Applications of Metal–Organic Frameworks and Their Derivatives in Electrochemical CO₂ Reduction

Highlights

• The electrochemical techniques utilizing metal-organic frameworks (MOFs)-based catalysts for converting CO₂ into chemical species are discussed.

• The structure–activity relationship of MOF-based catalysts in electrocatalytic CO₂ reduction reactions is thoroughly reviewed

• The challenges and opportunities of large-scale applications of MOF-based materials in electrochemical CO₂ reduction reactions are discussed, and possible directions for the future development of MOFs and their derivatives are outlined.

Abstract

Electrochemically reducing CO₂ to more reduced chemical species is a promising way that not only enables the conversion of intermittent energy resources to stable fuels, but also helps to build a closed-loop anthropogenic carbon cycle. Among various electrocatalysts for electrochemical CO₂ reduction, multifunctional metal–organic frameworks (MOFs) have been employed as highly efficient and selective heterogeneous electrocatalysts due to their ultrahigh porosity and topologically diverse structures. Up to now, great progress has been achieved in the design and synthesis of highly active and selective MOF-related catalysts for electrochemical CO₂ reduction reaction (CO₂RR), and their corresponding reaction mechanisms have been thoroughly studied. In this review, we summarize the recent progress of applying MOFs and their derivatives in CO₂RR, with a focus on the design strategies for electrocatalysts and electrolyzers. We first discussed the reaction mechanisms for different CO₂RR products and introduced the commonly applied electrolyzer configurations in the current CO₂RR system. Then, an overview of several categories of products (CO, HCOOH, CH₄, CH₃OH, and multi-carbon chemicals) generated from MOFs or their derivatives via CO₂RR was discussed. Finally, we offer some insights and perspectives for the future development of MOFs and their derivatives in electrochemical CO₂ reduction. We aim to provide new insights into this field and further guide future research for large-scale applications.

1 Introduction

The overuse of fossil fuels has caused a rapid increase in atmospheric CO₂ concentrations and disrupted the natural equilibrium of the carbon cycle, leading the global warming and subsequent consequences such as frequent storms, drought, and rising sea levels [1]. As climate change is getting worse, it is urgent to protect the ecological environment by developing advanced technologies to close the carbon loop. One potential approach is shifting current industries’ energy dependence from fossil fuels to renewable sources such as solar, wind, and thermal [2, 3]. However, unlike the controllable base-load energy in modern power grids, the inherent intermittency of renewable energy sources greatly limits their use or efficiency [4]. To solve the energy fluctuation problem, storing renewable energy in the chemicals form by converting CO₂ into chemical feedstocks and fuels is a more promising way, which not only efficiently utilizes renewable energy but also reduces carbon emissions [5, 6].

Electrocatalytic reduction of CO₂ to value-added chemicals using renewable electricity provides a practical solution to offset the extra carbon footprint [7,8,9,10,11,12]. However, the high activation barrier to break the symmetrical linear structure and stable C = O bonding (750 kJ mol⁻¹) make the electrochemical process difficult [13,14,15,16,17]. Furthermore, the CO₂ reduction reaction (CO₂RR) process involves multi-electron/proton transfer processes and generates many different reaction intermediates, which are associated with a number of different reactions pathways and result in a large variety of products [18,19,20,21,22,23,24]. In recent years, efforts have been extensively dedicated to the design of catalysts with high efficiency and selectivity, resulting in noteworthy progress. Metal–organic frameworks (MOFs) are a class of porous coordination polymers consisting of metal ions/clusters coordinated to organic ligands, the open frameworks have high permanent porosity, high crystallinity, and long-range order [25,26,27,28,29,30]. The high surface area, tunable porosity, diversity in inorganic units and organic linkers make them attractive in catalytic reactions especially in the field of electrochemical CO₂RR [31,32,33,34,35,36,37]: (1) The remarkably high porosity and surface area of MOFs facilitate the exposure of more active sites and boost rapid mass transport. (2) The electronic structure of MOFs can be precisely modulated via metal ion doping or functionalization of organic ligands, which influences the reaction pathways for CO₂ electroreduction. (3) The varied organic linkers and abundant metal ions/clusters make MOFs attractive precursors or templates to fabricate numerous advanced MOF derivatives [such as metal/metal oxide nanoparticles, carbon materials and single-atom catalysts materials (SACs)] [38,39,40,41,42,43,44,45,46,47] as electrocatalysts for selective CO₂RR.

There are various methods used for the synthesis of MOFs, including solvothermal, hydrothermal, microwave, and mechanochemical methods. The most common method is solvothermal synthesis, which involves the reaction of metal salts and organic ligands in a solvent at high temperature and pressure. The solvothermal method allows for precise control of the size, shape, and properties of the resulting MOFs. Hydrothermal synthesis is similar to solvothermal synthesis, but the reaction takes place in an aqueous solution instead of an organic solvent. This method is advantageous for the synthesis of water-stable MOFs. Microwave synthesis involves the use of microwave radiation to heat the reaction mixture and promote the formation of MOFs. This method is rapid and efficient and allows for the synthesis of MOFs in minutes rather than hours. Mechanochemical synthesis involves the use of mechanical force to induce chemical reactions between metal ions and ligands. This method is advantageous for the synthesis of MOFs with high thermal stability and is more environmentally friendly than traditional synthesis methods. Overall, the method chosen for the synthesis of MOFs for CO₂RR involves a careful balance between catalytic activity, stability, and selectivity, with consideration given to the properties of the metal, ligand, and pore structure.

Primitive MOFs usually have drawbacks including low electrical conductivity, instability and inactive metal nodes due to the blockage of metal centers by organic ligands, which result in poor CO₂RR performance. These limitations can be addressed by converting unstable MOFs into MOF derivatives, which can maintain the highly porous structure of the original MOF while also providing improved electrical conductivity and stability. MOF derivatives can be produced through electrochemical/chemistry reduction, surface modification, pyrolysis under specific conditions, and so on, which allow for precise control of material morphology, composition, surface area, and electronic structure of the metal nodes. For example, by controlling decomposition of MOFs under an inert atmosphere at high temperatures or under specific conditions can produce MOF-derived carbon materials or metal/metal oxide nanoparticles, respectively. Notably, MOF precursors with high thermal stability and metal loading can be thermally decomposed under an inert atmosphere to form SACs where isolated metal atoms are embedded in a carbon matrix, greatly enhancing the utilization efficiency of metal atoms. Overall, MOF derivatives largely expands the family of MOF materials and deliver superior performance to the pristine MOFs.

In the past years, MOF-related catalysts in electrochemical CO₂RR systems have been thoroughly investigated, it is timely for us to take a systematic review to summarize the recent advances and pertinent challenges in this field (Fig. 1). The review starts with an introduction of the CO₂RR reaction mechanisms at the molecular level, as well as a brief summary the electrolyzer structures. Then we discuss how to improve the selectivity and activity in electrochemical CO₂RR toward different products from the perspective of materials design strategies such as pore structure modification, central metal atom substitution, and coordination environment adjustments. Finally, this review concludes with some of our insights about the research challenges and future directions, hoping to stimulate continuous innovations for advancing MOF-derived functional materials for electrochemical CO₂RR.

Fig. 1

Schematic of MOFs and their derivatives for CO₂RR

2 Mechanism and Electrolyzer Types of Electrocatalytic reduction of CO₂

2.1 Mechanism of Electrocatalytic Reduction of CO₂

CO₂ reduction was first reported by Royer in 1870 when they observed the formation of formic acid in an aqueous medium [48]. Since then, the mechanism of electrochemical CO₂RR has been gradually discovered and a number of different possible reaction intermediates are identified, which essentially determine the reaction routes and final products [49, 50]. According to the number of carbon atoms, these products are typically divided into C₁, C₂, and C₃ molecules, such as carbon monoxide (CO), formic acid (HCOOH), methanol (CH₃OH), methane (CH₄), ethylene (C₂H₄), acetic acid (CH₃COOH), ethanol (CH₃CH₂OH), n-propanol (CH₃CH₂CH₂OH), and acetone (CH₃COCH₃). Table 1 shows the corresponding half-reactions for different products as well as their standard reduction potentials versus reversible hydrogen electrodes (RHE) in both acidic and basic conditions [51, 52]. The large variety of possible CO₂RR reaction pathways and their similar reduction potentials make the selective reduction to specific products a great challenge. Furthermore, in an aqueous electrolyte, the competition of hydrogen evolution reaction (HER) is a thorny issue that needs to be addressed as well.

Table 1 The half-reactions of electrochemical CO₂RR as well as their standard reduction potentials versus reversible hydrogen electrode (RHE)

2.1.1 C₁ Pathways

In general, the first step of electrochemical CO₂RR is to absorb CO₂ molecules onto the catalyst surface and form *CO₂⁻, which can further accept protons and/or electrons to form various intermediates that determine the final products [49, 53]. Figure 2 shows the possible mechanistic pathways of electrochemical CO₂ reduction to C₁ and C₂ products, respectively. For C₁ products especially two-electron products, such as CO and HCOOH (or HCOO⁻), the rate-limiting step is usually identified to be the formation process of *CO₂⁻ [49, 52, 54]. The reaction pathway toward CO or HCOO⁻ is typically determined by the absorption configurations of the intermediate on the catalyst surface [55, 56]. Typically, when *CO₂⁻ binds to the catalyst surface through two oxygen atoms and forms the *OCHO intermediate, then HCOOH is preferably formed by a proton-coupled electron transfer step. The reaction pathway toward CO is similar, a number of studies, which combines experimental and theoretical evidence, found that *COOH is a key descriptor for CO production [49, 52, 55, 57,58,59,60]. A carbon-bound *COOH intermediate is generated through the initial binding of CO₂ to the surface of the catalyst. Subsequently, the concerted proton-electron transfer (CPET) steps are readily triggered by attacking the oxygen atoms and forming H₂O, after dehydration, *CO intermediate can be easily desorbed as gaseous CO molecules when the binding energy between the catalyst and the intermediate is relatively low. In contrast, if the binding energy is strong, *CO can get further protonated to two intermediates *CHO or *COH. Then, the *CHO intermediate may be subsequently reduced via the protonation of its carbon atom, resulting in the formation of *CH₂O (desorb as HCOH) and *CH₃O. And *CH₃O can be further reduced via the protonation of its carbon atom or oxygen atom to get CH₄ or CH₃OH, respectively. On the other hand, *COH also can be further reduced to *C and then hydrogenation to CH₄ through the CPET steps [49, 52, 57].

Fig. 2

Reproduced with permission from Ref. [48]

Possible mechanistic pathways of CO₂ reduction to C₁ and C₂ products on catalysts.

2.1.2 C₂ Pathways

*CO is a crucial intermediate for electrochemical CO₂RR to get C₂ product, and the reaction pathways of *CO intermediates via multiple CPET or other steps determine the final C₂ products. Cu-based materials are the most efficient catalysts explored that show appreciable selectivity and faradaic efficiency for C₂ products. On Cu(100) surface, *CO dimerization occurs prior to protonation at low overpotential [61], and is the rate-determining step for the formation of C₂H₄, CH₃CH₂OH, and CH₃CH₂CH₂OH [13]. Besides the dimerization of *CO, the C–C bond could be formed by the coupling of other further protonated species. For example, on Cu(100) at high overpotential or on Cu(111) surface, *CO is protonated to *CHO before coupling with CO to form *COCHO, which is a key intermediate for getting C₂ products [61, 62]. Another example is that *CH₂ intermediate could be generated by the formation of C-H bonds from PECT steps, and it is a pivotal intermediate in the generation of C₂H₄ and CH₃COOH [13, 63]. *CH₂ can also be further protonated to *CH₃ intermediate, which leads to C₂H₆ product via *CH₃ dimerization. In addition, the reaction pathway for the formation of both C₂H₄ and CH₃CH₂OH productions involves CO insertion into *CH₂, which is more commonly known as a Fischer–Tropsch-like step [64, 65]. Due to the abundance of intermediates and their protonation possibilities, there are many reaction pathways that form different C₂ products, leading to uncontrollable product generation. Significantly, both experiments and density functional theory (DFT) calculations suggest that the rate-determining step in C–C coupling involves a decoupled proton-electron transfer, while whether the C–C coupling is an electrochemical step or a chemical step is still up for debate [66,67,68,69].

2.2 Electrolyzer Types

In the research of electrochemical CO₂RR, there are mainly three types of electrolyzers: H-type cells, flow cells, and membrane electrode assembly (MEA) cells. Figure 3a shows a simplified configuration of a traditional H-type cell. The cell is composed of independent cathode and anode chambers, with the characteristics of easy operation, facile assembly, and low cost [13, 70, 71]. The two chambers are separated by an ion-exchange membrane, which allows the flow of ions while preventing the oxidation of the CO₂ products by limiting their transport from cathode to anode. Catalysts are usually deposited or coated on conductive but inactive substrates (glassy carbon or carbon paper (CP)) and serve as working electrodes. In an H-type cell, CO₂ molecules are commonly bubbled from the bottom of the aqueous electrolyte and saturate the electrolyte, then transferred to the interface of the working electrode where CO₂RR takes place. However, the finite CO₂ solubility in aqueous electrolyte (only 0.034 M under ambient conditions) results in limited CO₂ reduction current densities (less than 100 mA cm⁻²) [57, 70]. Furthermore, the thick diffusion layer (> 50 μm) leads to poor mass transport between the catalyst surface and bulk electrolyte, resulting in a slow reaction rate [70, 72, 73]. Although H-type cells are hindered by such limitations for practical applications, it provides valuable information for evaluating the intrinsic catalytic performance of the catalysts.

