms and reduced electron density around the Cu atoms, which was not conducive to the coordination of CO₂ on Cu atoms.

H₂Pc^[198] and porphyrin^[199,200] polymers containing SA centers (M-N₄) are promising electrocatalysts for CO₂RR, and their electrocatalytic activities can be enhanced by tuning ligands or metal centers. Wei et al. prepared a 2D conjugated NiPc polymer (NiPcP) with sheet structure under vacuum at high temperature of 400 °C, which served as an efficient electrocatalyst for CO₂RR^[201]. The conjugated structure endowed strong binding affinity between the Ni sites in the polymer and CO₂ and fast electron transfer within the catalyst, ensuring high selectivity (> 98%) of CO for NiPcP at potential of -0.15~-0.60 V vs. RHE. NiPcP had ~100% FE at -0.5 V, maximum CO current density of 232 mA cm^-2 at -0.60 V, and high TOF of 23,148 h^-1 at 0.39 V overpotential. The intrinsic electronic structure, conjugation properties of Ni, and hydrophobic properties of NiPcP promoted CO formation during CO₂RR and inhibited the competitive HER.

The strong π-π stacking between layers in polymer structure may cause the aggregation of the bulk structure, which hinders the full play of the catalytic activity of the catalytic center and reduces the electrocatalytic performance. The construction of an ultra-thin polymer structure may be a good solution^[198]. Song et al. prepared an ultrathin covalent organic polymer (COP) with a thickness of ~1.7 nm by condensation of Co tetraaminophthalocyanine (CoPc-(NH₂)₄) and squaric acid (SA) as monomers [Figure 13A and B]^[198]. However, the COP-BDA prepared by replacing SA with 1,4-benzenedicarboxaldehyde (BDA) exhibited a significant stacking structure with a thickness of > 4 nm [Figure 13C]. Compared with COP-BDA, ultrathin COP-SA ensured higher active site exposure and electron transfer efficiency. As a result, the CO FE of COP-SA reached 96.5% at -0.65 V_RHE, slightly higher than that of COP-BDA (84.4%) [Figure 13D]. It was worth noting that the CO partial current density of COP-SA was 8.16 mA cm^-2, far exceeding that of COP-BDA (0.59 mA cm^-2) [Figure 13E]. Moreover, the TOF of COP-SA was 46.3 s^-1, 30.8 times that of COP-BDA. A smaller Tafel slope of COP-SA (118 mV Dec^-1) compared to COP-BDA (265 mV Dec^-1) suggested that the ultrathin structure contributed to enhanced CO₂RR kinetics [Figure 13F]. In addition, theoretical calculations suggested that higher binding energy of the Co center with *COOH (0.86 eV) and the weaker binding energy with *CO (0.12 eV) favored the desorption of the product CO [Figure 13G].

Figure 13. (A) The synthesis of COP-SA and COP-BDA. TEM and AFM images of (B) COP-SA and (C) COP-BDA. (D) FE and (E) CO current density from -0.4 to -0.65 V_RHE. (F) Tafel plots. (G) The binding strength of *COOH^[198]. Copyright 2021 Elsevier.

The regulation of CO₂RR performance can be achieved by adjusting the microenvironment around the electrocatalytic site. Pz-linked NiPc COFs (NiPc-NH-TFPN, NiPc-NH-TFPN-COOH, NiPc-NH-TFPN-NH₂, TFPN: tetrafluoroterephthalonitrile) with different ligands comprising -CN, -COOH, and -CH₂NH₂ groups were prepared to achieve the regulation of microenvironment near the active sites^[202]. The presence of Pz linkage made COF have better conductivity and stability, while the adsorption capacity of NiPc-NH-TFPN-NH₂ (80.6 cm³ g^-1) for CO₂ at 298 K was significantly higher than that of NiPc-NH-TFPN (19.4 cm³ g^-1) and NiPc-NH-TFPN-COOH (38.2 cm³ g^-1) due to the interaction between -NH₂ groups in NiPc-NH-TFPN-NH₂ and CO₂. The FE of CO for these three catalysts from -0.4 to -0.7 V_RHE showed an increasing trend with the rise of potential. The highest CO FE of NiPcNH-TFPN-NH₂ was 99.6% at -1.0 V, higher than the maximum FE of NiPc-NH-TFPN-COOH (71.4%). Meanwhile, NiPc-NH-TFPN-NH₂ exhibited a higher CO partial current density compared to NiPc-NH-TFPN and NiPc-NH-TFPN-COOH. In addition, the FE_CO and current density of the NiPc-NH-TFPN-NH₂ at -0.8 V showed almost no decay within 60 h, indicating good long-term stability. Compared with -CN and -COOH as electron-withdrawing groups, the introduction of electron-donating group -CH₂NH₂ can make more electrons participate in the electrocatalytic process. Moreover, the energy barrier of *COOH formation near the active site (-CH₂NH₂) was 1.35 eV, lower than that of NiPc-NH-TFPN (1.63 eV) and NiPc-NH-TFPN-COOH (1.57 eV), promoting the formation of CO.

