Confinement Effects in Well-Defined Metal–Organic Frameworks (MOFs) for Selective CO₂ Hydrogenation: A Review

Abstract

Decarbonization has become an urgent affair to restrain global warming. CO₂ hydrogenation coupled with H₂ derived from water electrolysis is considered a promising route to mitigate the negative impact of carbon emission and also promote the application of hydrogen. It is of great significance to develop catalysts with excellent performance and large-scale implementation. In the past decades, metal–organic frameworks (MOFs) have been widely involved in the rational design of catalysts for CO₂ hydrogenation due to their high surface areas, tunable porosities, well-ordered pore structures, and diversities in metals and functional groups. Confinement effects in MOFs or MOF-derived materials have been reported to promote the stability of CO₂ hydrogenation catalysts, such as molecular complexes of immobilization effect, active sites in size effect, stabilization in the encapsulation effect, and electron transfer and interfacial catalysis in the synergistic effect. This review attempts to summarize the progress of MOF-based CO₂ hydrogenation catalysts up to now, and demonstrate the synthetic strategies, unique features, and enhancement mechanisms compared with traditionally supported catalysts. Great emphasis will be placed on various confinement effects in CO₂ hydrogenation. The challenges and opportunities in precise design, synthesis, and applications of MOF-confined catalysis for CO₂ hydrogenation are also summarized.

1. Introduction

Valorization of CO₂ is a prerequisite for achieving a “carbon-neutral society” as declared in the 2015 Paris Agreement. The chemical conversion of CO₂ into value-added and energy-intensive chemicals is a promising approach to achieve this goal, especially when driven with the green hydrogen from water electrolysis powered by renewable electricity. Accordingly, intensive research efforts have been dedicated to selective CO₂ hydrogenation to CO, formic acid, hydrocarbons, or oxygenates [1,2,3,4,5]. However, these processes suffer from slow kinetics and low selectivity due to the strong strength of the C=O double bond (806 kJ mol⁻¹) [6]. Many advances have been made in various heterogeneous CO₂ hydrogenation catalysts, including oxide (e.g., Cu/ZnO/Al₂O₃ and ZnO-ZrO₂), oxide/zeolite (e.g., Na–Fe₃O₄/HZSM-5 and ZnZrO_x/SAPO-34), supported monometallic or bimetallic catalysts (Ru, Ir, Pt, Rh, PtCo, and PdRh), and immobilized molecular complexes [7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23]. The active sites are carefully designed to process special structural and electronic properties for desired products. For example, the inverse ZrO₂/Cu was designed to maximize the Cu-ZrO₂ interface, which promoted CH₃OH synthesis activity and selectivity of supported Cu [24]. The 8 nmsized Ni supported on CeO₂ exhibited superior CH₄ selectivity over the 4 nm and 2 nm counterparts in CO₂ hydrogenation due to the structure and size dependency [25]. The artificial (C₂₊) hydrocarbons or their oxygenated derivatives could be produced via the rational design of catalysts using zeolite [26,27]. Hence, the rational design of efficient catalysts is essential for achieving high catalytic activity and selectivity during CO₂ hydrogenation [28,29].

Besides the rational design of efficient catalysts, the confinement effects are also widely utilized to stabilize active sites during catalytic CO₂ hydrogenation [21,30,31,32]. The strong interactions between metal nanoparticles (NPs) and supports (e.g., Cu/β-Mo₂C, Ir/CeO₂, Ni/γ-Mo₂N) are confirmed as an effective confinement strategy to prevent the aggregation of NPs [33,34,35]. It modifies the electronic and geometric structures of catalytic active sites. However, the catalyst structure derived from the strong metal–support interaction (SMSI) effect is limited to certain metal and support combinations. Encapsulating the NPs with porous materials or membranes (e.g., zeolite), and thus increasing their resistance to sintering, is another effective and versatile strategy [36,37,38,39]. The fixation of metal NPs with diameters of 0.8–3.6 nm into zeolite crystals (NPs@zeolite) was demonstrated to be sinter-resistant at 600–700 °C and outperformed conventionally supported metal catalysts (NPs/zeolite) during CO₂ hydrogenation [36,37,40]. Unlike the utilization of SMSI or the encapsulation by the zeolite (with fixed tetrahedral Si/Al coordination and pore sizes <1 nm), the ordered three-dimensional metal–organic frameworks (MOFs), with uniform cavities, tunable local environments, tailorable composites, and versatility to introduce functional coordinators/groups, are highly attractive for the rational design of heterogeneous catalysis with confinement effects.

The last decades have witnessed intensive research efforts in the reticular design of various MOFs (>100,000), which have been explored in gas storage and separation, vapor sorption, catalysis, biomedical application, and chemical sensing [41,42]. This review only focuses on CO₂ hydrogenation, where the high surface area and porosity of MOFs are composed of various organic linkers, which can strongly bind metal atoms/clusters to serve as active sites, proffering the ability to adjust the structural and electronic modularity [5,43,44]. These composite materials can combine both the confinement effect of the pores and the SMSI or synergistic effects between active sites and organic chelates/metal-oxo clusters, providing new opportunities in finely tuning the performance of CO₂ hydrogenation [45,46,47]. Tsung and coworkers successfully encapsulated atomic sites or ultra-small NPs in MOFs via a post-synthesis method, which exhibited good activity and selectivity for CO₂ hydrogenation to formic acid, CH₃OH, and C₂H₅OH, respectively [48,49,50,51,52]. Moreover, metal or metal oxide NPs embedded in a porous carbon matrix can be produced by the pyrolysis of MOFs. Since various elements in MOFs are initially well ordered, the new structures can also contain highly dispersed single atoms, metal NPs, alloys, or dopants (e.g., N, P, S, or B) (in the case of dopant-containing organic linkers) [53,54,55]. Compared with conventional catalysts, the metal species are partially encapsulated by the deposited carbon, which exhibits enhanced thermal stability in high-temperature CO₂ hydrogenation reaction conditions. For instance, iron (Fe)- and cobalt (Co)-containing nanomaterials derived from MOFs showed superior selectivity and stability for CO₂ hydrogenation compared with traditional metal catalysts [56,57,58]. These pieces of evidence indicate that the introduction of well-defined MOFs or MOF-derived materials with identical coordinated environments is an effective strategy to unravel the key active sites for CO₂ hydrogenation catalysts, and then enhance the catalytic activity and selectivity.