Fig. 3

Schematic diagrams of the a H-types cell, b structure of GDEs, c flow cell, and d MEA cell

To meet the industrial utilization of the electrochemical CO₂RR, flow cells were developed so that CO₂ can be efficiently delivered to the cathode continuously. Before introducing flow cells in more detail, it is necessary to learn about the structure and function of gas diffusion electrodes (GDEs). Figure 3b is a schematic of GDEs, which consists of a porous catalyst layer (CL) and a diffusion medium (DM) [74, 75]. The DM typically serves as the gas-permeable and electron-conductive substrate on which the CL is deposited. The substrate not only plays a role in determining the local electronic environment of the catalysts, but also influences the mass transfer of the reactants and products to and from the CL. Most DM consists of two parts, the gas diffusion layer (GDL) and the microporous layer (MPL). The GDL acts as a porous medium, which permits diffusion of both gaseous CO₂ to the CL and gaseous products (H₂, CO, CH₄, C₂H₄) away from the CL surface [76, 77]. The unique structure of GDEs forms a gas–liquid-solid three-phase interface where the electrochemical CO₂RR occurs without solubility limitations [78, 79]. To stabilize the triple-phase interface, MPL composed of carbon black nanoparticles is commonly treated with hydrophobic polytetrafluoroethylene to prevent flooding of electrolytes into GDEs, resulting in efficient mass transport for gaseous CO₂.

In flow cells, GDEs are often used for better control of the three-phase interface where the gaseous CO₂ feedstock can be directly reduced without mass diffusion limitation. Typical membranes in flow cells include the cation exchange membrane (CEM), anion exchange membrane (AEM), or a bipolar membrane (BPM), the type of membrane affects the applicable electrolyte conditions and the ion transport kinetics. As shown in Fig. 3c, a polymer electrolyte membrane is sandwiched between two electrolyte channels, dividing the cell into two parts: the anodic and cathodic compartments. Both cathode and anode electrolytes are continuously flowed through electrolyte channels by a pump. The anode side typically carries out a complementary oxidation reaction, which is most commonly water oxidation, typically using IrO_x nanoparticles on Ti mesh/GDEs [80,81,82]. In the cathode chamber, a GDE sits at the interface of the inbound CO₂ and the flowing aqueous electrolyte, which largely improves the mass transport of gas–liquid interface, making it possible to conduct electrochemical CO₂RR at industrial-scale current densities [70, 83, 84].

The MEA cell is another emerging class of electrolyzers that has been commonly applied to electrochemical CO₂RR. As shown in Fig. 3d, by removing the flowing electrolyte channel between the GDE and membrane, it directly combines GDEs and ion exchange membrane into one unit. Benefiting from the uniqueness of this configuration, the MEA cell could significantly decrease the distance between the cathode and anode and thus boot the mass/electron transfer, resulting in high energy efficiency [85,86,87,88]. In addition, the removal of the flowing liquid electrolyte could relieve GDE flooding and reduce contamination of the cathode catalyst from impurities in electrolyte [89, 90].

3 Using MOFs and their Derivatives as Catalysts for CO₂ Electroreduction to Value-Added Chemicals

MOFs are porous crystalline materials synthesized from the coordination bonds of metal ions/clusters and organic ligands, and they possess open frameworks with tunable porous properties. In 2012, a copper rubeanate MOF was first reported as a catalyst for electrochemical CO₂RR and was shown to effectively convert CO₂ into HCOOH in 0.5 M KHCO₃ solution, with an HCOOH selectivity above 98% [91]. Subsequently, Senthil Kumar et al. reported the generation of oxalic acid from CO₂RR using Cu₃(BTC)₂ (BTC = 1,3,5-ben-zenetricarboxylicacid) catalyst in N,N-dimethylformamide, and a high Faradic efficiency (FE) of 51% [92] was achieved. Since then, MOF-related catalysts have received significant attention in electrochemical CO₂RR and become a rapidly growing research area. Some of these materials have demonstrated obvious activity, selectivity, and stability for electrochemical CO₂RR in various electrolytes. C₁ chemicals, such as CO, HCOOH, CH₄, and CH₃OH, are generally the main products in current MOF-catalyzed CO₂RR systems, whereas the desirable C₂ products are relatively rare. Thus, this paper mainly focuses on the C₁ products produced from MOF-related materials and discusses strategies to enhance their electrocatalytic performance, while state-of-art MOF-related materials for catalyzing CO₂ to C₂₊ are also briefly summarized.

3.1 Carbon Monoxide

CO is a crucial building block for the large-scale production of commodity and specialty chemicals [93]. Among all the products from electrochemical CO₂RR, CO is one of the most economically viable products and has a high ratio of molecular weight per electron [94, 95]. As mentioned above, an ideal catalyst that selectively catalyzes the electrochemical reduction of CO₂ into CO should possess not only strong adsorption energies of *COOH intermediate but also weak adsorption energies of *CO. However, the binding energies of *COOH and *CO are generally proportionally related and follow the scaling relations, it is hard to alter the reaction pathways individually. Fortunately, the linear relationship can be broken by regulating the intrinsic physical/chemical properties and electrochemical microenvironments of the catalytic materials. Owing to their tunable chemical properties, MOF-related materials are perfect candidates for CO₂RR because of their tunable structure and compositions. MOFs also have the advantage of well-defined single-atom sites; this is helpful to elucidate the surface dynamic changes and chemical adsorption of reaction intermediates. So far, a number of MOF-related materials have been extensively explored for the reduction of CO₂ to CO (Table 2), we will discuss their structure–activity relationships from the perspective of morphology, conductivity, and the coordination environment of metal center.

Table 2 Representative MOF-related catalysts for the electrochemical reduction of CO₂ to CO

3.1.1 Pristine MOFs

Morphology and size are two important characteristics of metal–organic framework (MOF) materials that influence their properties and performance in various applications. In the case of MOFs, the morphology can vary from a crystalline powder to a dense film or even an ordered nanostructured material. The morphology of MOFs can be influenced by several factors such as the synthesis method, the precursor choice, and the reaction conditions. On the other hand, the dimensions of the MOF particles or the size of the pores within the MOF structure determines their adsorption properties. Controlling the morphology and size of MOF material not only maximizes the active sites but also balances both charge and mass transport, resulting in high catalytic activity. For example, Kornienko et al. [96] fabricated a porous thin film of Al₂(OH)₂TCPP-Co(TCPP-H₂ = 4,4′,4″,4‴-(porphyrin-5,10,15,20-tetrayl)tetrabenzoate) MOF and employed it as an electrocatalyst for the efficient and selective reduction of CO₂ to CO in aqueous electrolytes, as early as 2015. By optimizing the catalytic linker unit (Fig. 4a) and assembling a porous thin film MOF (Fig. 4b) on a conductive substrate, a three-dimensional (3D) porous working electrode was successfully prepared. As shown in Fig. 4c, when tested in CO₂-saturated KHCO₃ electrolyte for continuous electrolysis, Al₂(OH)₂TCPP-Co exhibited good catalytic activity and stability, with a maximum CO FE of over 76% and good stability over 7 h at −0.7 V vs. RHE. The cobalt centers are electrically linked to the electrode and are reduced to the catalytically active Co(I) state (Fig. 4d–f). And the rate-limiting step might be either a CO₂ molecule adsorbing onto a Co(I) porphyrin coupled with a one-electron reduction or a one-electron reduction of a Co(I)-CO₂ adduct. Thanks to the nanoscopic MOF morphology and thickness, the number of active sites is maximized and both charge and mass transported could be simultaneously balanced, thus improving the electrocatalytic performance. Moreover, Dong et al. [97] also use the same ligand TCPP as a building block to synthesize a 3D highly stable porphyrin-based MOF of PCN-222(Fe) (Zr₆O₈(OH)₄(H₂O)₄][(TCPP-Fe-(III)-Cl)₂). As shown in Fig. 4g, it consists of Zr₆ clusters and square planar Fe-TCPP ligands, resulting in a star-shaped network. Thanks to the micromorphology, PCN-222(Fe) extensively exposes the porphyrin active sites, which have a substantial catalytic effect on the electrochemical reduction of CO₂. Furthermore, the permanent porosity of PCN-222(Fe) conferred a promising CO₂ adsorption ability, which is also advantageous in CO₂ electrochemical reduction since the reduction kinetics are closely related to the CO₂ concentration. When tested in CO₂-saturated 0.5 M KHCO₃ aqueous solution at various steady-state potentials from -0.45 to -0.85 V vs. RHE, PCN-222(Fe)/C (PCN-222(Fe) mixed with carbon black) showed excellent catalytic selectivity and activity toward CO. As shown in Fig. 4h, the product distribution in the gas phase was dependent on the applied potentials, and the maximum FE of CO occurs at -0.60 V vs. RHE (91%). In the low overpotential region, the Tafel slope of PCN-222(Fe)/C is 188 mV dec⁻¹, which indicates that a one-electron reduction of CO₂ to CO₂⁻ radical is a probable rate-limiting step. PCN-222(Fe)/C also showed high structural stability after continuous electrolysis for 10 h, however, the low CO₂-to-CO current density (< 10 mA cm⁻²) limited its large-scale industrial applications.

Fig. 4

Reproduced with permission from Ref. [96]. g 3D crystal structure of PCN-222(Fe). h Steady-state current density and the selectivity for each gas product in a potential range from -0.45 to -0.85 V vs. RHE. Reproduced with permission from Ref. [97]

a MOF catalyst allows for the modulation of metal centers, molecular linkers, and functional groups at the molecular level. b Organic building units, in the form of cobalt-metallated TCPP, are assembled into a 3D MOF, Al₂(OH)₂TCPP-Co with variable inorganic building blocks. c Stability of the MOF catalyst is evaluated through chronoamperometric measurements in combination with faradaic efficiency measurements. d In situ spectroelectrochemical analysis reveals the oxidation state of the cobalt catalytic unit of the MOF under reaction conditions. Upon varying the voltage from 0.2 to -0.7 V vs RHE, the Co(II) Soret band decreases at 422 nm and is accompanied by a C5o(I) Soret band increase at 408 nm. This change is quantified and plotted e to elucidate a formal redox potential of the Co center, which is deemed to be at the peak of the first derivative f of the Co(II) bleach and Co(I) enhancement.