In addition, the reduction of H₂Pc ligands during the electroreduction of CO₂RR at high overpotential may greatly affect the long-term stability of the electrocatalyst. The COP-CoPc was prepared by reacting viologen ligands with different redox states with cobalt polyphthalocyanine (CoPc-NH₂), which showed good electrocatalytic selectivity (FE_CO > 90%) for CO₂RR at a range of -0.68~-1.28 V_RHE in the H-type cell, reaching up the highest FE_CO of 97.3% at -0.88 V^[203]. The CO current density of COP-CoPc reached 90 mA cm^-2 at -1.4 V_RHE, higher than that of CoPc (45 mA cm^-2) and CoPc-NH₂ (57 mA cm^-2). Remarkably, a high FE_CO of more than 95% can also be obtained under the condition of 150 mA cm^-2 in the MEA system, showing great application potential. More importantly, at a constant current of 0.25 A, COP-CoPc maintained more than 90% FE_CO over 48 h period.

A porphyrin complex with M-N₄ sites is also a promising CO₂RR electrocatalyst. Lu et al. prepared a Ni porphyrin CTF (NiPor-CTF) with NiN₄ units for the effective reduction of CO₂ to CO^[204]. The CO FE of NiPor -CTF exceeded 90% at -0.6~-0.9 V_RHE, reaching up the highest FE of 97% at -0.9 V with current density of 52.9 mA cm^-2. The calculations indicated that the kinetic energy barrier of *CO₂ to *COOH transition on NiN₄ active sites could be reduced. Furthermore, Wang et al. prepared a large π-conjugated polymer (CoPor-N₃) by coupling an electron-rich ligand (2, 6-bis (5-amino-1h-benzimidazole-2-yl) pyridine, N₃) with meso-tetra (p-formylphenyl) porphyrin cobalt (CoPor) containing Co-N₄ unit via strong covalent imine bonds (C=N)^[205]. H₂Por-N₃ with no Co sites only had a low FE_CO of 5.4% at -0.65 V_RHE, while the highest FE of the CoPor monomer for CO reached 78% at -0.55 V_RHE, suggesting that the high CO selectivity of CoPor-N₃ may be mainly contributed by the CoPor unit as an active site for CO₂RR. It was worth noting that CoPor-N₃ showed a significantly higher FE_CO of 96% at a lower potential of -0.50 V_RHE than that of CoPor, which was mainly attributed to coupling between electron-rich N₃ ligands and CoPor ensuring that electrons can be enriched around the Co center, facilitating the electron transfer from the Co center to CO₂ for activation and reduction. In addition, the superior optical driving characteristics of the CoPor unit enabled CoPor-N₃ to be used as a photosensitive electrocatalyst for the reduction of CO₂ to CO, so that partial current density of CO reached -42.1 mA cm^-2 at -0.60 V_RHE under illumination, which was higher than the partial current density (-28.2 mA cm^-2) with no illumination.