In recent years, several reviews have introduced advances in the synthesis, characterization, and application of zeolite-fixed metal NPs as metal@zeolite catalysts [8,37,39,59]. For example, Xiao and coworkers reviewed the reported strategies for the construction of metal@zeolite hybrid materials and described how the zeolite micropore and metal nanoparticle synergistically worked to improve the catalytic performance [37,39]. Janiak and coworkers highlighted the confinement of noble nanometals in a zeolite matrix in heterogeneous catalysis [59]. Li et al. summarized the development of MOFs encapsulating active nanoparticles as emerging composites for various catalysis, including thermal catalysis, electrocatalysis, and photocatalysis [60,61]. These works all produced excellent reviews on the preparation of porous materials–confined metals and their applications in heterogeneous catalysis. However, a comprehensive review on confined catalysis in MOFs or MOF-derived materials for CO₂ hydrogenation is still lacking. The current review mainly focuses on the confinement effects in MOFs or MOF-derived materials during CO₂ hydrogenation, as shown in Figure 1. Beginning with the introduction of the preparation and application of MOF-confined catalysts, the confinement effects on CO₂ hydrogenation will be discussed including molecular complexes in the immobilization effect, atomic sites in the size effect, cage stabilization in the encapsulation effect, and adsorption and activation in the synergy effect. Furthermore, the product distribution perturbed by MOF-derived materials during CO₂ hydrogenation will be interpreted. Finally, an outlook on the future development of MOF-based confined catalysis for CO₂ hydrogenation will be given. We hope that the mechanistic insights interpreted by confinement effects in MOFs or MOF-derived catalysts for CO₂ hydrogenation could facilitate the development of clear structure–activity relationships for rational catalyst design.

2. Confined Synthesis and Confinement Effects of MOFs as Host Materials

MOFs represent a frontier in advanced host materials for CO₂ hydrogenation due to their tunable properties including: (1) semi-rigid and highly porous structures, (2) high physical and chemical host stability, (3) strong interaction for the stable dispersion of active species, and (4) solid Lewis or Brønsted acid characteristic. These properties provide an opportunity to synthesize stable and active catalysts using MOFs as host materials in CO₂ hydrogenation. In this section, various synthetic strategies for MOF-confined molecular complexes, single atoms, NPs, or even micro-scale oxides were systematically summarized, including a one-pot synthetic approach (active sites confined in MOF cavities or channels), post-synthetic modification (active sites interacting with functional cavities or channels), and a two-step synthetic approach (active species surrounded by MOF membranes). For further details regarding the synthesis and characterization of various MOFs, the readers are referred to previous reviews and literature cited herein [62,63]. Moreover, the corresponding catalytic performance affected by confinement effects are detailed, indicating that rational designs of the structures of MOFs and active sites are essential in the valorization of CO₂.

2.1. Molecular Complex Encapsulation in MOFs

Compared with heterogeneous CO₂ hydrogenation conducted at higher temperature and pressure conditions (e.g., >200 °C and >3 MPa), homogenous molecular complexes can work at mild conditions (e.g., <150 °C) [64]. The well-defined structures of molecular complexes facilitated the interpretation of reaction mechanisms at the molecular level, which promotes the optimization of catalyst design. A remaining challenge for applying molecular complexes is their immobilization to fabricate stable and recyclable catalysts for practical applications. The porosity and highly ordered structure of MOFs enable their potential as ideal host materials to improve the accessibility of catalytic sites, which also makes molecular complexes recyclable and productive.

The pioneering studies by Yoshio and coworkers demonstrated that homogeneous ruthenium–phosphine catalysts could selectively catalyze CO₂ hydrogenation to formic acid at low temperatures [65]. Moreover, homogeneous molecular complexes containing n-heterocyclic carbenes (NHCs) or phosphorus–nitrogen–phosphorus (PNP) pincer-type ligands could facilitate CO₂ hydrogenation due to the strong electron-donating ability, where the Ru PNP-pincer catalyst could deliver a turnover frequency (TOF) value of 1,100,000 h⁻¹ at 65 °C and 4 MPa (H₂/CO₂ (v/v) = 3:1) [66,67,68,69,70]. However, the challenges including catalysts and products separation and recyclability hindered their industrial applications. Accordingly, the heterogenization of molecular complexes is an effective pathway, which could integrate the distinctive activity of homogeneous catalysts with the advantages of heterogeneous catalysts [71,72]. MOFs are the potential candidates to host molecular complexes. For example, Wu et al. reported various heterogeneous Ru-based molecular complexes immobilized on an azolium-based MOF via post-synthetic metalation which were examined for CO₂ hydrogenation toward formic acid, as shown in Figure 2. The examined Ru molecular complexes included RuCl₃, [RuCp*Cl₂]₂ (Cp* = pentamethylcyclopentadienyl), and [Ru(C₆Me₆)Cl₂]₂ (C₆Me₆ = hexamethylbenzene); the corresponding hybrid catalysts were named Ru_x-NHC-MOF (x = 1, 2, 3). The surface areas and total pore volumes of the three as-prepared catalysts decreased due to the occupancy in the MOF pores (Figure 2b). The as-obtained Ru₃-NHC-MOF catalyst exhibited the highest activity due to the stronger electron-donating ability of the C₆Me₆ ligand in the [Ru(C₆Me₆)Cl₂]₂ complex. Specifically, values of the turnover number (TON) up to 3803 were obtained at 120 °C under a total pressure of 8 MPa (H₂/CO₂ (v/v) = 1) for 2 h in the presence of K₂CO₃ as the base in N, N-Dimethylformamide (DMF) solvent. The strong polarity of the DMF solvent also facilitated the insertion of CO₂ into the Ru−H bond, which is the rate-determining step (RDS) for CO₂ hydrogenation [73].