Choosing electron-donating ligands as well as their orientation and bonding arrangements is an efficient strategy to improve the conductivity of MOFs, these functional groups can increase the conductivity of MOFs by creating pathways for the flow of electrons and increasing the density of free electrons in the material [114,115,116]. Owing to the high overlap of d-π conjugation orbitals between the nickel node and the planar Ni-phthalocyanine-substituted X (X: o-phenylenediamine or catechol), Zhang et al. [98] employed Ni-phthalocyanines (NiPc) as the building block for the construction of a porous intrinsically conductive two-dimensional (2D) MOF (NiPc-Ni(NH)₄). The 2D NiPc-Ni(NH)₄ MOF showed a high electrical conductivity of 2.39 × 10^–4 S m⁻¹, and NiPc-Ni(NH)₄ nanosheets showed outstanding CO₂-to-CO electrocatalytic performance with a high CO selectivity of 96.4% at −0.7 V vs. RHE in CO₂-saturated 0.5 M KHCO₃ electrolyte. DFT calculations revealed that the active site is Ni-N₄ moiety in NiPc. The presence of square planar Ni(NH)₄ nodes can efficiently accelerate the proton/electron transport to the active sites, thus accelerating the reaction kinetics during the electrochemical CO₂RR. Following the same strategy, Yi et al. [99] employed NiPc as the building block to prepare the phthalocyanine-based MOF (Ni–Pc–NiO₄) via the solvothermal synthesis method (Fig. 5a), and then exfoliated the bulk powder into 2D nanosheets through high-frequency sonication at room temperature. A two-contact probe method was applied to test the electrical conductivity of NiPc-NiO nanosheets at room temperature. The 2D NiPc-NiO nanosheets showed good electrical conductivity (4.8 × 10^–5 S m⁻¹). Such good electrical conductivity would be beneficial for the electron transfer to the active sites during CO₂RR, thereby improving electrochemical activity and energy conversion efficiency. As expected, when tested in CO₂-saturated 0.5 M KHCO₃ electrolyte, NiPc-NiO nanosheets showed a high CO FE of > 90% in a wide potential range from -0.65 to -1.1 V vs. RHE, reaching the maximum of 98.4% at -0.85 V vs. RHE, surpassing NiPc-OH monomer (Fig. 5b). X-ray absorption spectroscopy (XAS) and X-ray photoelectron spectroscopy (XPS) analysis confirmed that the NiPc sites and NiNO₄ nodes in NiPc-NiO₄ were well maintained after the CO₂RR, demonstrating the superior structural stability of NiPc-NiO₄. Theoretical calculations show that the Gibbs free energy of the rate-determining step (RDS) (formation of *COOH intermediate) on NiPC (1.93 eV) is significantly lower than that on the NiO₄ node (2.53 eV), revealing that NiPc is the active site for the electrochemical conversion of CO₂ to CO (Fig. 5c). Compared to NiO₄ node, NiPc showed stronger Van der Waals interaction with CO₂ molecules (Fig. 5d) in the more electron-rich environment (Fig. 5e). It is worth noting that the lowest unoccupied molecular orbital (LUMO) energy level of NiPc-NiO₄ shifts from −4.22 to −4.62 eV when the CO₂ molecule moves from NiPc to NiO₄ node (Fig. 5f), revealing the excellent reducibility of NiPc. In addition, another 2D conductive MOF was reported by Majidi et al. [100]. They use a catechol-based linker, tetrahydroxyquinone (THQ), to synthesize the Cu-based 2D conductive MOF (Cu-THQ) nanoflakes, and the product shows good electrical conductivity of 1.5 × 10^–7 S cm⁻¹. The presence of the THQ linker in Cu-THQ nanoflakes resulted in a large distance between Cu centers, which not only keeps the Cu center from agglomeration but also ensures the reoxidation of the reduced Cu center during the CO₂RR process. When tested in a hybrid electrolyte (1 M C₅H₁₄ClNO + 1 M KOH), Cu-THQ showed high CO current densities and low overpotential. DFT calculations revealed that the higher the CO coverage, the lower the free energy for CO adsorption on the Cu surface, resulting in a high CO production rate. Zhong et al. [101] also designed layered 2D conductive MOFs (PcCu-O₈-Zn) with bimetallic centers (ZnO₄/CuN₄) to improve electrocatalytic CO₂RR activity (Fig. 5g). They immobilized a mixture of PcCu-O₈-Zn and carbon nanotubes (CNTs) onto a carbon paper substrate as the working electrode and tested its CO₂RR performance in 0.1 M KHCO₃. As shown in Fig. 5h, at −0.7 V vs. RHE, the PcCu-O₈-Zn/CNT can effectively reduce CO₂ to CO with a FE of 88%. Moreover, such performance could be sustained for over 10 h, demonstrating excellent stability. They conducted a series of operando experiments and theoretical calculation to demonstrate that the synergistic effect of Cu and Zn sites in PcCu-O₈-Zn is essential for the high catalytic activity and selectivity toward CO production (Fig. 5i). The ZnO₄ units facilitate CO₂ reduction, while the CuN₄ units promote proton and electron transfer during the reaction process (Fig. 5j).

Fig. 5

Reproduced with permission from Ref. [99]. g Schematic structure of PcCu-O₈-Zn (the dashed rectangular indicates the unit cell). h Faradaic efficiency of CO for PcCu-O₈-Zn/CNT, PcCu-O₈-Cu/CNT, PcZn-O₈-Zn/CNT and PcZn-O₈-Cu/CNT at different potentials. i Operando surface-enhanced infrared absorption (SEIRA) spectro-electrochemical analysis of PcCu-O₈-Zn/CNT in CO₂-saturated 0.1 M KHCO₃. j Schematic HER and CO₂RR reaction process of PcCu-O₈-Zn. Reproduced with permission from Ref. [101]

a Illustration of the preparation steps of NiPc-NiO₄. Top and side view of their structures with 2 × 2 square grids in AA-stacking mode. b Faradaic efficiencies of CO. c Calculated energy diagrams for CO₂-to-CO conversion on two proposed active sites in NiPc-NiO₄. d The non-covalent interaction (NCI) between CO₂ and NiPc-NiO₄ structure. e Mulliken charge of different Ni atoms in NiPc-NiO₄. f Energy level of HOMO and LUMO of different Ni atoms in NiPc-NiO₄ when introducing CO₂.

Ligand engineering can also modulate the coordination environment on metal center, thus boosting the activity of electrochemical CO₂RR. By virtue of 1,10-phenanthroline doping, Dou et al. [102] synthesized a ligand-doped product (ZIF-A-LD, ZIF-A: ZIF-8 was activated to generate open Zn sites) with excellent charge transfer ability. Then, the ZIF-A-LD was mixed with carbon black to prepare a working electrode (ZIF-A-LD/CB). When tested in 0.1 M KHCO₃, ZIF-A-LD/CB exhibited a higher CO FE compared to pristine ZIF-8. DFT calculations revealed that the charge transfer from the dopant phenanthroline molecule (excellent electron-donating ability) to the adjacent sp² C sites in the imidazolate enables stronger electrons movement from the active sites to the antibonding orbitals of CO₂, which facilitates *COOH formation and boost CO₂RR activity. Huang et al. [103] reported a stable MOF (NNU-15), Co(OH)₂(H₂O)₂(Co-TIPP) (TIPP = [5,10,15,20-tetra (4-imidazol-1-yl)phenyl]porphyrin), which contains two OH⁻ coordinated Co ions to mimic the active surface status of the catalysts under alkaline CO₂RR conditions (Fig. 6a). It exhibits an outstanding ability to capture and convert the CO₂ molecule. When tested in CO₂-saturated 0.5 M KHCO₃ electrolyte, NNU-15 shows a high FE of 99.2% for CO at -0.6 V vs. RHE (Fig. 6b) and excellent long-term stability of 110 h (Fig. 6c). The NNU-15-CO₂ intermediate can be detected during the CO₂RR process, demonstrating that the metal catalytic center of MOF can cooperate with OH⁻ to capture and activate the CO₂ molecule to form the HCO₃⁻-chelating active center, and thus getting a remarkable activity for CO₂ conversion to CO. In another study, Al-attas et al. [104] take a ligand engineering strategy to enhance the efficiency of CO₂ conversion to CO for Zn-based MOFs. They prepared two types of Zn-based MOFs with different ligands. As shown in Fig. 6d, one is Zn₂(Tz)₂Ox, which is assembled from Zn ions and 1,2,4-triazolate ligand (Calgary Framework 20, CALF20), and the other is Zn(MeIm)₂ that is assembled from Zn ions and 2-methylimidazolate ligand (zeolitic imidazolate framework-8, ZIF-8). They used a gas-diffusion electrode in a flow electrolyzer (circulated with 1.0 M KOH solution) to assess the activity and selectivity of Zn-based MOFs toward CO₂RR. As shown in Fig. 6e, CALF20 shows higher CO FE than ZIF-8 over the whole potential range and the highest FE for CO production is located at −0.97 V vs. RHE with a partial current density of − 32.8 mA cm⁻². The difference in the Zn oxidation state between CALF20 and ZIF-8 is confirmed by soft X-ray spectra, indicating that the Zn site in CALF20 exhibited a lower electron density due to the electron-withdrawing character of the triazole ligand. DFT calculations revealed that the C₂O₂ bridge promotes stronger charge transfer, and the sp² C sites in azolate are the active sites with low overpotential for CO₂RR. More electrons are produced on the adjacent active sites of the azole ligand and facilitated *COOH formation (Fig. 6f), resulting in the high current density and FE toward CO production.

Fig. 6

Reproduced with permission from Ref. [103]. d Schematic illustration of the crystal structure of ZIF-8 and CALF20, including the chemical structure of polyhedron zinc nodes. e CO Faradaic efficiencies at different applied potentials for CALF20 and ZIF-8 in 1.0 M KOH. f Calculated free energy diagram for electrochemical reduction of CO₂ to CO over CALF20 at U = 0.00 V vs RHE (the dashed lines are simply to guide the eye). Reproduced with permission from Ref. [104]

a 3D frameworks for NNU-15 along the b-axis and 3D frameworks with the open holes in NNU-15. b Faradic efficiencies for CO of TIPP, NNU-16 and NNU-15. c Durability test of NNU-15 at the potential of -0.6 V versus RHE (inset: CO FF at different time).

3.1.2 MOF-Derived Materials

Generally speaking, MOFs are considered to be insulators or poor conductors due to the insulating nature of the organic ligands. The poor conductivity of MOFs is an important limitation for their CO₂ catalysis applications. One simple strategy to improve charge transport is mixing MOFs with conductive carbon, but this can result in the loss of mass activities. A more effective way to overcome charge transport is introducing guest redox molecules into the frameworks. For example, Xin et al. [105] introduced cobaltocene into MOF-545-Co to prepare CoCp₂@MOF-545-Co through a facile chemical vapor deposition method (Fig. 7a). Compared with MOF-545-Co, CoCp₂@MOF-545-Co shows high CO₂-to-CO activity and the maximum FE of CO was dramatically increased from 55% (at −0.8 V vs. RHE) to 97% (at −0.7 V vs. RHE) (Fig. 7b). The excellent CO₂ reduction performance of CoCp₂@MOF-545-Co was demonstrated to originate from the synergy effect of MOF-545-Co and CoCp₂. There are several reasons accounting for this consequence. Firstly, in metallocene electron orbits, the participation of d-orbits enlarges the delocalized π-electron from the cyclopentadienyl ring, thus making metallocene a perfect electron-donating or carrying unit. Secondly, the presence of porphyrin ring ligands in MOF-545-Co are ideal charge transfer carriers, and metal in the center of the porphyrin ring might serve as an efficient electron acceptor and active catalysis center. DFT calculations revealed that the introduction of metallocene can create continuous electron transfer channels in MOFs and the strong binding interaction between metalloporphyrin and reactants during the CO₂RR process can largely reduce the adsorption energy of CO₂. This work shows that the conductivity of MOF materials is a crucial factor that affects the efficiency of electrochemical CO₂RR. Furthermore, it could open a new avenue to develop highly selective CO₂RR electrocatalysts. In another study, Xin et al. [106] followed the same strategy to enhance the conductivity of MOFs materials. They insert the conducting polypyrrole (PPy) molecule into the channel of MOF-545-Co through a facile in-situ low-temperature polymerization method. As shown in Fig. 7c, in the obtained PPy@MOF-545-Co, the PPy molecule acts like a cable in the MOF channel, it can boost the conductivity of MOFs and facilitate electrons transfer to the active center of Co-TCPP. The charge-transfer resistance of PPy@MOF-545-Co (7.5 Ω) was proved to be much lower than that of MOF-545-Co (12.5 Ω), indicating that it possessed a better electron transfer ability. Subsequently, PPy@MOF-545-Co was applied as the working electrode and its electrocatalytic CO₂RR activity was tested using an H-type cell in the CO₂-saturated 0.5 M KHCO₃ electrolyte. As shown in Fig. 7d, PPy@MOF-545-Co exhibits excellent catalytic activity toward CO with a maximum FE up to 98%, which is much higher than its MOF-545-Co counterpart. The CO partial current density was also remarkably higher than the samples in the controlled experiments. Furthermore, after continuous electrolysis at -0.8 V vs. RHE for 10 h, the chemical structure of PPy@MOF-545-Co remained unchanged, revealing its excellent stability. The charge-transfer resistance of PPy@MOF-545-Co (7.5 Ω) was also lower than that of MOF-545-Co (12.5 Ω), indicating that it possessed a better electron transfer ability. By impregnating guest redox molecules into the framework structure, the charge can transfer directly from the guest molecule to the metalloporphyrin center and the strategy was demonstrated to significantly enhance the charge transfer efficiency.

Fig. 7

Reproduced with permission from Ref. [105]. c Schematic presentation for the advantages of PPy in the channel of MOF-545-Co for electrocatalytic CO₂RR. d FE_CO of PPy@MOF-545-Co and contrastive samples measured under different voltages. Reproduced with permission from Ref. [106]

a Comparison of MCp₂@MOF and MOF in electrocatalytic CO₂RR. b Electrocatalytic CO₂RR performances of CoCp₂@MOF-545-Co and comparative samples.