The use of different linkers can adjust the catalytic effect of metal centers and control CO₂RR performance to a certain extent. However, the linker itself is generally not catalytically active. The performance of CO₂RR can be improved in principle by linking metalloporphyrin groups with catalytic monomers (such as metal bipyridine) to synthesize polymers with two metal sites^[206,207]. Porphyrin-bipyridine bifunctional COF (COF_Re) was constructed by Schiff base reaction between bipyridine rhenium [obtained by the reaction of 2,2′-bipyridine-5,5′-formaldehyde and pentacarbonyl chlororhenium (I)] and 5,10,15,20-tetra(4-aminophenyl) porphyrin (TAPP)^[206]. Further modifications with cobalt chloride (CoCl₂) or ferric chloride (FeCl₃) led to the addition of the Co²⁺ or Fe³⁺ into the porphyrin unit, resulting in COFs with bimetallic sites (COF-Re_Co or COF-Re_Fe). COF-Re_Co had higher FE_CO of 18(2)%. The low selectivity may be related to the competition between the two metal centers, while the considerable distance between the two sites limited the migration of carriers. Based on the above considerations, a 3D bimetallic catalyst labeled as Co/Ni-TPNB-COF was constructed by covalently linking Ni(II) -porphyrin and Co(II)-porphyrin units with 4′,4″,4‴-nitrilotris (1,1′-biphenyl) (NB) units^[208], which showed higher FE_CO of 95% and TOF of 4.10 s^-1. Theoretical calculation and experimental characterization showed that the triphenylamine unit as a linker could promote the charge migration between two metal sites, as well as the charge redistribution, which could help improve the electrocatalytic activity.

The particle size of COF is also a critical factor affecting its electrocatalytic performance, which is usually limited by the large particle size caused by particle aggregation, while the precise control of the morphology of COF catalyst is still a difficult problem. Endo et al. employed a trityl group-protected cobalt porphyrin precursor [Co(ttpp), ttpp = 5, 10, 15, 20-tetrakis (4-(tritylamino) phenyl) porphinato (2-)] for the preparation of small particle COF (COF-366-Co)^[209]. Due to the trityl protection, Co(ttpp) exhibited good solubility in the solvent, which could inhibit intermolecular packing and uniform nucleation of COF particles; therefore, the size of the crystals (COF-T) prepared was reduced to 162 ± 41 nm, much smaller than that of the COF particles (COF-A, about 10 μm) prepared by the unprotected porphyrin precursor. As a result, the smaller size of COF showed excellent electrocatalytic performance of CO₂-to-CO with FE of 86%-95% at -0.58~-0.88 V vs. RHE, much greater than that of COF-A (38%~67%).

The binding strength between reactants and intermediates is considered to be one of the factors affecting the catalytic performance of COFs for CO₂RR. Many methods have been used to promote the charge transfer within the framework, thereby regulating the binding strength of intermediates, and ultimately improving the electrocatalytic performance. Bimetallic COFs with dioxin-linkage were synthesized to improve the electron transfer within the framework, guaranteeing FE_CO of 97% and TOF of 2.87 s^-1[210]. A conductive 2D H₂Pc COF with isolated Cu-H₂Pc active sites, as an electrocatalyst, can convert CO₂ to acetate with FE of 90.3(2)% and current density of 12.5 mA cm^-2 at -0.8 V vs. RHE^[211]. Theoretical calculations showed that isolated Cu-H₂Pc had a high electron density, ensuring the migration of d electrons from Cu to C. At the same time, due to the strong interaction between COF and *CH₃ intermediates, the C-C coupling between *CH₃ and CO₂ was promoted, and the formation of ethylene and ethanol by the interaction between *CO and other intermediates was avoided. Furthermore, the COF containing an imidazole group as the catalyst for electrocatalytic CO₂RR enhanced the stabilization of the key intermediate *COOH and the electron transfer, thereby facilitating the catalytic reaction and ultimately achieving nearly 100% FE_CO^[212,213]. Due to the single donor and acceptor units of the above COF, the regulation of the binding strength of intermediates was inaccurate. To accurately adjust the interaction between the catalyst and the intermediate, Liu et al. constructed a three-component COF by adding the electron-rich diarylamine unit, the electron-deficient benzothiazole and the Co-porphyrin unit^[214]. The COF had higher electronic conductivity and more efficient charge transfer, and thus high CO FE of 97.2% at -0.8 V vs. RHE, which was attributed to favorable *COOH formation and *CO desorption.