Host (MOF)-guest (molecular complexes) composites have already demonstrated their advantages in CO₂ hydrogenation. However, the successful synthesis of host–guest composites is limited, and they can only be prepared via de novo synthesis and post-synthetic modification [49,74,75], where the size of molecular complexes should be smaller than the pore of MOFs. Recently, the Tsung group developed a post-synthetic approach to synthesize host–guest composites, where guest molecular complexes were larger (e.g., 3–4 times) than the aperture size of the MOF host but could also be encapsulated into MOFs via aperture-opening events [50]. Guest molecular complexes were encapsulated effectively due to the short-lived “open” states of the pores formed upon linker dissociation, which circumvented the disadvantages of previous strategies including tedious synthetic processes and poor encapsulation efficiency. The aperture-opening process occurred even in the robust MOF, which was significantly affected by the identity of the selected solvent, as shown in Figure 3. A Zr-based MOF consisting of a cubic framework of cationic Zr₆O₄(OH)₄ nodes and 1,4-benzenedicarboxylate linkers (BDC) was selected as the host material, named UiO-66. The (tBuPNP)Ru(CO)HCl (tBuPNP = 2,6-bis((di-tert-butyl-phosphino)methyl)pyridine) was successfully encapsulated into UiO-66 (named [Ru]@UiO-66) by exposing UiO-66 to methanol solvent containing (tBuPNP)Ru(CO)HCl at 55 °C for 5 days in Figure 3. [Ru]@UiO-66 exhibited a comparable TON of ca. 280,000 with that of homogeneous molecular catalysts for CO₂ hydrogenation to formic acid in DMF at 27 °C and 1.5 MPa (H₂/CO₂ (v/v) = 4:1). Interestingly, the recyclability and stability of the encapsulated composites were much better than those of homogeneous ones, where the activity of [Ru]@UiO-66 was maintained after five successive cycles. The homogeneous composite lost more than half of its original activity in the second cycle, which was likely due to the biomolecular deactivation [48].

Apart from formic acid as the main product during CO₂ hydrogenation, CH₃OH also could be produced on MOF-confined molecular complexes. Inspired by biological organisms [76,77], Rayder et al. developed a multi-component catalyst system by integrating [Ru]@MOF and another Ru molecular catalyst to achieve a three-step cascade CO₂ hydrogenation toward CH₃OH. More specifically, the molecular complexes (tBuPNP)Ru(CO)-HCl (Ru-1) were firstly encapsulated within the pores of UiO-66 (named Ru-1@UiO-66) by the aperture-opening strategy, where Ru-1 served as an active site to hydrogenate CO₂ to formic acid [71], and the zirconium oxide nodes with Lewis acidity in UiO-66 served as active sites to catalyze formic acid to a formate ester [78]. Then, another homogenous Ru-based complex (tBuPNN)RuH(CO)Cl (Ru-2) was necessary to catalyze ester hydrogenation to CH₃OH, as illustrated in Figure 4a [79]. This multi-component catalyst system could deliver a TON of 4710 ± 150 in DMF in the presence of ethanol at 70 °C and 4 MPa (H₂/CO₂ (v/v) = 37:3) after 16 h. In contrast, Ru-1, UiO-66, or Ru-2 ([Ru-1, Ru-2]@UiO-66) were found to be inactive, when independently or physically mixed for CO₂ hydrogenation. More interestingly, the co-encapsulation of Ru-1 and Ru-2 in UiO-66 was also prepared successfully, which exhibited higher activity toward CH₃OH than that of mixtures of Ru-1@UiO-66 and Ru-2@UiO-66 as shown in Figure 4b. Even though some activity loss was observed over [Ru-1, Ru-2]@UiO-66, the durability was enhanced and no significant loss was found during the five successive cycles, which was likely due to the isolated sites achieved by the MOF avoiding possible bimolecular decomposition pathways [51]. In this study, UiO-66 not only served as the host material, but also provided a solid Lewis site to transform formic acid to formate ester.

Beyond the primary coordination sphere, second-sphere interactions in host–guest multicomponent catalysts also significantly affect the catalytic activity and selectivity [80,81,82,83,84,85,86]. Tsung and coworkers extensively investigated the second sphere interactions in multicomponent catalyst systems by combining ligand design in the MOFs and the aperture-opening encapsulation strategy for CO₂ hydrogenation to CH₃OH [52]. The structure–activity relationships during CO₂ hydrogenation were efficiently established using various functionalized UiO-66-X hosts (X = -CH₃, -F, -Br, -NO₂, -NH₂, and -NH₃⁺), as shown in Figure 5, where the UiO-66-NH₃⁺ host was found to significantly increase CH₃OH activity compared with that utilized by other UiO-66-Xs as the host. Mechanistic experiments revealed that the NH₃⁺ functionality could serve as Brønsted acid and facilitate the CO₂ hydrogenation to formic acid with autocatalytic features. Importantly, the synergistic effect from the host worked only when the functional group was physically close to the encapsulated molecular complexes. The combination of Ru-1@UiO-66-NH₃⁺ and Ru-2@UiO-66 could achieve the highest TON of 10,900 at 70 °C and 4 MPa (H₂/CO₂ (v:v) = 37:3) in molecular sieves treated with DMF with 10⁻⁵ mmol 2,2,2-trifluoroethanol after 16 h [52].