3.1.3 MOF-Derived Single-Atom Catalysts Materials

Recently, single-atom catalysts have been demonstrated to show excellent catalytic performance for various reactions due to their controllable properties and high atom utilization efficiency. Particularly, MOFs are ideal sacrificial templates to fabricate SACs because of the uniformly dispersed metal sites and abundant heteroatoms to immobilize the single metal site, resulting in a number of catalysts with high selectivity and remarkable activity for CO₂RR. For instance, Gong et al. [107] employ MgNi-MOF-74 as precursors to produce Ni_SA-Nx-C SACs (x is N coordination numbers) (Fig. 8a). The presence of Mg²⁺ in MgNi-MOF-74 can regulate and control the interatomic distance between adjacent Ni atoms, while N atoms from pyrolyzed polypyrrole (PPy) serve as anchoring sites to stabilize the Ni atoms. By controlling the pyrolysis temperature, they prepared three single-atom Ni catalysts with different N coordination numbers. Among them, Ni_SA-N₂-C (Fig. 8b) exhibits the highest selectivity for CO and the maximum FE is 98% at −0.8 V vs. RHE in CO₂-saturated 0.5 M KHCO₃ (Fig. 8c). This work not only provides a strategy for the fabrication of SACs, but also opens an avenue to enhance the activity of SACs for CO₂RR by controlling the metal coordination environment. Similarly, Chen et al. [108] report an amination strategy to regulate the electronic structure of SACs. As shown in Fig. 8d, they successful synthesized Ni–N₄/C–NH₂ SACs by a two-step method. A gas-tight H-type cell containing CO₂-saturated 0.5 M KHCO₃ electrolyte was used to evaluate the electrocatalytic activity of Ni–N₄/C–NH₂ and a maximum CO FE of 96.2% was achieved. Although the CO FE is slightly lower than that of Ni–N₄/C (98.1%), the CO partial current density of Ni–N₄/C–NH₂ is found to be significantly enhanced, which is 2.5 times that of Ni–N₄/C at −1.0 V vs. RHE (Fig. 8e). In order to meet requirements for industrial CO₂-to-CO production, they applied a gas-fed flow cell equipped with a GDE to solve the mass transfer limitation caused by low solubility of CO₂ in aqueous electrolysis. As shown in Fig. 8f, Ni–N₄/C–NH₂ achieves a remarkable CO partial current density of 447.6 mA cm⁻² at −1.0 V vs. RHE, which is 7.0 times that in an H-type cell and is much larger than that of Ni–N₄/C (250 mA cm⁻²), suggesting that amination treatment can indeed boost catalytic activity. DFT calculations revealed that the electronic structure of M–N/C catalysts are regulated by amino-modification (Fig. 8g), which enhances the adsorption energies of the reaction intermediates and accelerates the charge transfer rate, thus promoting the CO₂ activation and transformation process. Moreover, phthalocyanine-based MOFs can also be used as catalyst regulators to modify the electronic structure of metal center and enhance the CO₂RR to CO. For example, Lin et al. [109] took a synergistic catalysis strategy to boost the CO₂RR activity by anchoring Fe–N sites into CoPc (CoPc©Fe–N–C) through a sequential pyrolysis and post-impregnation method (Fig. 8h). CoPc©Fe–N–C shows CO FE of above 90% over the measured potential range of -0.13 to -0.84 V vs. RHE, exceeding its counterpart. Significantly, CO current density reaches 275.6 ± 27.0 mA cm⁻² at −0.84 V vs. RHE, which is sufficient to satisfy industrial requirements. The strong interaction between CoPc and Fe–N-C also reduces the *CO poisoning and accelerates desorption of CO. DFT calculations revealed that the adsorption energy of *CO and *H on CoPc©Fe–N–C is lower than that of Fe–N–C, while the *COOH formation energy does not change much (Fig. 8i), demonstrating unprecedented synergistic catalysis effect toward CO₂RR.

Fig. 8

Reproduced with permission from Ref. [107]. d Schematic of the synthesis process for Ni–N4/C–NH₂. e CO partial current density of Ni–N₄/C and Ni–N₄/C–NH₂. f Electrocatalytic activity of Ni–N₄/C–NH₂ in the flow cell. g Projected DOS of Ni 3d in Ni–N₄/C–NH₂ and Ni–N₄/C. Reproduced with permission from Ref. [108]. h Schematic illustration for the preparation of CoPc©Fe–N-C. The image on the far right is the calculated electron density difference of the CoPc©Fe–N-C structure. (Blue and yellow contours present electron depletion and electron accumulation, respectively. The isosurface level is set to be 0.0006 e Bohr-3). i Calculated free energy diagram for the CO₂RR to CO at U = -0.7 V versus RHE on the Fe site in CoPc©Fe–N-C, Fe site in Fe–N-C, Co site in CoPc©N–C, and Zn site in CoPc©Zn-N–C, respectively. Reproduced with permission from Ref. [109]

a Illustration showing the host–guest cooperative protection strategy for the fabrication of Ni_SA-N_x-C catalysts for electrocatalytic CO₂ reduction. b EXAFS fitting and optimized model for NiSA-N₂-C. c FEs of CO at different applied potentials in the CO₂-saturated 0.5 M KHCO₃ electrolyte.

3.2 Formate/Formic acid

HCOOH/HCOO⁻ is not only a hydrogen storage chemical with high energy density, but is also widely applied in the green synthesis of a range of petrochemicals in modern energy systems [117,118,119]. Moreover, among all kinds of products from CO₂RR, HCOOH possesses the highest profit per mole of electrons, exhibiting good economic prospects in large-scale industrial production [75]. To date, there are many MOF-related catalysts showing remarkable activity and selectivity for the production of HCOOH by CO₂RR. Several representative examples are shown in Table 3, we will discuss their structure–activity relationships from the perspective of metal composition, such as In, Sn, and Bi.

Table 3 Representative MOF-related catalysts for selectively catalyzing the electrochemical reduction of CO₂ to HCOOH

3.2.1 In-based MOF

Among the numerous catalyst materials that were studied, In-based catalysts were found to be selective toward the production of HCOOH. Recently, with the marriage of biomimetic and catalytic technology, a new revolution has emerged in structure design of In-based catalysts, which has significantly promoted the development of the CO₂RR field. For example, by mimicking the active [NiS₄] sites of formate dehydrogenase and CO-dehydrogenase, Zhou et al. [120] prepared a crystalline enzyme-mimicking three-dimensional In-based MOF, (Me₂NH₂⁺){In^III-[Ni(C₂S₂(C₆H₄COO)₂)₂]} 3DMF 1.5H₂O (In-MOF 1, DMF = N,N-dimethylformamide), with excellent chemical and thermal stabilities. Compared to the isomorphic MOF, (Me₂NH₂⁺) [In^III-(TTFTB)]·0.7C₂H₅OH·DMF (In-MOF 2, TTFTB = tetrathiafulvalene-tetrabenzoate), In-MOF 1 exhibited higher HCOOH selectivity during CO₂RR and the FE increased from 54.7% to 89.6% (-1.3 V vs. RHE) (Fig. 9a), demonstrating that the presence of unsaturated [NiS₄] sites in MOFs can significantly enhance the CO₂RR activity (Fig. 9b). DFT calculations further revealed that the formation energy of *HCOO intermediate on In-MOF 1 is much lower than that of In-MOF 2 (Fig. 9c), the introduced unsaturated [NiS₄] sites are proved to be the absorption and catalytic site for CO₂-to-HCOOH. Another In-based MOF was reported by Zhu et al. [121]. They constructed a stable 3D In-MOF (V11), {(Me₂NH₂) [In(BCP)]·2DMF}_n (H₄BCP = isophthalic acid) consisting two types of channels (1.6 and 1.2 nm diameter) (Fig. 9d). Subsequently, methylene blue molecules were introduced into the framework of V11 and then converted to carbon nanoparticles (CPs), forming V11-supported CPs (CPs@V11) (Fig. 9e). When tested in CO₂-saturated 0.5 M KHCO₃ electrolyte, the catalytic performance of CPs@V11 (methylene blue mass load of 10%) was significantly improved and the highest FE_HCOOH is 90.1% at −0.84 V vs. RHE. In addition, it also exhibited higher HCOOH partial current density than that of pure V11 at the same potential (Fig. 9f), obviously, the introduction of CPs via pyrolysis of MB greatly enhances the catalytic activity of the MOF. In-situ Fourier Transform infrared spectroscopy (FT-IR) spectra at different potentials showed that there are obvious absorption peaks of *HCOO intermediate at 1394 cm⁻¹, which gradually increase with decreasing potential (Fig. 9g). The introduced CPs not only improves electrochemical active surface area (ECSA) but also increases the conductivity, thus facilitating charge transfer (Fig. 9h) and enhancing the catalytic performance of the MOFs in terms of activity and selectivity. In addition, Qiu et al. [122] also synthesized an efficient In-based electrocatalyst (In₂O₃-x@C nanocorn) for converting CO₂ to HCOOH through a two-step process involving In MOF preparation and carbonization. When tested in 1 M KOH electrolyte, such electrocatalyst exhibited excellent catalytic activity and stability, due to its unique nanocorn structure, high concentration of active sites, and favorable electronic transfer properties. The operando experiments have confirmed that In³⁺ species as the catalytic active sites for the production of HCOOH. Furthermore, DFT calculations have revealed that the presence of oxygen vacancies creates an electron-rich environment for the In³⁺ active sites, leading not only to an enhancement of the reducing power at the active sites but also to a reduction in the energy barrier for electron transfer.

Fig. 9

Reproduced with permission from Ref. [120]. d The view of the two types of 1D channels in V11. e Illustration of the preparation process of CPs@V11. f The comparison of the FE HCOO- and the j HCOO- for various samples. g Potential-dependent in situ FTIR spectra of CPs@V11. h Nyquist plots for the samples in CO₂-saturated 0.5 M KHCO₃ electrolyte. Reproduced with permission from Ref. [121]

a Plots of FE_HCOO– for In-MOF 1 and In-MOF 2 versus applied potential. b Crystal structure of In-MOF 1 viewed along the a-axis, showing three rhombic pores and a two-fold interpenetrated framework (hydrogen atoms have been removed for clarity), and structures of ligands [Ni(C₂S₂(C₆H₄COOH)₂)₂] and H₄TTFTB with different conformations. c Proposed reaction paths for the formation of HCOOH on the [NiS₄] and [In(COO)₄].^– sites.

3.2.2 Sn-based MOF

Sn-based MOFs are another class of catalysts with excellent HCOOH selectivity. Geng et al. [123] prepared Sn-doped ZIF8 catalysts via an ion-exchange strategy, the method can efficiently integrate Sn into the node of ZIF-8 while preserving the whole framework structure. When tested in CO₂-saturated 0.5 M KHCO₃, Sn-doped ZIF8 showed high HCOOH activity and selectivity of 74% FE at −1.1 V vs. RHE. And the spatially separated Sn atoms are proposed to be responsible for the superior CO₂RR activity of Sn-doped ZIF8. Similarly, Deng et al. [124] also replaces the Zn metal nodes in ZIF8 by Sn doping, and then follows a solvent-assisted linker exchange (SALE) process to obtain the Sn-N6-MOF catalyst (Fig. 10a). When tested in CO₂-saturated 0.5 M KHCO₃, such catalyst achieved excellent selectivity for HCOOH with FE_HCOOH up to 85.1% at −1.23 V vs. RHE. In-situ Raman spectra indicate the organic ligands in MOF are gradually lost during the CO₂RR (Fig. 10b). Furthermore, ex-situ ¹¹⁹Sn Mössbauer results demonstrated the presence of zero-valence Sn metal after continuous electrolysis for 1 h at −1.23 V vs. RHE (Fig. 10c). A series of experiments results have revealed that Sn-N6-MOF catalyst undergoes in-situ structural reconstruction during the CO₂RR process and then generates Sn nanoclusters, which are the real active sites for producing HCOOH (Fig. 10d). In a previous study, Wu et al. [125] also discovered that Sn(101) crystal plane is more favorable for the formation of HCOOH (Fig. 10e). They synthesized a Sn-MOF (Sn₃O(1,4-BDC)₂) precursor to prepare a series of Sn(101)/SnO₂/C composites via the calcination and acidic etching processes (Fig. 10f). The Sn(101)/SnO₂/C-500 (500 is the carbonization temperature) exhibits high selectivity toward HCOOH and achieved a FE_HCOOH up to 93.3% at −0.8 V vs. RHE (Fig. 10g). However, the FE_HCOOH of such catalyst is gradually reduced from 93.3% to 72.8% after continuous electrolysis for 13 h at −0.8 V vs. RHE.

Fig. 10

Reproduced with permission from Ref. [124]. e Calculated free-energy diagrams for HCOO⁻,CO formation. f Schematic illustration showing the preparation of Sn(101)/SnO₂/C catalysts. g Electrocatalytic property of Sn(101)/SnO₂/C-500. Reproduced with permission from Ref. [125]

a Diagram of the synthetic procedures for Sn-N6-MOF. b In-situ Raman spectra measured with varying the acquiring time at -1.23 V vs. RHE in CO₂-saturated 0.5 M KHCO₃ aqueous. c Room temperature 119 Sn Mössbauer spectra acquired after maintained at -1.23 V vs. RHE in CO₂-saturated 0.5 M KHCO₃ aqueous for 1 h. d Proposed reaction pathway for the formation of HCOOH over Sn-N₆-MOF.