POPs/C composites

Although the electrocatalysts of metalized POPs have well-defined and tunable active sites, the efficiency of their CO₂RR is usually much lower than that of metal catalysts due to poor conductivity, which largely limits their practical applications. The combination of POPs with substrates to construct composites is seeking to confront poor electrical conductivity of POPs and obtain high electrocatalytic performance. The interaction of POPs and graphene or CNTs has been used to enhance the electron transport capability^[215]. Lu et al. prepared a composite material (COF-366-Co@CNTs) by in-situ polymerization of COF with Co porphyrin units on amino-functionalized CNTs via covalent bonding to ensure close contact between COFs and CNTs^[216], thereby promoting rapid electron transfer from the Co center to CNTs [Figure 14A]. To further adjust electronic structure of the Co center, COF-366-(OMe)₂-Co@CNT was prepared by introducing an electron-withdrawing group (methoxy) into the porphyrin units during the preparation of the composite. In addition, COF-366-(OMe)₂-Co@CNT had higher binding capacity for CO₂ due to the hydrophobicity of the methoxy group and CO₂. The ultrathin COF nanolayer with a thickness of 0.9 nm wrapped on the CNT greatly shortened the electron transport distance from the COF to the CNT [Figure 14B]. For these reasons, COF-366-(OMe)₂-Co@CNT showed the best performance of CO₂→CO, with FE of 93.6% at -0.68 V_RHE, current density of 40 mA cm^-2 at -1.05 V and TOF of 11,877 h^-1 at 770 mV overpotential [Figure 14C-E]. For the same purpose, the composites CoP@CNT, CoP-Ph@CNT and CoP-F@CNT were prepared by in-situ growth of covalent Co porphyrin polymers with different porphyrin substituents on CNTs through the coupling between the alkynyl groups of porphyrins and CNTs to adjust the electronic structure of the polymers^[217] [Figure 14F]. As a result, FE_CO of CoP@CNT (~95%) and CoP-Ph@CNT (~80%) was significantly higher than that of CoP-F@CNT (~20%) in a large potential window of -0.57~-0.87 V [Figure 14G], attributed to electron-rich Co^I center in CoP and CoP-Ph which facilitated the binding and activation of CO₂ to improve CO₂RR selectivity compared with electron-deficient CoP-F. In addition, the activity and selectivity of CoP@CNT hardly decreased at -0.57 V within 100 h, exhibiting excellent stability.

Figure 14. (A) The construction of COF-366-Co@CNT. (B) HRTEM images of COF-366-(OMe)₂-Co@CNT. (C) FEs (D) Current densities and (E) TOFs^[216]. Copyright 2020 the American Chemical Society. (F) The synthesis of CoP@CNT, CoP-Ph@CNT, and CoP-F@CNT. (G) FEs at -0.47~ -0.87 V^[217]. Copyright 2022 Chinese Chemical Society.

MePc and its derivatives as electrocatalysts are capable of selectively electroreducing CO₂ to CO. In contrast, the polymerization of metal H₂Pc can improve the stability of the catalyst, but it can not effectively control electronic properties and microenvironment of the metal active center, so the catalytic activity is often weaker than that of small H₂Pc molecules. Therefore, it is of great significance to regulate the polymerization of metal H₂Pc molecules so as to give consideration to both stability and activity in the selective electrocatalytic reduction reaction of CO₂. Guo et al. synthesized a CNT@conjugated microporous polymer (CMP) (CoPc-H₂Pc) composite electrocatalyst with both catalytic activity and stability by copolymerizing H₂Pc and cobalt H₂Pc (CoPc) molecules on the CNT surface through Scholl coupling reaction^[218]. The addition of H₂Pc in the polymerization process can not only effectively avoid the metal shedding of CoPc in the reaction, but also promote the dispersion of metal Co in the ultra-thin shell of CMP at atomic level. The catalytic performance results suggested that the material had the lowest onset potential, achieving CO current density of 15.2 mA cm^-2 at -0.6 V_RHE, far exceeding of CNT@CMP(CoPc) with 9.9 mA cm^-2. In addition, FE_CO of 96% can be maintained even at high current densities (> 200 mA cm^-2) from -0.6 to -1.1 V_RHE. In the study of catalytic mechanism, it was found that, on the one hand, the H₂Pc unit as electron donor improved the nucleophilicity of the neighboring monatomic Co and enhanced the interaction between the metal and CO₂. On the other hand, the H₂Pc unit can also be used as a proton donor to stabilize the intermediate of CO₂ through intermolecular hydrogen bonding, improve the proton transfer ability of the electrocatalyst, and promoted the activation process of CO₂, thus improving the selective reduction of CO₂ by cooperating with the active metal center.