Despite the above advances, it still remains challenging to realize the precise control of the position of molecular complexes in MOFs [52,87,88,89,90]. To achieve the controllable modulation of the local environment of active sites in MOFs, organic linkers with some functional groups have been introduced to anchor the metal ions [75,91,92], mimicking ligands in organometallic complexes to stabilize homogeneous organometallic complex catalysts in a solid matrix [93]. An et al. reported a novel heterogenized molecular catalyst for CO₂ hydrogenation to formic acid via incorporating Ir^III ions into UiO-type MOFs using post-synthetic metalation followed by NaBhEt₃ treatment, where 2,2′-bipyridine-5,5′-dicarboxylate ligands (bpydc) with or without −OH substitution of the 6-position were synthesized, as shown in Figure 6a. The TON of as-obtained bpydcOH-Ir^III-UiO and bpydc-Ir^III-UiO achieved 6149 and 417 at 85 °C and 0.1 MPa (H₂/CO₂ (v:v) = 1:1) over 15 h, respectively. The isotopic effect and DFT calculations revealed that concerted proton–hydride transfer was the RDS of CO₂ hydrogenation. Then, the electron-donating groups from -OH and the pyridyl nitrogen were ascribed to facilitate the RDS [94]. Tshuma et al. presented the rational design and synthesis of novel isostructural MOFs containing catalytically active Pd(II) sites by using 2,2′-bipyridine-4,4′-dicarboxylate linkers, named (Mg(bpdc)(DMF)₂PdCl₂]n (Pd@Mg:JMS-2) and [Mn(bpdc)(DMF)₂PdCl₂]n (Pd@Mn:JMS-2)), respectively. Figure 6b shows the packing diagram and channel structure of Pd@Mg:JMS-2. For CO₂ hydrogenation to formic acid in ethanol with KOH as the base at 100 °C and 5 MPa (H₂/CO₂ (v:v) = 4:1), Pd@Mn:JMS-2 and Pd@Mg:JMS-2 gave the TOF values of 409 h⁻¹ and 303 h⁻¹, respectively, which were higher than that of the homogeneous complex of 170 h⁻¹ (Pd directly anchored on 2,2′-bipyridine-4,4′-dicarboxylate). The presence of open metal sites and the encapsulation effect are beneficial to concentrate the CO₂ and H₂ gases and avoid deactivation. The reaction mechanism was proposed as follows: (1) H₂ was first activated to form Pd-dihydride intermediate; (2) the Pd-hydride complex attacked CO₂ to generate the formate complex, then desorbed to regenerate the active sites [95].

It was also noticed that different linkers could significantly affect the catalytic performance during CO₂ hydrogenation. Wang et al. successfully immobilized Ru-based molecular complexes into terminal amino functionalized MIL-101(Cr)-NH₂ by post-synthetic modification, where salicylaldehyde (Sal) and 2-diphenylphosphinobenzaldehyde (DPPBde) were utilized as bridging linkers between MOF and Ru (III) ions. The structure of active sites is shown in Figure 6c, where RuCl₃ coordinated with the amino group in the MIL-101 (Cr)-NH₂ (RuCl₃@MOF), the N atom of carbon–nitrogen double bonds and the O atom of Sal (RuCl₃@MIL-101(Cr)-Sal), and the N atom of carbon–nitrogen double bonds and the P atom from linker DPPBde (RuCl₃@MIL-101(Cr)-DPPB), respectively. Among them, RuCl₃@MIL-101(Cr)-DPPB delivered the highest TON of 242 under 120 °C and 6 MPa (H₂/CO₂ (v:v) = 4:1) with triethylamine (Et₃N) as the base in ethanol during CO₂ hydrogenation toward formic acid. Mechanistic insights suggested that the stronger electron-donating ability of P sites in DPPBde promoted the RDS, namely, the insertion of CO₂ into Ru-H [96]. Similarly, ZIF-8 was also utilized as the host, and 3-methyl-1,2,4-triazole with uncoordinated N sites was introduced to anchor Ru by post-synthetic modification, which was examined for CO₂ hydrogenation to formic acid. The donated electron to Ru from uncoordinated N sites in linkers decreased the activation energy of CO₂ activation, as revealed by experiments and theoretical calculations [97].

Overall, the host–guest composites not only maintain the distinctive activity in low-temperature CO₂ hydrogenation and increase the stability and recyclability of molecular complexes but also provide a platform to further enhance their activity and/or selectivity via engineering the primary coordination sphere and second sphere interactions. The one-pot synthesis is relatively simple to encapsulate molecular complexes into an MOF cage or channel, but it is limited by the size of the MOF channel and remains challenging to tune the encapsulation location and uniformity. It is worth noting that guest molecule leaching always hinders this kind of catalyst in large-scale applications. The post-synthetic modification via the aperture-opening strategy could break the limitation of the MOF channel and guest molecule leaching, but the underlying mechanism remains elusive, impeding the rational design. Moreover, these strategies are applicable only after sophisticated synthetic techniques to construct the desired structure, which would limit large-scale applications. Thus, the development of novel and efficient strategies to prepare MOF-based host–guest composites is highly desired.