3.2.3 Bi-based MOF

Bi is also a promising electrocatalyst for CO₂RR to produce formate in aqueous solutions because of its large overpotential for HER in the aqueous electrolyte (Bi is situated at the bottom corner point of the volcano plot), low toxicity, and good stability [138, 139]. Therefore, Bi-based MOFs have drawn tremendous attention as catalysts for CO₂ reduction to HCOOH [140]. Zhang et al. [126] synthesized a Bi-MOF (Bi-BTC-D) by the hydrothermal method and evaluated its CO₂RR performance in the CO₂-saturated 0.5 M KHCO₃ electrolyte. Electrochemical experiments showed HCOOH selectivity with a maximum FE of 95.5% at -0.86 V vs. RHE. DFT calculations revealed that BTC ligands in MOF structure can effectively regulate the catalytic activity of the Bi atoms. Li et al. [127] also use the same ligand (H₃BTC) to construct a helical rod-based 2D Bi-MOF (CAU-17) with permanent crystallographic-independent channels (Fig. 11a), and it exhibits high CO₂-to-HCOOH activity in the H-cell. Such catalyst achieves an excellent FE of 92.2% at −0.9 V vs. RHE (Fig. 11b, c). It is worth noting that the type of catholyte significantly affects the selectivity of HCOOH (Fig. 11d). When SO₄²⁻ was added in the catholyte, the accumulated negative charges on the electrode surface can generate a potential difference in the electric double layer and facilitates the transport of polar water molecules instead of nonpolar CO₂ molecules, leading to the dominating competitive HER. Operando XAS (Fig. 11e) and DFT calculations were also conducted to explore the origin of the high HCOOH selectivity on the Bi-MOF, and the results showed that the highly accessible Bi³⁺ and the unique channels played vital roles in the enhancement of CO₂ adsorption and HCOOH conversion. Nevertheless, the maximum HCOOH partial current density is only 15 mA cm⁻¹ at −1.1 V vs. RHE, which is far from meeting the requirements for industrial applications (Fig. 11c). Yang et al. [128] also synthesized a Bi-MOF (CAU-17) with claviform shape and then spray on CP as the precursor, following an in-situ electroreduction process to fabricate leafy Bi nanosheets (Bi NSs) (Fig. 11f). Electrochemical experiments, which were conducted in a flow cell reactor, revealed that Bi NSs are excellent catalyst with high HCOOH activity and selectivity in both 1 M KHCO₃ or KOH electrolytes. The maximum FE of Bi NSs is up to 98% at the potential of −0.48 V vs. RHE (total current density up to 133 mA cm⁻²) (Fig. 11g). Significantly, the HCOOH partial current density reaches 374 mA cm⁻² at −1.51 V vs. RHE in 1 M KHCO₃. The outstanding performance may be associated with the hybrid Bi/Bi-O species on the surface of Bi NSs. DFT calculations also confirm that O atoms of the Bi-O surface may be beneficial to reduce the free energy barrier for *OCHO formation (Fig. 11h).

Fig. 11

Reproduced with permission from Ref. [127]. f Schematic illustration of the preparation procedure of Bi NSs. (purple, gray, orange, and yellow balls represent Bi, C, O, and H, respectively). g FEs and cathodic energetic efficiency (CEEs) of formic acid over two electrocatalysts in 1 M KOH. h Gibbs free energy profiles for CO₂ electroreduction to HCOOH on Bi NPs and Bi NSs. Reproduced with permission from Ref. [128]

a Schematic depiction of the formation of Bi-MOF. Bi-O polyhedra shown in the MOF crystal structure are indicated in purple. b HCOOH FEs, c HCOOH partial current densities of Bi-MOF, Bi sheets, bulk Bi₂O₃, and carbon paper electrodes within a potential window of −0.6 to −1.1 V in CO₂-saturated 0.1 M KHCO₃ solution. d HCOOH FEs and HCOOH partial current densities of Bi-MOF in various electrolytes. e Comparison of Bi L₃-edge X-ray absorption fine structure (XAFS) of Bi-MOF along with those for Bi metal and Bi₂O₃ as reference standards and represent their respective XANES and Fourier transform of EXAFS spectrum as a function of electrochemical bias and with electroreduction time under in-situ electrochemical CO₂ reduction conditions.

Thanks to the tunable porosity and well-defined architectures, MOFs are particularly appealing with regard to fabricating versatile carbon hybrid nanostructures with well-defined compositions and morphologies. CAU-17, which is an ideal sacrificial template to fabricate various metal/carbon hybrids catalysts, has been widely used in CO₂RR. For example, Deng et al. [129] construct carbon-nanorods-encapsulated bismuth oxides catalysts (Bi@C and Bi₂O₃@C) via the carbonization of CAU-17 in Ar and air atmosphere, respectively (Fig. 12a). The Bi₂O₃@C-800 shows a high FE_HCOOH of 92% at −0.9 V vs. RHE (Fig. 12b) in CO₂-saturated 0.5 M KHCO₃ electrolyte using an H-type reactor. However, the HCOOH partial current density is only 7.5 mA cm⁻², which was attributed to the low solubility of CO₂ in aqueous electrolytes. When they employ a flow cell configuration using 1 M KOH electrolyte, the HCOOH FE of Bi₂O₃@C-800 stays above 93% at −0.3 to −1.1 V vs. RHE and the HCOOH partial current density is significantly enhanced (Fig. 12c). The presence of crystalline Bi-O structure in the Bi₂O₃@C-800 is proved to be beneficial for promoting the reaction kinetics (Fig. 12d). The carbon matrix can significantly reduce the charge transfer resistance, which can promote the formation of *CO₂⁻ intermediates. Therefore, the synergistic effect of Bi₂O₃ nanoparticles and carbon matrix is beneficial to improve the activity and selectivity for CO₂-to-HCOOH. Ying et al. [130] prepared the CAU-17 fiber with a larger accessible surface area and abundant active catalytic sites via morphology engineering. Then, they calcined the fiber-shaped MOFs in an inert gas atmosphere to prepare Bi/C hybrids (CAU-17-fiber-x, x is the calcination temperature) for catalyzing CO₂ to HCOOH (Fig. 12e). The CAU-17-fiber-400 gets the highest CO₂-to-formate FE (96.4%), with a high partial current density (20.4 mA cm⁻²) at −0.9 V vs. RHE in CO₂-saturated 0.1 M KHCO₃ aqueous electrolyte using H-type cell (Fig. 12f). In the Bi/C hybrids, Bi nanoparticles (NPs) were encapsulated inside the CAU-17 derived porous fiber-shaped carbon framework, such a novel structure brought many unique benefits, such as larger accessible surface area and higher Bi content, thus improving the activity of CO₂RR. Wang et al. [131] used another H₃BTC ligand to prepare a Bi-based bimetallic MOF (BiIn-MOF), which was then calcined to obtain MOF-derived Bi/In bimetallic oxide nanoparticles/carbon (BiInO-x@C, x representative the ratio of Bi/In) in Ar atmosphere at 600 °C (Fig. 12g). Benefiting from the synergistic effect of bimetallic components, BiInO-0.67@C exhibits excellent activity and selectivity for the electroreduction of CO₂ to formate. They also applied in-situ FT-IR to reveal the catalyst mechanism of BiInO-0.67@C. As shown in Fig. 12h, the peak at 1430 cm⁻¹ is attributed to the symmetric stretching mode of OCO in the *HCOO species of the dioxygen bridge, indicating that the *HCOO pathway is the preferred route to produce formate on the catalyst.

Fig. 12

Reproduced with permission from Ref. [129]. e Schematic representation of preparation of CAU-17-derived electrocatalysts. f CAU-17-fiber series at various potentials in CO₂-saturated electrolyte based on a 2,500 s experiment. Reproduced with permission from Ref. [130]. g Schematic diagram of the preparation process of the BiInO-x@C catalyst. h In situ FT-IR spectra of BiInO -0.67@C at 1200 − 2200 cm⁻¹. Reproduced with permission from Ref. [131]

a Schematic illustration of the preparation procedure of Bi@C and Bi₂O₃@C catalysts. b FE of formate for various Bi-based composites in all potentials range. c FE and partial current density of HCOOH for Bi₂O₃@C-800 in all potentials range. d Tafel plots.

In addition to carbon-based materials, MOFs can also be used to prepare ultrathin metallene materials. As demonstrated by Cao et al. [132], ultrathin Bi-based metal–organic layers could serve as a pre-catalyst to produce atomically thin bismuthene (Bi-ene) following in-situ electrochemical reconstruction (Fig. 13a). The as-obtained Bi-ene shows an average thickness of 1.28–1.45 nm (Fig. 13b) and exposes more active sites due to the two-dimensional nature. As a result, Bi-ene could deliver a FE_HCOOH close to 100% at a wide potential range in both KHCO₃ and KOH electrolytes. Notably, the total current density can reach 200 mA cm⁻² at −0.75 V vs. RHE in 1 M KOH (Fig. 13c). In-situ ATR-IR spectroscopy and DFT analysis confirmed that HCOOH is generated through the *HCOO intermediate on Bi-ene. Furthermore, Yuan et al. [133] prepared a Bi-1,3,5-tris(4-carboxy-phenyl) benzene (Bi-BTB) and discovered that the bismuth-carboxylate MOFs can be in-situ transformed to Bi₂O₂CO₃ in an HCO₃⁻ electrolyte. Bi-BTB exhibits an outstanding CO₂-to-HCOOH performance (Fig. 13d). After electrolysis, the crystalline phase of Bi-BTB disappeared from the X-ray diffraction (XRD), while the peaks intensities of a new crystalline phase of Bi₂O₂CO₃ dramatically increase, indicating that the true catalytic species are Bi₂O₂CO₃ (Fig. 13e). Since Bi³⁺ ion and carboxylate belong to intermediate acid and hard base, Bi-BTB is not stable. Therefore, in the presence of HCO₃⁻, the Bi–O bonds in Bi-BTB can be broken, resulting in the structural evolution from Bi-BTB to Bi₂O₂CO₃ (Fig. 13f). A Tafel slope of 119 mV dec⁻¹ for Bi₂O₂CO₃ indicate that the initial electron transfer step is the RDS for the CO₂RR (Fig. 13g). This work shows a good example of surface reconstruction and gives a strong signal that careful evaluation is required to distinguish the real reactive sites for different MOF electrocatalysts.

Fig. 13

Reproduced with permission from Ref. [132]. d FE of production and the current densities of HCOOH with Bi-BTB at different working potentials in CO₂-saturated 0.5 M KHCO₃ electrolyte. e The changes of XRD patterns for various samples. f Electrochemical cell for the electrolysis experiments and proposed mechanism for the formation of MOF-derived Bi₂O₂CO₃. g Tafel plot for MOF-derived Bi₂O₂CO₃ in CO₂-saturated 0.5 M KHCO₃ electrolyte. Reproduced with permission from Ref. [133]

a Synthesis and characterizations of Bi-MOLs and Bi-ene. b Atomic force microscopy images of Bi-ene. c Chronopotentiometric curves at 100 and 200 mA cm.⁻² in 1.0 M KHCO₃ and KOH.

3.3 Methane

Among all products derived from CO₂RR, CH₄ has attracted significant attentions because of its high values of mass heat (56 kJ g⁻¹), good compatibility with current energy infrastructure [141,142,143]. Furthermore, the CO₂RR reaction for CH₄ formation is thermodynamically more favored than the reaction for both CO and HCOOH. However, as a deepest reduction product, the formation of CH₄ involves eight electrons and sluggish kinetics, resulting in high overpotential and low selectivity. Therefore, it is extremely attractive to design catalysts with high activity and selectivity for CO₂ reduction to CH₄. To date, a number of MOF-related catalysts have been proved to be capable of promoting CO₂RR toward CH₄ (Table 4).

Table 4 Representative MOF-related catalysts for selectively catalyzing the electrochemical reduction of CO₂ to CH₄

Generally, Cu-based catalysts show the best catalytic activity for the selective production of CH₄ in the CO₂RR system. For example, Kim et al. [144] prepared highly isolated Cu nanoparticles (Cu NPs) clusters with an average size of 30–50 nm for CO₂-to-CH₄, which were synthesized from Cu-MOF-74 precursor via an electroreduction process (Fig. 14a). When tested in 0.1 M KHCO₃, the as-obtained Cu NPs clusters exhibited good catalytic activity toward CH₄. As shown in Fig. 14b, the maximum FE of CH₄ is approximately 50% at −1.3 V vs. RHE, while only 35% is achieved by commercial Cu NPs at the same conditions. It should be also noted that Cu NPs show the highest activity for CH₄ at −1.3 V vs. RHE, and the CH₄ partial current densities on Cu NPs are 2.3 times that of the commercial Cu NPs. It is believed that the activity difference is attributed to the extent of aggregation of Cu particles at nanoscales and MOF-derived Cu NPs are found to be less aggregated. Yang et al. [145] employed adenine and acetic acid ligands to fabricate Cu-ade MOF with different thicknesses (Cu-ade nanosheets (s-Cu-ade), nanoplates (p-Cu-ade) and nanocuboids (c-Cu-ade)). Figure 14c shows the molecular structure of the Cu-ade monomer, where stable MOFs were formed by the bonding of N with C and Cu. Electrochemical measurements demonstrated that s-Cu-ade has the best CO₂ reduction performance in CO₂-saturated 0.1 M KHCO₃ electrolyte and the maximum FE of CH₄ is over 50% at -1.6 V vs. RHE (Fig. 14d). It worth noting that the CO₂ electroreduction process can induce the structure evolution of the Cu-MOF to form Cu nanoparticles functionalized by the nitrogen containing ligands (Fig. 14e, f). The presence of N-containing functional groups would activate the protons to obtain COH* or CHO* intermediate [153,154,155,156], which are critical intermediates for further hydrogenation to form CH₄, thus boosting the conversion of CO₂ to CH₄ (Fig. 14g).

Fig. 14

Reproduced with permission from Ref. [144]. c Molecular structure of the Cu-ade monomer. d FE of CH₄ and C₂H₄ for the s-Cu-ade MOF. e N 1s scan XPS patterns of the initial and cathodized Cu-ade MOFs. f XRD pattern of the cathodized s-Cu-ade MOF. g Proposed Cu-ade MOF evolution. Reproduced with permission from Ref. [145]

a Schematic illustration of the hydrothermal synthesis of Cu-MOF-74 and preparation of Cu NPs from Cu-MOF-74 by electroreduction. b Faradaic efficiencies for C₁ and C₂ hydrocarbons production and partial current density for CH₄ production on Cu NP electrodes at the applied potential.