The electrocatalytic activity of M-N_x sites in POPs may be affected by the carbon substrate, and it is reported that CNTs with highly curved surfaces may provide a unique curvature effect for M-N_x, which in turn improves the electrocatalytic performance^[219]. Ao et al. prepared NiN_x by high temperature pyrolysis of Ni particles in N-doped CNTs, and then coated CTF nanosheets to synthesize electrocatalyst Ni-N/CNT@CTF^[220]. The close contact between ultra-thin 2D CTFs and CNTs could inhibit the accumulation of carbon layers and give full play to the curvature effect of CNTs. The Ni-N bond length (1.90 Å) in the Ni-N/CNT@CTF was slightly larger than that of the common NiN_x site (1.84~1.88 Å) due to the compressive strain of the curved CNTs on the Ni-N sites. DFT data showed that NiN₃ site was active site for electrocatalytic CO₂RR, and its electronic structure could be adjusted by the curvature effect of CNTs, thereby improving the kinetics of electrocatalytic CO₂RR. The current density and FE of CO were 201 mA cm^-2 and 98% at -0.9 V vs. RHE for Ni-N/CNT@CTF in flow cells.

Metal/POP composites

Metal NPs suffer from problems such as easy aggregation, low mass and electron transfer, and poor stability in the reaction process, which significantly reduce their catalytic performance^[221]. Metal/polymer hybridization is a feasible way to solve the above problems. The modification of metal NPs by POPs with large SSA and excellent stability is not only beneficial to the high dispersion of metal particles, thereby improving the accessibility of active sites, but also can improve the stability of metal catalysts in the catalytic process. The introduction of POPs can also regulate the catalytic microenvironment of metal sites. Therefore, it is highly desirable to design metal/polymers with higher mass and electron transfer efficiency through different strategies. Polymer ligands can control the surface accessibility of metal nanocatalysts, which provides a way to improve their cycle life and catalytic performance. Zhang et al. used Au and palladium (Pd) nanocatalysts stabilized by two polymer NHCs (mono- and polydentate) for CO₂RR^[222]. Compared to traditional thiols, amines, and other ligands, the stable metal-carbene bond at the reduction potential prevented the aggregation of metal NPs. The NHC-ligand-modified Au catalyst had 86% CO FE at -0.9 V for 11 h, whereas that of the unmodified Au was < 10%. Moreover, the hydrophobicity of polymer ligands and the surface electron density of NPs greatly improved the selectivity of C via the σ-donation of NHCs.