2.2. Active NPs Confined into MOFs’ Cavities or Channels

Nanometals or oxides have been demonstrated to be effective catalysts for CO₂ hydrogenation [3,98,99]. Nanoscale effects have been well examined in the rational design of high-performance catalysts due to their specific structural and electronic effects (e.g., facet, coordination, or unsaturated sites) [58,100,101]. High surface-area supports (e.g., zeolite, carbon, oxide, and MOF) are needed to stabilize active sites due to their thermodynamic instability at the nano-/atomic scale, especially at high temperature and pressure conditions in the presence of H₂O. Among them, the active sites confined in MOFs provide more opportunities, resulting from their chemically tunable pore surfaces [102], framework flexibility [103], and various modification strategies (e.g., post-synthetic modification and exchangeable ligands) [104,105]. Apart from active NPs directly confined in MOFs, the functionalization of linkers with catalytically active species [93,106] and utilization of the modified MOF nodes as the active sites [107] were also designed. The functionalized linkers with different chemical groups could also enhance reactant adsorption [108,109], promoting CO₂ hydrogenation. In this section, the strategies for MOFs confined in NPs catalysts will be reviewed and the confinement effects will be presented, including size effects, encapsulation effects, and synergistic effects, to provide deeper insights into the advantages and features of confined catalysis.

2.2.1. Atomically Active Sites Confined in MOFs

The electronic and geometric structures of supported NPs can result in dramatic changes in catalytic performance [110,111,112,113,114]. Atomically active sites have been well studied for hydrogenation reactions due to their unique electronic and geometric structures, 100% atom utilization, and unique reaction micro-environments [115,116,117,118]. Recent theoretical and experimental studies have found that atomically active sites exhibit better catalytic activity or selectivity than their nanometer-sized counterparts [118,119,120]. Moreover, the rapid development of related synthetic strategies, characterizations, and theoretical interpretation of atomically active sites provide an ideal platform to understand the structure–performance relationship and reaction mechanisms.

A unique example is that a well-defined MOF containing functionalized linkers with catalytically active species was utilized as a model material to interpret the reaction mechanism during CO₂ hydrogenation to examine the structural requirements for CH₃OH production on ZrZnO_x. The well-defined Zn²⁺–O–Zr⁴⁺ sites were constructed by post-synthetic treatment of Zr₆(μ₃-O)₄(μ₃-OH)₄ nodes of MOF-808 with ZnEt₂ and then followed by a mild thermal treatment to remove capping formats on Zr SBUs. The as-prepared composites delivered a high CH₃OH yield of 5.72 mmol_CH3OH g_Zn⁻¹ h⁻¹, >99% CH₃OH selectivity, and excellent stability (e.g., >100 h) in CO₂ hydrogenation at 250 °C. A synergistic effect between open Zr⁴⁺ sites and Zn²⁺ centers was proven to be indispensable to the generation of CH₃OH. Mechanistic investigations disclosed that (1) Zn²⁺ was responsible for H₂ activation by TPD of H₂ and H/D exchange tests; (2) CO₂ was adsorbed and activated on open Zr⁴⁺ sites in the nearby Zr⁴⁺–O–Zn²⁺ interface, as shown in Figure 7. In situ DRIFT and DFT calculations further showed that the activated CO₂ reacted with the heterolytic splitting H⁻ to form *HCOO intermediate on Zn²⁺ sites, followed by hydrogenation to dioxomethylene (*H₂COO), formaldehyde (*H₂CO), methoxy (*H₃CO), and finally methanol [121]. These findings demonstrated that Zn was active for H₂ activation; meanwhile, both Zn and Zr participated in CO₂ activation, in agreement with previous studies [10,31,122]. Moreover, this study revealed the precise chemical structures of Zn and Zr in their active forms during CO₂ hydrogenation.

Multifunctional sites in MOFs were also facilitated, enabling cascade CO₂ hydrogenation to multi-carbon products (e.g., C₂H₅OH and C₂H₄). An et al. designed well-defined Cu active sites on deprotonated [Zr₁₂O₈(μ₃-O)₈(μ₂-O)₆(carboxylate)₁₈]₁₄⁻ SBUs in the Zr₁₂-MOF for selective CO₂ hydrogenation to ethanol [123]. Regarding the synthesis, Zr₁₂-SBUs with an average thickness of 50 nm were prepared and the SBUs were deprotonated by LiCH₂SiMe₃, followed by metalation with [Cu^I(CH₃CN)₄](BF₄), which was confirmed by infrared (IR) spectrum. Two types of Cu species were formed: 1) four-coordinated tetrahedral Cu^I centers([(μ₄-O⁻)(μ₂OCO carboxylate)₂ Cu^I(THF)]) (site 1, 3) and [(μ₄-O−Li⁺)(μ₃-O−)(μ₄-O−)Cu^I(THF)] (site 2), as shown in Figure 8. Moreover, different alkali cations were also introduced into the MOF structure to modify the local environment of Cu, which was confirmed by XAS analysis. Among them, the Cs⁺-modified MOF catalyst achieved a TON of 4080 and >99% selectivity of ethanol in supercritical CO₂ (30 MPa CO₂ /5 MPa H₂) at 85 °C. In contrast, the randomly supporting active Cu sites on ZrO₂ preferred CH₃OH formation under identical reaction conditions, as the aggregation occurred. The experimental results and DFT calculations disclosed that the H₂ was activated on (Cu^I)₂ sites by a bimetallic oxidative addition process to form (Cu²⁺–H⁻)₂ and then reacted with CO₂ to produce methanol and *CHO species, where electron-rich Cu^I induced by the effective electron-donating effect of Cs could facilitate these steps. The C–C bond was formed to get the CH₃CHO intermediate via the nucleophilic attack on methanol by the CHO intermediate on bimetallic Cu sites; ethanol was formed by further hydrogenation. Additionally, alkali metals could stabilize *CHO intermediate to direct the C₂H₅OH formation. Overall, the synergistic effect between the cooperative nature of the bimetallic Cu^I₂ centers and the electron-rich environment for the Cu center was interpreted to contribute to the higher activity for C₂H₅OH formation.