As discussed above, achieving high hydrocarbon selectivity remains a great challenge. Fortunately, the synergistic strategy that combines Cu₂O with Cu MOFs seems to be an effective way to enhance hydrocarbon selectivity, which has been confirmed by recent studies. For instance, Tan et al. [146] prepared an all-in-one hybrid Cu₂O@Cu-MOF by time-resolved controllable restructuration. Briefly, the surface of Cu₂O spheres can be oxidized to Cu²⁺ in the mixed alcohol solution at 80 °C, and then Cu²⁺ can further coordinate with H₃BTC to form Cu-MOFs on the surface of Cu₂O, resulting in Cu₂O@Cu-MOF (Fig. 15a). When tested in CO₂-saturated 0.1 M KHCO₃ solution, Cu₂O@Cu-MOF showed considerable FE_CH4, which was significantly higher than that of both Cu₂O and Cu-MOF, and the maximum FE was up to 63.2% at −1.71 V vs. RHE (Fig. 15b). Due to the porous nature of the framework, Cu₂O@Cu-MOF exhibited considerable adsorption capacity of CO₂ molecules (Fig. 15c), which significantly enlarge the local CO₂ concentration on the active sites of the electrode, while the intrinsic catalytic activity of Cu₂O can be maintained well simultaneously. Moreover, the Cu₂O core embedded in the Cu-MOF accelerates charge transfer. On the contrary, Yi et al. [147] fabricated Cu₂O(111) quantum dots with an average size of 3.5 nm on a porous conductive Cu-MOFs (CuHHTP) via an electroreduction process (Fig. 15d). Linear sweep voltammetry (LSV) tests are conducted in both Ar and CO₂-saturated 0.1 M KCl + 0.1 M KHCO₃ mixture solution. As shown in Fig. 15e, compared with pristine Cu-MOFs, the Cu₂O(111)@CuHHTP shows larger current densities at same conditions, implying that Cu₂O(111) quantum dots have high electrocatalysis activity for the CO₂RR. Notably, Cu₂O(111)@CuHHTP achieved high selectivity of 73% at -1.4 V vs. RHE toward CH₄ with partial current density up to -10.8 mA cm⁻² (Fig. 15f). The superior electrochemical performance was attributed to the following two reasons: 1) Due to the strong charge delocalization between Cu²⁺ and HHTP, Cu-MOFs shows excellent electronic conductivity (5.1 × 10^–5 S m⁻¹) and serves as a conductive substrate, which can accelerate the electron transfer to Cu₂O(111) during the CO₂RR process. 2) During electroreduction, Cu-O₄ nodes in Cu-MOFs were partially reduced to Cu₂O, thus exposing a number of hydroxyl groups of the uncoordinated HHTP ligand. The hydroxyl-rich environment was assumed to stabilize the *CO intermediate by hydrogen bonding (Fig. 15g) and improves the selectivity toward CH₄ [157,158,159].

Fig. 15

Reproduced with permission from Ref. [146]. d Illustration of the solvothermal synthesis of CuHHTP and preparation of Cu₂O@CuHHTP via electrochemical treatment of CuHHTP at the applied potential of -1.2 V vs. RHE for 30 min. e LSV curves of CuHHTP and Cu₂O@CuHHTP in 0.1 M KCl/0.1 M KHCO₃ electrolyte under Ar and CO₂. f Comparison of CH₄ FE between Cu₂O@CuHHTP, Cu₂O on conductive carbon black(Cu₂O@CCB), and commercial Cu₂O. g Proposed mechanism of Cu₂O@CuHHTP for the formation of CH₄. Reproduced with permission from Ref. [147]

a Schematic illustration of the process to synthesize Cu₂O@Cu-MOF. b FEs of CH₄ and C₂H₄ and the ratio of CH₄ to C₂H₄ for Cu₂O@Cu-MOF, Cu-MOF, and Cu₂O at − 1.71 V versus RHE in CO₂-saturated 0.1 M KHCO₃ solution. c CO₂ adsorption curves of Cu₂O@Cu-MOF, Cu-MOF, and Cu₂O.

Apart from Cu₂O, many other Cu(I)-based catalysts have also been demonstrated to be active for the selective reduction of CO₂ to CH₄. For example, Zhu et al. [148] prepared a Cu-MOF, Cu₄ZnCl₄(btdd)₃ (Cu₄^II-MFU-4 l, H₂btdd = bis(1H-1,2,3-triazolo-[4,5-b],[4′,5′-i])dibenzo-[1, 4]-dioxin), by an ion exchange process that uses Cu(II) ions to replace outer sphere Zn(II) ions in Zn₅Cl₄(btdd)₃ (MFU-4 l) cluster (Fig. 16a). When tested in CO₂-saturated 0.5 M NaHCO₃ solution, Cu₄^II-MFU-4 l showed obvious electrocatalytic CO₂RR activity and achieved a high CH₄ FE of 92% at -1.2 V vs. RHE (Fig. 16b), while MFU-4 l only yielded minor CO product at the same condition. After long-term electrolysis, the Cu(II) ions were found to be reduced to Cu(I) ions (Cu₄Zn-(btdd)₃) with the trigonal pyramidal Cu(I)N₃ sites, indicating that the Cu(I)N₃ sites is the actual active centers for the CO₂RR (Fig. 16c). Figure 16d shows the conversion process from Cu(II) to Cu(I) ions during the electrocatalysis. Moreover, in situ experiments and DFT calculations revealed that the synergistic interactions between Cu(I)N₃ sites and adjacent aromatic hydrogen atoms can stabilize the key CO₂-to-CH₄ intermediates via hydrogen bonding. In order to understand the effect of intrinsic cuprophilic interactions inside the Cu(I) catalysts, Zhang et al. [149] used 2-(5-(3-(5-(pyridin-2-yl)-1H-,2,4-triazol-3-yl)phenyl)-1H-1,2,4-triazol-3-yl) pyridine (H₂bptb) and Cu(I) ions to construct two MOFs (NNU-32 and NNU-33(S) (S = sulfate radical)). Then they transformed NNU-33(S) into NNU-33(H) (H = hydroxyl) through an anion exchange process in 1 M KOH solution (Fig. 16e). After structural transformation, NNU-33(H) showed enhanced cuprophilic interactions and expanded interlayer distances. The electrocatalytic CO₂RR performance was tested in 1 M KOH solution using a flow cell and the results showed that NNU-33(H) exhibited a high FE_CH4 value of 82.17% at -0.9 V vs. RHE (Fig. 16f). Significantly, the current density is up to 391.79 mA cm⁻² at -0.9 V vs. RHE, which is sufficient to satisfy the requirements of industrial applications. The XAS, in situ Raman, XPS, and in situ FTIR were performed to confirm the structural stability of NNU-33(H) and those results suggested that the properties of NNU-33(H) are stable during CO₂RR. The distances of Cu(I)-Cu(I) in both NNU-32 and NNU-33(H) are a little larger than twice the covalent radius of Cu but significantly shorter than twice its van der Waals radius, suggesting the existence of cuprophilic interactions in the crystal. Moreover, the distances of Cu ions in NNU-33(H) are shorter than that NNU-32, illustrating stronger cuprophilic interactions in NNU-33(H), which may be the key factor that influences the CH₄ selectivity. DFT calculations revealed the fourth hydrogenation step (*H₂COOH → *OCH₂) is the potential determining step (PDS) (Fig. 16g). Significantly, the free energy of the PDS process is increased significantly from 0.74 to 1.11 eV after removing the Cu(I)-Cu(I) interaction, confirming that the internal Cu-Cu interaction plays an essential role in the CO₂RR process. However, the durability of NNU-33(H) still needs to be improved (Fig. 16h).

Fig. 16

Reproduced with permission from Ref. [148]. e Structures of {Cu₈} clusters and unit cell in NNU-33(S) and NNU-33(H), respectively. f Electrocatalytic performances of NNU-33(H) FE of H₂, CO, CH₄, and C₂H₄ products. g Calculated free energy diagram and the corresponding intermediates for CO₂ electrocatalytic reduction to CH₄ on the Cu₈ model catalyst. h Current profile and FEs of CH₄ at a constant voltage of -0.9 V vs RHE. Reproduced with permission from Ref. [149]

a Structures of local coordination environments, metal transformation process, and Cu₄^II-MFU-4 l (3D channel surface highlighted in yellow). b FE and TOF by Cu₄^II-MFU-4 l. c Cu K-edge EXAFS spectra and fitting for CuN₃C at -1.2 V under CO₂. d Illustration of the conversion from Cu(II) to Cu(I) ions and the formation of intermediates during the electrocatalysis.

As discussed before, reasonable regulation of the Cu coordination environment in MOFs is a common method to modulate the selectivity of electroreduction catalysis, this is also applicable for the CO₂-to-CH₄ process using MOF-based electrocatalysts. For example, Zhang et al. [150] employed the highly conjugated organic ligand (dibenzo-[g,p] chrysene-2,3,6,7,10,11,14,15-octaol, 8OH-DBC) to construct a Cu-based conductive MOFs (Cu- DBC) with abundant and uniformly distributed Cu-O₄ sites (Fig. 17a). Cu-DBC exhibits an electrical conductivity of 1.2 × 10^–2 S m⁻¹ due to the charge delocalization between metal ions and conjugated ligands. Electrochemical experiments reveal that Cu-DBC delivers obvious CO₂RR activity with a maximum CH₄ FE of 80% at -0.9 V vs. RHE (Fig. 17b). The experimental measurements and DFT calculations further revealed that the Cu-O₄ site in Cu-DBC is easier to be reduced into low-valence Cu sites during the activation process and is more energetically favorable for the following CO₂ reduction compared to nitrogen-coordinated Cu sites (Fig. 17c). Liu et al. [151] designed a conjugated, nitrogen-containing ligand hexahydroxyl-hexaazatrinaphthylene (HATNA-6OH) and further synthesized 2D conductive MOFs (Cu₃(C₂₄H₆O₆N₆)₂ 1.5(NH₃CH₂CH₂NH₃), HATNA-Cu-MOF) by a solvothermal method in the presence of ethylenediamine (Fig. 17d). Owing to the synergic effects of the redox active copper catecholate nodes and favorable p-p stacking of the ligand, HATNA-Cu-MOF exhibits high CH₄ selectivity with a FE of 78% at -1.5 V vs. RHE (Fig. 17e). Unfortunately, the CH₄ partial current density is only -8.2 mA cm⁻². In addition, Zhang et al. [152] prepared a series of Cu₄X cluster-based MOFs ([Cu₄X(TIPE)₃]·3X, [X = Cl, Br, I, TIPE = 1,1,2,2-tetrakis(4-(imidazol-1-yl)phenyl)ethene], named as Cu-Cl, Cu-Br, Cu-I) to investigate the effect of different halogens atoms on the activity and selectivity of CO₂RR products (Fig. 17f). Electrochemical experiments revealed that Cu-I is the optimal catalyst for CO₂ reduction to CH₄ with the highest FE_CH4 of 57.2% at −1.08 V vs. RHE (Fig. 17g). Meanwhile, the CH₄ partial current density is up to 60.7 mA cm⁻² at the same potential. In order to explore the origin of the high activity and selectivity of Cu-I, DFT calculations was also conducted. As shown in Fig. 17h, when the radius of halogen atom increase, the d-band center of Cu site is shifting closer to the Fermi level, which can significantly reduce the formation energy of the PDS and promote subsequent CH₄ formation.

Fig. 17

Reproduced with permission from Ref. [150]. d Synthesis of HATNA-Cu-MOF. e Potential dependent FE of different reduction products. Reproduced with permission from Ref. [151]. f Schematic illustration of the synthesis and structure of Cu-X. g Average FEs of CH₄ at different potentials over Cu-Cl, Cu-Br, and Cu-I catalysts. h Gibbs free energy profiles of CO₂ reduction reaction on Cu-Cl, Cu-Br, and Cu-I. Reproduced with permission from Ref. [152]

a Structure obtained by Cu ions and 8OH-DBC. b FEs of CO₂RR products at different applied potentials. c Free energy profiles for the CO₂RR-to-CH₄ reaction pathway.

3.4 Methanol

CH₃OH is one of the important chemicals in the production of organic compounds and synthetic gasoline [160]. Moreover, CH₃OH is a promising liquid fuel to replace fossil fuels because of its environmental friendliness and ease of transportation [161, 162]. CH₃OH can also be directly used in conventional internal combustion engines or in direct methanol fuel cells, making it stand out as the alternative fuel for building a sustainable society [163]. Recently, the catalytic CO₂ hydrogenation to CH₃OH by MOF-related materials has been investigated and several representative examples are listed in Table 5.