Cu-based catalysts can be used in the electrocatalytic reduction of CO₂ to produce CH₄, etc.^[223]. However, due to complicated reaction process and poor structural stability, the selectivity and stability of CH₄ are poor, which limits its practical application^[224]. By adjusting the catalytic microenvironment of Cu sites, including the coverage of reaction intermediates^[225], the adsorption of CO₂, H₂O^[226], etc., the path of CO₂RR can be regulated to improve the selectivity of CH₄ products. The sheet-like structure and open channels of 2D COFs provide an ideal platform for regulating the catalytic microenvironment^[227]. Cu NPs were modified by 2D COF nanosheets (NUS9) with high density of heteroatoms and sulfonic acid groups to prepare Cu-NUS9, which can adjust the selectivity and activity of its products. As a result, the Cu-NUS9 catalyst reduced CO₂ to CH₄ with 66% FE and 296 mA cm^-2 of current density in an acidic electrolyte (pH = 2) and 61.6% FE and 308 mA cm^-2 of current density in an alkaline electrolyte (pH = 14). The introduction of NUS9 with a porous structure and ionic group could promote the local accumulation of CO₂ and K⁺ around the surface of the catalyst, which significantly reduced adsorption energy of *CO on Cu-NUS9 (-0.22 eV) compared with Cu (-0.27 eV) and prevented the occurrence of C-C coupling reaction. On the other hand, the enhanced adsorption of H₂O significantly promoted the dissociation of H₂O which was beneficial to the hydrogenation reaction in the CO₂RR process. Additionally, the presence of NUS9 significantly lowered energy barrier of *CO + *H→*CHO, which was RDS for CH₄ production (Cu-NUS9: 1.02 eV, Cu: 1.35 eV), and promoted the reduction of CO₂ to CH₄.

In addition, Meng et al. proposed an effective tandem catalytic strategy to improve the selectivity of reducing CO₂ to C₂H₄^[228]. Combined with the advantages of non-noble metal monatomic Ni to produce CO and Cu nanocatalyst to conduct CO-CO coupling, the selectivity of CO₂ to ethylene was improved by tandem catalysis. Based on the above considerations, non-noble PTF (Ni)/Cu was prepared by loading Cu NPs on Ni monatomic porphyrin-triazine framework (PTF-Ni) [Figure 15A and B]. In the catalytic process, monatomic Ni efficiently reduced CO₂ to intermediate CO, and the generated CO was immediately converted to ethylene by the adjacent Cu nanocatalyst through C-C coupling reaction. As a result, compared with non-tandem catalyst PTF/Cu which mainly produced CH₄, FE of C₂H₄ increased from 9.6% to 57.3% at -1.1 V vs. RHE [Figure 15C]. Furthermore, PTF (Ni)/Cu could still maintain ~91% of the initial activity after 11 h [Figure 15D]. PTF (Ni) was beneficial to increasing the generation of active intermediates of *CO on Cu NPs, thereby improving the probability of C-C coupling, and significantly reducing energy barrier required for the production of C₂H₄, thus improving activity of converting CO₂ to C₂H₄ through tandem catalysis.

Figure 15. (A) Fabrication of PTF(Ni)/Cu. (B) HRTEM images of PTF(Ni)/Cu. (C) FEs of C₂H₄ and CH₄ at different potential. (D) The long-term stability for 11 h on PTF(Ni)/Cu catalyst^[228]. Copyright 2021 Wiley-VCH.

Molecular catalyst/POP composites

The porous structure of POPs allows the insertion of molecular electrocatalysts via rapid diffusion to obtain sufficient active sites, thereby improving the electrocatalytic activity and stability^[229]. To verify these characteristics, rhenium (Re)-based Re (2,2′-bpy-5,5′-diamine) (CO)₃Cl was inserted into the framework of COF-2,2′-bp with non-metallic 2,2′-bpy moiety to segregate COF-2,2′-bpy-Re^[230]. COF-2,2′-bpy-Re combined with carbon black and polyvinylidene fluoride (PVDF) actively catalyzes the reduction of CO₂ to CO at -2.8 V only for a Re concentration of 29.38 wt%. After 30 m of controlled potential electrolysis (CPE), the CO FE for the catalyst reaches 81%. After 30 m, the electrocatalytic performance of the composite decreases due to the inhibition of mass transfer and substrate diffusion. Moreover, Co porphyrin molecules have been embedded into cationic POPs with bipyridyl ligands (POP-Py) by anion exchange to generate strong secondary coordination sphere interactions to improve CO₂RR performance^[231]. In this study, Co meso-tetra (4-carboxyphenyl) porphyrine (Co-TCPP) was uniformly encapsulated in POP-Py under electrostatic attraction. The hybrid exhibited a CO FE of 83%, current density of 4.0 mA cm^-2, and TOF of 1.4 s^-1 at 0.49 V of overpotential, which was attributed to the synergistic interaction between the cationic polymer substrate and the molecular catalyst. The electrostatic effect of the positively charged framework of the cationic POPs could not only promote the dispersion and stable loading of the negatively charged molecular catalyst, but also stabilize the intermediate of the electrocatalytic reaction and improve the catalytic activity.