To better understand the multi-functionality of MOFs, MOFs involved in catalytic cycles were also investigated. For example, Zeng et al. developed a cascade catalyst, Cu^I centers supported on MIL-125 ([Ti₈(μ₂-O)₈(μ₂-OH)₄(BDC)₆]), where the Cu^I catalyzed CO₂ hydrogenation to C₂H₅OH, but C₂H₅OH dehydration on Ti₂-μ₂-O–M⁺ (M⁺ = H⁺, Li⁺) sites occurred. The post-synthetic metalation of SBUs was utilized for catalyst preparation and MIL-125-NH₂-Cu^I-4 was obtained, as shown in Figure 9a [124,125]. The as-obtained composites exhibited >90% selectivity toward C₂H₄ during CO₂ hydrogenation with a reaction rate of up to 514 μmol g_cat^–1 h^–1 at 100 °C and 5 MPa (H₂/CO₂ = 3:1) in THF. The choice of MIL-125 as the support was critical to the tandem CO₂-to-C₂H₄ transformation: (1) the Ti₈(μ₂-O)₈(μ₂-OH)₄ SBU rather than other SBUs stabilized Cu^I to avoid Cu⁰ formation; (2) the short Cu^I–Cu^I distance in Ti₈(μ₂-O)₈(μ₂-O-CuI)₄ enhanced the synergistic effect on C₂ formation; (3) the strong Lewis acid Ti^IV catalyzed C₂H₅OH dehydration at low temperatures (∼100 °C); and (4) C₂H₅OH was easily adsorbed on the Ti₂(μ₂-O–M⁺) site in Ti₈(μ₂-O)₈(μ₂-O–CuI)₄, which further induced β-elimination to produce C₂H₄ (Figure 9b,c). This work highlighted new opportunities in using MOFs as novel supports for CO₂ hydrogenation to C₂H₄ [126]. Moreover, this study indicated the pore-dependent activity and selectivity during CO₂ hydrogenation in tunable and ordered structure MOFs.

Apart from the transition metals (Zn or Cu), noble metals (e.g., Pt) are also able to encapsulate into MOFs, where the atomically dispersed Pt₁ atom was coordinated by four O atoms in MIL-101, as shown in Figure 10a. The active sites created by metal–ligand cooperativity led to the dissociation of H₂ to form hydroxyl groups, attacking CO₂ to produce the key intermediates of HCOO*, evidenced by DFT calculations and operando spectroscopies (Figure 10b). In contrast, the Pt cluster (Pt_n) encapsulated in MIL-101 preferred hydride formation and then hydrogenated CO₂ into COOH* as key intermediates. The divergence in reaction paths resulted in different catalytic selectivity between Pt₁@MIL-101(90%) and Pt_n@MIL-101 (58%), where the *HCOO intermediates went through stepwise hydrogenation to produce CH₃OH, while *COOH was hydrogenated into various products including HCOOH, CH₃OH, CO, and CH₄ (Figure 10c,d) [127].

Accordingly, atomically dispersed active sites in MOFs are an ideal target in catalyst design, especially for noble metals, which not only increase the exposed active sites per molar amount of metals but also create new prospects for the controllable surface free energy of the metal atoms. The synergistic effect in MOFs is of great importance for the rational design of catalysts to achieve excellent performance for desired products. The design of atomically distributed active metals in MOFs with reasonable catalytic performance has been demonstrated in the lab; however, no related application was found on a large scale, hence, more efforts should be directed to developing more efficient strategies for confined catalysts using MOFs as hosts.

2.2.2. Active NPs Confined in Monocrystal MOF

Heterogeneous catalysts are usually prepared by loading active metal sites onto a support, where both the structure of active metal sites (e.g., facet, morphology, and electronic state) and properties of supports (physical properties and chemical properties) will affect the catalytic performance [116,128,129,130,131,132,133]. For example, the different facets of active NPs result in different unsaturated metal surface atoms, which can change the catalytic processes; the supports not only affect the dispersion of active NPs but also tune the electronic and/or geometrical structure of active NPs via SMSI [132,134,135,136]. Actually, the sintering under harsh reaction conditions due to Ostwald ripening or particle migration and coalescence makes it challenging for industrial application [137,138]. An efficient and sustainable strategy, encapsulation has been well investigated to increase their long-term stability [139].

Rungtaweevoranit et al. successfully synthesized single nanocrystalline UiO-66 encapsulated 18 nm Cu via the bottom-up method using Zr(OPrⁿ)₄ as a precursor (Figure 11a,b). The as-prepared Cu@UiO-66 was demonstrated to be active and selective for CO₂ hydrogenation toward CH₃OH, and the TOF was 3.7 × 10⁻³ s⁻¹ at 175 °C and 1 MPa (H₂/CO₂ = 3/1). The formation rate of CH₃OH on Cu@UiO-66 was almost two and nine times higher than those of Cu supported on UiO-66 and commercial Cu/Zn/Al₂O₃, respectively. Moreover, 100% selectivity was observed on Cu@UiO-66. The strong interaction between Cu NPs and Zr oxide SBUs of the MOF was interpreted as leading to the higher catalytic performance [140]. Kobayashi et al. investigated the coated effects of MOFs including ZIF-8, MIL-100, and functionalized UiO-66 on CO₂ hydrogenation for CH₃OH formation, as shown in Figure 11c. The charge transfer was observed between Cu and MOF substrates, and UiO-66 was found to be the most active support. Interestingly, the replacement of Zr⁴⁺ with Hf⁴⁺ or functional groups (e.g., from -NH₂ to -COOH) in UiO-66 enhanced the rate of CH₃OH production [141]. Truhlar and coworkers provided deeper insight into the molecular interactions between Cu and Zr in UiO-66. The experimental and theoretical results disclosed that the direct interaction between Cu NPs and ZrO₂ nodes was essential for CO₂ hydrogenation to CH₃OH (Figure 11d) [142]. The quantitative effects of the missing-linker defects on H₂ and CO₂ activation were carried out using a detailed quantum mechanical study. It was found that the optimum number of missing linkers (ca. 5–7 per unit cell) balanced the steric effects and the strong CO₂ binding on the ZrO₂ node delivered a maximum TOF of CH₃OH formation [143].