Table 5 Representative MOF-related catalysts for selectively catalyzing the electrochemical reduction of CO₂ to CH₃OH

Two impressive studies, which employed Cu-BTC as the precursor to fabricate catalysts with high selectivity for CH₃OH were reported. In one study, Zhao et al. [164] carbonized Cu-BTC to synthesize oxide-derived Cu nanoparticles in a porous carbon matrix (OD Cu/C) under Ar atmosphere (Fig. 18a). Subsequently, they loaded OD Cu/C catalysts on CP as a working electrode and tested its CO₂RR performance in CO₂-saturated 0.1 M KHCO₃ solution. As shown in Fig. 18b, OD Cu/C-1000 (1000 is the carbonization temperature) exhibits high selectivity and activity for CO₂ reduction to CH₃OH, with a maximum FE of 43.2% at −0.3 V vs. RHE. Significantly, the overpotential for CH₃OH formation is only 190 mV. The synergistic effect between the highly dispersed copper and the porous carbon is beneficial for converting the adsorbed CO to alcohol, thus improving the activity and selectivity for CH₃OH. Furthermore, the existence of a carbon matrix can protect the active sites' deactivation during the electrochemical reduction of CO₂, thus enhancing the durability of OD Cu/C-1000. Another study was reported by Yang et al. [165]. Briefly, the PVP-modified Cu-BTC was synthesized by hydrothermal method and then calcined in air to obtain a carbon-supported Cu@Cu₂O catalyst (Fig. 18c). Three samples were calcined at different temperatures, and Cu@Cu₂O−400 ℃ was found to show distinct activity for catalyzing CO₂ to CH₃OH. The maximum FE_CH3OH was up to 45% at −0.7 V vs. RHE and continuous electrolysis for 2 h was demonstrated in CO₂-saturated 0.5 M KHCO₃ aqueous solution (Fig. 18d). Compared with other samples, Cu@Cu₂O-400℃ has the highest concentration of Cu⁺, which could not only enhance the absorption capacity of the CO* intermediate but also promote its protonation on C site to form CHO*. Furthermore, the synergistic effect between Cu⁰ and Cu⁺ could adjust CO* binding energy, and the surface OH groups could help to increase the concentration of the *H, both facilitating the CH₃OH formation following the protonation or CPET process. These conclusions are also confirmed by in-situ ATR-IR measurements. In addition, Payra et al. [166] also prepared carbon-supported intermetallic alloys as electrocatalysts toward CO₂-to-CH₃OH. Intermetallic PtZn/C, Pt₃Zn/C, and Pt_xZn/C (1 < x < 3) NPs on N-doped carbon were synthesized by thermal decomposition of ZIF-8 under an inert atmosphere (Fig. 18e). Three synthesized intermetallic nano-alloys were deposited on glassy carbon electrode and tested in CO₂-saturated 0.1 M NaHCO₃ solution. The mixed-phase Pt_xZn/C was found to show the highest FE of up to 81.4% at −0.9 V vs. RHE (Fig. 18f). Figure 18g shows the possible reaction paths for CO₂ reduction over the intermetallic nano-alloys and the CH₃OH selectivity is governed by the bonding strength of the surface *–OCH₃ species. A weak interaction between O and the catalysts surface could facilitate desorption of the whole *–OCH₃ group, and thus increasing the selectivity of CH₃OH. Theoretical calculations revealed that Pt_xZn has the lowest bonding energies of *OCH₃ compared to the other nano-alloys, which was also confirmed by the results of *–OH adsorption strength on the surface of the catalytic electrode (Fig. 18h).

Fig. 18

Reproduced with permission from Ref. [164]. c Scanning electron microscope (SEM) images of Cu@Cu₂O electrocatalysts derived from Cu-BTC pyrolysis at 400 °C. d CH₃OH over Cu@Cu₂O-T electrocatalysts at various applied potentials in CO₂-saturated 0.5 M KHCO₃. Reproduced with permission from Ref. [165]. e High resolution transmission electron microscopy (HR-TEM) images of PtxZn/C and corresponding particle size distribution. f FE of CH₃OH in CO₂RR over the intermetallic nano-alloys as a function of potentials. g Possible mechanism of CO₂RR with product distribution over intermetallic nano-alloys. h CV traces for hydroxide adsorption in 0.1 M NaOH solution over the intermetallic nano-alloys. Reproduced with permission from Ref. [166]

a Synthesis process of oxide-derived Cu/Carbon catalysts. b EE for CO₂ electrochemical reduction of OD Cu/C-1000.

Downsizing the active metal component to the atomic level is also a promising way to improve the CH₃OH selectivity. For example, Yang et al. [167] dispersed numerous isolated Cu atoms in through-hole carbon nanofibers to synthesize flexible and self-supported single-atom catalysts (CuSAs/TCNFs) with high CO₂-to-CH₃OH efficiency. Briefly, they embedded Cu/ZIF-8 precursor into PAN nanofiber and then carbonize it to form the isolated Cu atoms (Fig. 19a). Notably, no metal clusters or nanoparticles are found in the CuSAs/TCNFs (Fig. 19b). Thanks to the good mechanical strength and high conductivity, CuSAs/TCNFs can be directly used as cathodes without binder or current collector. As shown in Fig. 19c, due to the synergetic effect of the both through-hole carbon structure and abundant isolated Cu active sites, the FEs for CH₃OH reached a maximum value of 44% at -0.9 V vs. RHE when tested in CO₂-saturated 0.1 M KHCO₃ electrolyte. DFT calculation showed that the Gibbs free energy for *COH to *CHOH (~ 0.86 eV) on the active sites (Cu-N₄) is significantly lower than that for *COH to *C (~ 1.88 eV), thus CH₃OH instead of CH₄ is more thermodynamically favored (Fig. 19d). In addition, Liu et al. [168] also reported a honeycomb-like single-atom catalyst (M₃(HHTQ)₂, M = Cu, Ni; HHTQ = C₃-symmetric 2,3,7,8,12,13-hexahydroxytricyclo-quinazoline), where the Cu or Ni atoms are uniformly anchored in the hexagonal lattices (Fig. 19e, f). The content of metal atoms was up to 20% in the M₃(HHTQ)₂ and the interplay between both metal centers and nitrogen-rich organic ligands in Cu₃(HHTQ)₂ significantly improved the CO₂-to-CH₃OH selectivity, showing a FE up to 53.6% at a low overpotential of 0.4 V (Fig. 19g). DFT calculations were also conducted to reveal the catalyst mechanism of Cu₃(HHTQ)₂ and the most favorable reacting pathway was proposed as follows: CO₂(g) → *OHCO → *HCOOH → *CHO → *CH₂O → *CH₂OH → *CH₃OH → * + CH₃OH (Fig. 19h, i). This work provides new insights into designing novel 2D conductive MOFs as electrocatalysts for CO₂RR by enhancing the interplay between metal centers and organic ligands.

Fig. 19

Reproduced with permission from Ref. [167]. e Synthesis of M₃(HHTQ)₂ (M = Cu, Ni), C gray, N blue, O red, Cu purple, H white spheres. f Zoom-in view of HR-TEM image for Cu₃(HHTQ)₂ taken along the c axis that shows a hexagonal pore and ligand termination, overlaid with a structure model. g CH₃OH FEs for Cu₃(HHTQ)₂, Ni₃(HHTQ)₂, Cu₃(HHTP)₂ at different potentials. h Free energy profiles for the CO₂RR on Cu₃(HHTQ)₂ and the proposed catalytic mechanism for electrochemical reduction of CO₂ to CH₃OH by CuO₄ sites of Cu₃(HHTQ)₂. i Structures of the catalyst and key reaction intermediates involved in the proposed reaction mechanism for the CO₂RR on Cu₃(HHTQ)₂. Reproduced with permission from Ref. [168]

a Synthesis procedure of CuSAs/THCF: I, adsorption of Cu ions; II, electrospinning of polymer fibers; III, carbonization and etching. b HR-TEM images of CuSAs/TCNFs; the inset of b shows the SAED pattern. c FEs of all products at CuSAs/TCNFs. d Free energies for conversion of *CO to CH₃OH on Cu-N₄ structure. Orange, gray, dark blue, red, and light blue spheres stand for Cu, C, N, O, and H atoms, respectively.

3.5 Multi-carbon (C₂₊) Species

C₂₊ chemicals such as C₂H₄ and CH₃CH₂OH, are attractive and desirable high-energy–density CO₂RR products [169,170,171]. However, it is still challenging to produce C₂₊ chemicals through CO₂RR because of the high energy barrier and low selectivity for C–C coupling in aqueous electrolytes [172]. MOF-related materials provide a promising way for catalyzing CO₂ to C₂₊ products, several representative examples are listed in Table 6.

Table 6 Representative MOF-related catalysts for selectively CO₂RR toward C₂₊ products

Cu-based MOF-related catalysts exhibit remarkable activity and selectivity for the production of C₂₊ chemicals in CO₂RR because of their moderate binding energy to *CO intermediates, which is critical for the formation of C–C coupling. Zhang et al. [173] designed a 3D Cu-base catalyst (Cu GNC-VL) by introducing copper precursors into the oriented ZIF-L-derived GO nanosheets (vZIF-L@GO). Amorphous Cu/Cu₂O NPs are found to uniformly disperse inside the Cu GNC-VL catalyst and facilitates electron transport. The synergistic effect between Cu (111) and Cu₂O (111) not only enhances the CO₂ adsorption but also facilitates the C–C coupling, thus resulting in a high CH₃CH₂OH FE of 70.52% and a total current density of 10.4 mA cm⁻² at -0.87 V vs. RHE. As mentioned above (Sect. 4.1), Cu-THQ exhibits excellent performance for the electroreduction of CO₂ to CO, thus creating a high local concentration of *CO on the catalyst surface. Zhao et al. [174] borrowed this idea and prepared a tandem catalyst (Cu(111)@Cu-THQ) via in-situ electroreduction. The integration of two kinds of active catalytic sites increased the *CO intermediate coverage on the Cu surface and also reduced the C–C coupling energy barrier. Owing to the synergistic effect of two active sites (Cu(111) and CuO₄ nodes), such catalyst delivers a high C₂H₄ FE of up to 44.2% at -1.2 V vs. RHE in CO₂-saturated 0.1 M KHCO₃ electrolyte. Han et al. [175] prepared an HKUST-1 thin film to investigate the relationship between the structural evolution of these materials and product selectivity during continuous electrolysis. Electrochemical experiments showed that the catalytic activity of HKUST-1 varies with electrolysis time. After continuous electrolysis for 15 min, HKUST-1 was found to evolve into 3D nanospheres made from numerous small fragments (HKUST-1₁₅). When electrolysis time increased to 120 min, HKUST-1 was thoroughly transformed into cross-linked nanobelts (HKUST-1₁₂₀). Compared with HKUST-1₁₂₀, HKUST-1₁₅ had a higher Cu⁺/Cu ratio on the surface, which greatly promotes CO₂ activation and facilitates the C–C coupling, thus exhibiting higher activity for CO₂ reduction to CH₃CH₂OH and C₂H₄. Another study reported Cu-MOF-derived mesoporous Cu nanoribbons by a similar in-situ electrochemical reduction method [176]. The inherited mesoporous structure of Cu contains many edges and pores that induce an enhanced electric field, which could not only reduce the thermodynamic energy barrier for the formation of CO by forming concentrated K⁺ on the active site [182], but also confine the OH⁻ ions diffusion and thus promote the local pH, leading to enhanced selectivity of C₂₊ chemicals [183,184,185]. A flow cell was used to evaluate the performance of the mesoporous Cu nanoribbons for CO₂RR in 1 M KOH electrolyte. The results show that such a catalyst can selectively produce C₂₊ product with FE up to 82.3% at the potential of −1.2 V vs RHE. At the same time, the partial current density toward C₂₊ product is up to 347.9 mA cm⁻², which is sufficient to satisfy the requirements for industrial applications.