The strategy of integrating the photosensitive donor Ru(bpy)₃Cl₂ and the acceptor CO₂ reduction activator Co-Por into the crystalline COF framework can effectively prolong excited state lifetime of Co-Por and improve the catalytic performance of electrocatalytic CO₂ reduction under illumination. Wu et al. constructed Ru (bpy)₃Cl₂ photosensitizer into 2,2′-bipyridine-functionalized Co porphyrin-based COF (Co-Bpy-COF) by post-synthesis method, and obtained Co-Bpy-COF-Ru_x(x represented molar ratio of Ru to Co)^[232]. Due to the large polarity difference of building units and the high crystallinity of 2D COF, the built-in electric field in Co-Bpy-COF-Ru_1/2 was created to facilitate the migration of photoexcited electrons from Ru (bpy)₃Cl₂ to Co-Por on the one hand, and on the other hand, the excited state lifetime of Co-Por was 3.4 times longer than that of Co-Bpy-COF, and therefore the formation energy of CO was effectively reduced under the illumination. Thus, Co-BPy-COF-Ru_1/2 had the highest FE_CO of 96.7% at -0.7 V_RHE, and CO partial current density of 16.3 mA cm^-2 at -1.1 V_RHE, both far exceeding the value with no illumination (84.6%, 11.9 mA cm^-2).

MOF/POP hybrids

COFs and MOFs, as two kinds of porous crystalline materials, have received extensive attention and research in electrocatalytic reduction of CO₂. Despite many similarities between them, there are still obvious differences in structure and functionality, and each has its own shortcomings. It is possible to combine their own advantages and form synergistic catalytic effect in electrocatalysis to overcome their respective shortcomings. In recent years, there have been many reports on MOF and COF hybrid materials, which have opened up a wide range of frontier applications for the synthesis and application of these materials^[33]. However, MOFs/COFs reported in the literature have some problems, such as less exposure of active sites and insufficient research on interface effect or synergistic mechanism. Therefore, it is of great value and significance to explore MOF@COF hybrid materials with fully exposed active sites to realize electrocatalytic CO₂RR and produce high-value-added products.

Based on the above considerations, Yang et al. prepared a series of composite electrocatalysts MCH-X by epitaxially growing an ultrathin Cu-Por-based COF (COF-366-OH-Cu) layer with nanoscale thickness on the surface of a mesoporous honeycomb UiO-66-NH₂ template via covalent bonds (imine linkage)^[233]. The morphology of MCH-X and the exposure of active sites were optimized by adjusting the amount of MOFs, which endowed the composite with high electrocatalytic performance. The optimum MCH-3 was still able to maintain the honeycomb structure due to the thinner COF layer with a thickness of ~8.5 nm, which was favorable for the exposure of the active sites. The CO₂ adsorption capacity of MCH-3 was higher than that of COF-366-OH-Cu but lower than that of UiO-66-NH₂, suggesting that the introduction of UiO-66-NH₂ improved the affinity of the hybrid material for CO₂. The results showed that UiO-66-NH₂, COF-366-OH-Cu and MCH-3 could reduce CO₂ to CH₄ and a very small amount of C₂H₄. The maximum FE_CH4 and total current density of MCH-3 was 76.7% and -398.1 mA cm^-2 at -1.0 V, far exceeding that of UiO-66-NH₂ (15.9%, -187.8 mA cm^-2) and COF-366-OH-Cu (37.5%, -374.4 mA cm^-2), demonstrating the advantage of the honeycomb heterostructure possessed by MCH-3. Due to the introduction of MOFs, the formation energy barrier of the *COOH, acting as RDS for the CH₄ generation for MCH-3, was 1.18 eV, lower than that of COF-366-OH-Cu (1.41 eV), further promoting the formation of the intermediates *CO, *CHO, and *CH₂O. This work provides a reference for applications of porous crystalline hybrid materials in the efficient electrocatalytic reduction of CO₂.