MOFs further provide a specific platform for NPs including protective microenvironments and functional sites. Zheng et al. reported that the core–shell monodispersed nanosphere Au@Pd could be encapsulated by the UiO-66 membrane to protect its morphology and then the microporous characteristic of UiO-66 preferred the adsorption of Pt NPs on its surface to impart its functionality. This kind of assembly enhanced the interaction between NPs and UiO-66, allowing for the spatial distribution of NPs in MOFs. Furthermore, the catalysts for the RWGS in a fixed-bed flow reactor exhibited high catalytic activity and CO selectivity [144].

MOFs can also serve as novel supports for mixed oxide catalysts via taking advantage of tunable and specific strong metal–support interactions. Lin and coworkers demonstrated a strategy to utilize the UiO-bpy MOF constructed by bpy and Zr₆(μ₃-O)₄(μ₃-OH)₄ to encapsulate ultrasmall Cu/ZnO_x nanoparticles, which selectively catalyzed high-rate CO₂ hydrogenation to CH₃OH with the yield up to 2.59 g_CH3OH kg_Cu^–1 h^–1, 100% CH₃OH selectivity, and high stability over 100 h (Figure 12a). The UiO-bpy MOF was chosen because of its exceptional hydrothermal stability. Post-synthetic metalation was developed to introduce Cu²⁺ and Zn²⁺ ions sequentially, which were coordinated to the bpy and μ₃-OH sites in the MOF, respectively. Finally, ultra-small Cu/ZnO_x NPs generated under reaction conditions were encapsulated in the tetrahedral and octahedral cages confined by the ligands as the active sites. The as-prepared composite exhibited much better CH₃OH selectivity and stability than those of commercial Cu/ZnO/Al₂O₃ (Figure 12b), likely resulting from the suppression of Cu NPs’ agglomeration and phase separation between Cu and ZnO_x. Regarding mechanistic understanding, the authors proposed that CO₂ was adsorbed on unsaturated ZrO_x and ZnO_x sites to form carbonates and bicarbonates; the homolytic dissociation of H on Cu spill over to the Zr sites on the SBUs and defect sites of ZnO_x (Figure 12c). Overall, the synergistic combination of hydrogen activation sites on Cu and CO₂ activation on ZnO_x and Zr SBUs contributes to superior performance of the CuZn@UiO-bpy catalyst [145]. Similarly, Yu et al. also reported that ultra-small bimetallic Cu/ZnO_x NPs was encapsulated in UiO-66 by the deposition–precipitation method; both metal loading and the Cu/Zn mole ratio could regulate the Cu-Zn interaction. The highest CH₃OH yield of 9.1% was obtained over the Cu-Zn@UiO-66 catalyst (Cu-Zn loading: 35 wt.% and Cu/Zn mole ratio: 2.5) [146].

2.2.3. Active Species Confined in MOF Membrane

The active species of CO₂ hydrogenation are not only encapsulated in monocrystal MOFs, but also confined by the MOF membrane. The combination of guest species and different types of MOF active sites (e.g., metal nodes, functional organic linkers) makes MOFs promising multifunctional materials for synergistic catalysis [47]. For example, the ZnZrO_x/SAPO-34 composite was able to selectively catalyze CO₂ hydrogenation to paraffin. Jiang et al. shifted the product distributions from paraffins to olefins via the introduction of the UiO-66 membrane as a coating layer on the SAPO-34 surface. It was demonstrated that the stable UiO-n membrane passivated the excessive Brønsted acid sites of SAPO-34, suppressing the hydrogenation of olefins to paraffins. Meanwhile, the uniform UiO-66 membrane had no effect on the diffusion step during CO₂ hydrogenation, where the selectivity of the C₂−C₄ olefins was increased from 57% on ZnZrO_x/SAPO-34 to 80% on that with a UiO-66 membrane under 380 °C, 3 MPa, and GHSV of 6000 h^–1 (CO₂/H₂/Ar = 24:72:4), as shown in Figure 13a [147]. Compared with the existing strategies of adjusting the acidity of zeolites by changing the structure of the zeolite framework (e.g., alkali treatment or increasing calcination temperature), the method of membranization by coating a layer of a functional MOF membrane on the zeolite surface did not affect the framework structure. Moreover, a UiO-66 membrane was epitaxially grown on the surface of nano-SAPO-34 clusters as a support for dispersed Pt for CO₂ hydrogenation. The MOF membrane not only overcomes the noble metal agglomeration and enhances the molecular sieve synergistic catalysis, but also provides sites for CO₂ adsorption, significantly enhancing the selectivity and yield [148]. Similarly, Pan et al. developed a strategy for the epitaxial growth of the UiO-66-NH₂ shell to obtain an MOF-membranized bicomponent core−shell catalyst, HZSM-5@UiO-66-NH₂/Pd. By coating UiO-66-NH₂ on the surface of HZSM-5, the highly dispersed and small Pd NPs can be anchored on the HZSM-5@UiO-66-NH₂ supports owing to the affinity between UiO-66-NH₂ and Pd NPs. The as-prepared HZSM-5@UiO-66-NH₂/Pd exhibited high CO selectivity (92.2%) with 17.1% CO₂ conversion in CO₂ hydrogenation toward CO at 320 °C [149].