It is universally acknowledged that the linker or ligand in MOFs plays an essential role in the selective reduction of CO₂ to value-added chemicals. Choosing an appropriate linker not only determines the distance between metal centers but also regulates the electronic structure of MOFs, thereby influencing the activity and selectivity of CO₂RR. For example, Qiu et al. [177] employed PcCu-(OH)₈ (2,3,9,10,16,17,23,24-octahydroxy- phthalo-cyaninato)copper(II)) and CuCl₂ to assemble a PcCu–Cu–O MOF (Fig. 20a). The as-obtained PcCu-Cu–O MOF was further exfoliated to small particles and coated on a glassy carbon electrode to prepare the working electrode. They used an H-type cell and CO₂-saturated 0.1 M KHCO₃ aqueous solution to evaluate the CO₂RR performance of such material. As shown in Fig. 20b, such catalyst exhibits high selectivity toward C₂H₄ (FE = 50% at -1.2 V vs. RHE) and shows good stability after 4 h continuous electrolysis, reflecting the significant synergistic effect between CuPc and CuO₄ nodes. Periodic DFT (PDFT) calculations show that the CuO₄ unit serves as an active site for the formation of *CO, while the CuPc unit has a high adsorption energy for *CO and is beneficial for the hydrogenation of *CO toward *CHO. Further, the CO molecule desorbed on the CuO₄ unit can easily dimerize with the absorbed *CHO intermediate on CuPc unit to form the *COCHO intermediate, and C₂H₂ is finally obtained through multi-step reactions (Fig. 20c). In situ ATR-FTIR analysis also confirmed the proposed C₂H₄ mechanism. As demonstrated by Wen et al. [178], sulfur doping in HKUST-1 can change the coordination environment of Cu nodes and generate a precatalyst (S-HKUST-1) with uniformly dispersed Cu–S motifs (Fig. 20d), and their in-situ reconstruction can yield abundant active biphasic Cu/Cu_xS_y interfaces, which show good C₂H₄ selectivity and activity. Electrochemical experiments revealed that such a catalyst gets the highest C₂H₄ FE of 57.2% with a total current density of up to 400 mA cm⁻². Compared with HKUST-1, S-HKUST-1 can derive more Cu^δ+ species at the Cu/Cu_xS_y interfaces after the reconstruction process. DFT calculations show that the S-stabilized Cu^δ+ at the Cu/Cu_xS_y interface has an optimized geometric and electronic structure for *CO dimerization (Fig. 20e), thus improving the CO₂-to-C₂H₄ selectivity. In addition, Xie et al. [179] demonstrated that impregnating Au nanoneedles (AuNN) into a PCN-222(Cu) MOF (AuNN@PCN-222(Cu)) is also an effective way to improve the CO₂-to-C₂H₄ performance (Fig. 20f). Electrochemical experiments revealed that the catalyst exhibits a maximal C₂H₄ FE of 52.5% at s−1.2 V vs. RHE (Fig. 20g). A series of experiments has been conducted to explore the reaction mechanisms and the results show that AuNN@PCN-222(Cu) serves as a tandem catalyst: CO generated by AuNN are further reduced on the metalloporphyrin sites, which is consistent with theoretical results (Fig. 20h). The AuNN plays a cable-like role in the PCN-222(Cu), which provides a charge transfer path to the metalloporphyrin center, thus enhancing the structural stability of AuNP@PCN-222(Cu).

Fig. 20

Reproduced with permission from Ref. [177]. d Schematic illustration of the preparation of S-HKUST-1, which indicates a local H₂O molecule might be replaced by S atoms. e The reaction barriers together with enthalpies and corresponding transition state configurations for *CO dimerization and hydrogenation over Cu(111) and Cu/Cu₂S surfaces, respectively. (Yellow, red, gray, white, orange and blue balls refer to S, O, C, H, Cu0 and Cu^δ+ atoms, respectively). Reproduced with permission from Ref. [178]. f Impregnation of Au nanoneedles into PCN-222(Cu) with cleaved ligand-node linkage to alter the charge conduction path and steer the CO₂RR pathway. g FEs of various reduction products for AuNN@PCN-222(Cu). h Formation energy of key intermediates along with the reaction coordinates for TCPP(Cu) and TCPP(Cu)-Au₁₂. Reproduced with permission from Ref. [179]

a Illustration of the structure of PcCu-Cu–O. b FEs of C₂H₄, CH₄, CO, and H₂ for PcCu-Cu–O. c Proposed CO₂RR mechanism of PcCu-Cu–O. (Color codes: carbon (gray), nitrogen (blue), oxygen (red), hydrogen (white), copper (orange)).

As mentioned above, SACs can effectively increase atomic utilization and enhance catalytic performance. Therefore, highly dispersed single-atom Cu catalysts with unique electronic structures may break the linear relationship between the intermediates and improve the selectivity of C₂₊ products by regulating the energy barrier of C–C dimerization. For example, by dispersing Cu atoms in a nitrogen-doped conductive carbon matrix, Karapinar et al. [180] synthesized a single-site Cu–N–C catalyst (Cu_0.5NC) via a simple pyrolysis strategy from a dry-phase mixture of ZIF-8 and Cu^II precursor (Fig. 21a). The Cu site with four nitrogen coordinations (CuN₄) enables selective CO₂RR to CH₃CH₂OH with a maximum FE of 55% and exhibits a stable average current density of 16.2 mA cm⁻² under the optimal conditions (potential of −1.2 V vs. RHE, 0.1 M CsHCO₃ electrolyte, CO₂ flow rate of 2.5 mL min⁻¹ and gas-phase recycling set up). It is worth noting that the CuN₄ coordination environment was destroyed during the electrochemical reaction and the oxidation state of Cu was shifted from + 2 to 0, resulting in Cu-Cu coordination, which was confirmed by XAS under operando electrolysis conditions (Fig. 21b). Therefore, the metallic Cu nanoparticles with an average size of 0.47 ± 0.04 nm were likely to be the real active species. Interestingly, the metallic Cu nanoparticles can be oxidized back to + 2 and the Cu–N₄ site can also be restored when the used electrocatalyst is exposed to air or a positive potential is applied (Fig. 21c). This is likely the consequence of the small size of the particles and the strong Cu^II-chelating capacity of the N₄ sites of the material. This work rendered the assignment of single Cu as sole active sites questionable. In addition, Zhao et al. [181] synthesized a single-atom Cu on an N-doped porous carbon catalyst (Cu-SA/NPC, Fig. 21d) by the carbonization of Cu-doped ZIF-8 precursor, and Cu-SA/NPC exhibited considerable CO₂–to–CH₃COCH₃ activity at a low overpotential. EXAFS fitting showed that the Cu species in Cu-SA/NPC are atomically dispersed and coordinated with four pyrrole N atoms, resulting in Cu-pyrrolic-N₄ active sites, which facilitates the C–C coupling and stabilizes the reaction intermediates for acetone production. The FE of CH₃COCH₃ reached a maximum value of 36.7% at the potential of −0.36 V vs. RHE in a CO₂− saturated 0.1 M KHCO₃ solution. DFT calculations revealed that the conversion of CO₂ to CH₃COCH₃ was thermodynamically favorable on the Cu-pyrrolic-N₄ sites of Cu-SA/NPC and the formation of *COOH is the RDS for CH₃COCH₃ production (Fig. 21e). Moreover, the synergy effect of Cu and coordinated pyrrolic N species optimize the adsorption configuration of reaction intermediates on the catalyst surface, and thus promoting the reaction toward the CH₃COCH₃ product (Fig. 21f). Notably, they also demonstrated that both uncoordinated pyrrolic-N₄ and oxidized N in Cu-SA/NPC do not contribute to acetone formation.

Fig. 21

Reproduced with permission from Ref. [180]. d High-angle annular dark-field scanning transmission electron microscopy (HAADF-STEM) images of Cu-SA/NPC. e Free energy diagrams calculated at a potential of -0.36 V for CO₂ reduction to CH₃COCH₃ on Cu-pyridinic-N₄ and Cu-pyrrolic-N₄ sites of Cu-SA/NPC. f Optimized structures of all reaction intermediates involved in the pathways of CO₂ reduction on the Cu-pyrrolic-N₄ site (gray: C of catalyst; black: C of adsorbate; red: O; orange: Cu; blue: N; white: H). Reproduced with permission from Ref. [181]

a Structural and morphological characterization of Cu_0.5NC. b Fourier transform of the experimental EXAFS spectra of Cu_0.5NC under operando electrolysis conditions. c Illustration of the reversible restructuration of metal sites.

4 Summary and Outlook

Transforming waste CO₂ into value-added fuels and chemicals by using renewable energy is an environmentally friendly way to reduce CO₂ emissions and help build a sustainable future. But implementing large-scale CO₂RR technologies still has a long way to go to achieve social and economic benefits. One of the most challenging parts is the design of efficient catalysts that can break the scaling relations and show high selectivity and activity toward a specific reaction pathway. In this review, we first took a detailed look at the CO₂RR mechanisms for different products and provide crucial insights for designing efficient catalysts, with a special focus on the design strategies for efficient and selective MOF-related catalysts such as pristine MOFs and MOFs-derived materials (single-atom catalysts, clusters, metallic NPs or hybrid). Then we discussed the current applications of these catalysts in the CO₂RR. Although tremendous progress has been achieved, there are still many challenges that need to be addressed before MOF-related catalysts can be applied in industrial applications.

First of all, the real active sites for the CO₂RR need to be clarified since the coordination between metal nodes and organic linkers plays a significant role in the formation of MOFs, which greatly affect the structural robustness of MOFs especially when negative potentials are applied. There are considerable MOFs that have been identified as being stable when they are used as electrocatalysts, but the stability of the electrocatalyst is only verified by the crystal structure of the MOFs in the CO₂RR reaction media and/or some post-electrocatalysis analysis (XRD, SEM, XPS, and TEM). To obtain a better understanding of the MOFs’ role in the targeted reactions (real catalysts or precatalysts), it is vital to combine different techniques (especially advanced in situ/operando characterization techniques) to monitor the possible reconstruction of MOFs that may occur during reductive electrolysis. A deep insight into the structure of MOFs during the CO₂RR process would also help to uncover the catalytic mechanism, and thus further guide the rational design of more efficient electrocatalysts.

Moreover, in order to avoid the stability problems, it is also expected that more MOFs with high stability toward water/moisture, acid, and base can be synthesized by incorporating the second type of metal center that is less affected by hydrolysis. The introduction of functional groups such as carboxyl, amino, or phosphonate groups could also improve the stability of MOFs by providing anchoring sites for the metal nodes and enhancing the interaction between the metal nodes and the organic linkers. Furthermore, MOFs are good platforms for modulating the electronic structure of the active sites. We can reasonably adjust the local environment of the active site by molecular or surface engineering, such as the reasonable design of organic ligands and incorporation of coordination heteroatoms, etc. For example, unsaturated sites can be formed by removing part of the linkers. The interaction between under-coordinated sites and reactants or reaction intermediates can tune the local concentration of reactants or intermediates around the active site, thus facilitating the intrinsic activity of CO₂RR. Functional groups on the MOF materials can also play an important role in regulating the selectivity of CO₂RR toward specific value-added products. It could provide additional sites for CO₂ adsorption and facilitate the activation of CO₂. For instance, the existence of N-containing functional groups would activate the protons to acquire COH* or CHO* intermediate [153,154,155,156], which are crucial intermediates for further hydrogenation to produce CH₄. Thus, incorporating amino or nitryl groups onto MOF-related materials may represent a viable strategy for improving the selectivity of carbon dioxide reduction to methane. In addition, as mentioned above, most MOFs are synthesized by harsh solvothermal methods using costly ligands and organic solvents, making the technology unaffordable both economically and environmentally. Therefore, a cost-effective and environment-friendly commercial synthetic approach needs to be explored in future research.

The low conductivity of MOFs is another substantial drawback that greatly hinders the electron transfer in the CO₂RR, resulting in low electrocatalytic activity. Combining MOFs with conductive materials (carbon black, carbon nanotube), introducing guest redox molecules into the frameworks, or choosing the electron-donating ligands are effective strategies to design more conductive MOFs. MOFs with high electrical conductivity can facilitate the transfer of electrons from the electrode to the catalytic sites, leading to enhanced electrocatalytic activity and selectivity toward specific products. Electrical or thermal treatment is also a promising way to convert MOFs into conductive and stable MOF-derived materials (such as metal oxides, alloy, metal on carbon supports, carbon-based SACs, metal-free materials, and nanocomposites). The thermal treatment under an inert atmosphere could retain the porous structure of the original MOF and uniformly dispersed metal sites. Moreover, it allows for the precise control of the material properties, such as the morphology, composition, surface area, and the electronic structure of the metal nodes, which can significantly improve the CO₂RR performance. To enhance atomic utilization efficiency, strategies are needed to anchor metal atoms on the available surface of catalysts to synthesize SACs with a high density of single-atom sites, while avoiding the formation of any clusters or particles encapsulated within the carbon matrix.

As we have discussed above, most studies are devoted to improving the selectivity toward target products, while strategies to improve the efficiency of CO₂ utilization are rarely mentioned. The CO₂ utilization efficiency should be further highlighted in future studies. It seems that the CO₂RR under acidic conditions could be a feasible way for industrial  applications. Moreover, in traditional H-cell or flow-cell reactors, the generated products, either gas or liquid, were mixed with the electrolytes (KHCO₃, KCl or KOH) or H₂ and surplus CO₂. An immediate outlook for large-scale industrial applications points toward the use of systems that directly obtain pure products, avoiding extra separation and concentration processes. For gas products (such as CO, CH₄, C₂H₄) we can combine the advantages of MOF materials, including catalysis and gas separation, to synthesize a functionalized catalyst. Such catalysts can not only effectively convert CO₂ into gas products, but also be used as adsorption and separation materials to purify the gas products, finally obtaining pure gas hydrocarbons. For liquid products, several studies have demonstrated that a modified solid-state electrolyte MEA electrolyzer can generate pure liquid products such as HCOOH and CH₃COOH [9, 186, 187]. However, other pure liquid products, such as alcohols, are rarely reported. It is worth noting that some organic products could damage the solid-state electrolyte and ion-exchange membranes during long-term operation, leading to poor durability. Therefore, optimizing the solid-state electrolyte electrolyzer system for different products is necessary.

Overall, it is not hard to predict a promising foreground of MOF-related materials as CO₂RR electrocatalysts. Although there are still many challenges in optimizing electrolyzer configurations and MOF-related catalysts for achieving higher activity, selectivity, and stability, it is believed that with continuous research efforts and development, MOF-related catalysts will find their unique positions in building a closed-loop anthropogenic carbon cycle in the near future.