Apart from Cu/ZnO catalysts, Pd/ZnO catalysts are also active for CH₃OH production in CO₂ hydrogenation [150,151]. Li et al. designed core–shell ZnO@ZIF-8 nanorods with an ultrathin ZIF-8 membrane overcoating to confine Pd NPs at the ZnO/ZIF-8 interface for CO₂ hydrogenation, as shown in Figure 13b. Surface oxygen defects of ZnO were generated during the formation of ZIF-8 shells, which facilitated the chemisorption of CO₂ by promoting electron transfer to CO₂. Sub-nano PdZn alloy was confined at the ZnO/ZIF-8 interface after H₂ reduction, giving a high CH₃OH selectivity and long-term stability. The thickness of ZIF-8 significantly dominated the activity of CH₃OH formation, which was related to the number of PdZn alloy sites at the interface with O-defective ZnO. The optimized catalyst achieved 66–78% of CH₃OH selectivity and 12.1–19.8 g g_Pd⁻¹ h⁻¹ of methanol yield at 250–290 °C and 4.5 MPa, showing much higher values than those of reported Pd-based catalysts under comparable conditions [1].

With the rapid development of synthetic chemistry, various MOF-confined catalysts with different structures and interfaces were successfully constructed via one-pot synthesis and post-synthetic modification, which impressively exhibited unique properties over traditionally supported catalysts. The catalytic performance of various catalysts in CO₂ hydrogenation is summarized in

Table 1. Summary of CO₂ hydrogenation over various catalysts.

Catalysts	Reaction Conditions			Performance		Ref.
	P (MPa)	T (°C)	CO2/H2	TON	Product
Ru3-NHC-MOF	8	120	1	3803	HCOOH	[65]
[Ru]@UiO-661	1.5	27	4	280,000	HCOOH	[48]
Ru-1@UiO-66+Ru-2	4	70	12	6600	CH3OH	[79]
Ru-2@UiO-66+Ru-1				5700
Ru-1@UiO-66+Ru-2@UiO-66				3500
[Ru-1, Ru-2]@UiO-66				4300
Ru-1@UiO-66-NH3+	4	70	12	10,900	CH3OH	[52]
Ru-1@UiO-66				6900
Ru-1@UiO-67				7100
Ru-1@UiO-66-NH3+				8300
bpydcOH-IrIII-UiO	0.1	85	1	6149	CH3OH	[87]
bpydc-IrIII-UiO				417
Pd@Mn:JMS-2	5	100	4	409	C2H5OH	[88]
RuCl3@MIL-101(Cr)-DPPB	6	120	4	242	HCOOH	[89]
Zn-MOF-88	4	250	3	5.9 a	CH3OH	[121]
[CuI]	35	85	0.2	4080	C2H5OH	[123]
Cu/MIL-125	5	100	3	514 a	C2H4	[126]
Pt1@MIL-101	3.2	150	3	0.6 a	CH3OH (90%)	[127]
CuNP@UiO-66	1	175	3	0.004 b	CH3OH	[140]
CuZn@UiO-bpy	4	250	3	2.6 c	CH3OH	[145]

Note: [Ru]: (tBuPNP)Ru(CO)HCl; Ru-1: (tBuPNP)Ru(CO)-HCl; Ru-2: (tBuPNN)RuH(CO)Cl; bpydcOH: 2,2′-bipyridine-5,5′-dicarboxylate ligands (bpydc) with −OH substitution on the 6-position; Pd@Mn:JMS-2: [Mn(bpdc)(DMF)2PdCl2]_n and 2,2′-bipyridine-4,4′-dicarboxylate (bpdc); DPPB: 2-diphenylphosphinobenzaldehyde (DPPBde); a: mgCH3OH gZn⁻¹ h⁻¹; [Cu^I]: [(μ₄-O−Li⁺)(μ₃-O−)(μ₄-O−)Cu^I(THF)]; b: s⁻¹; c: g_MeOH kgCu^–1 h^–1.

. However, they still suffered from big challenges, especially MOF stability in fixed-bed type reactors. Most MOFs were easily decomposed in high-polar solvents or at temperatures higher than 300 °C, especially in the presence of water [152,153,154].

3. Confined Synthesis and Confinement Effects for MOF-Derived Materials

MOFs have been explored as potential templates or precursors to prepare highly dispersed NPs via thermal decomposition due to their unique structure and textural properties [57,155,156,157]. Carbon-encapsulated NPs derived from MOFs as emerging catalysts received tremendous research efforts [158,159,160]. The framework of MOFs could protect the active site aggregation to generate highly dispersed NPs during high-temperature pyrolysis. Moreover, the confinement effects induced by the encapsulation carbon layer provide robust microenvironments for CO₂ hydrogenation, which could tune the binding energy of key intermediates and affect the adsorption of CO₂ and transportation of products. In this section, we will simply highlight some important merits of the encapsulated metal nanoparticles for CO₂ hydrogenation, including modification of local reaction environment, electronic transfer, and interfacial catalysis to enhance catalytic performance.

Carbon-encapsulated highly dispersed NPs derived from MOFs were widely investigated for selective CO₂ hydrogenation. Lu et al. reported a one-step pyrolysis of MOFs containing both an N-free and an N-containing linker to obtain highly dispersed cobalt NPs embedded in a carbon matrix for selective CO₂ hydrogenation to CO (Figure 14a,b) [161]. The presence of N in the MOF precursor could not only decrease the size of generated Co [162,163], but also modulate local reaction conditions because of the formation of a Mott–Schottky interface, which lowered the energy barrier for the formation of formate intermediates [164,165,166]. Similar enhancement was also observed in Fischer–Tropsch synthesis [167]. Moreover, Ni-based MOFs could generate hierarchical Ni@C hollow spheres composed of highly dispersed Ni NPs confined in carbon shells, which was active and selective for CO₂ hydrogenation to CH₄ (Figure 14c,d). The high surface area and highly dispersed active sites were ascribed to promote CO₂ adsorption and redox catalysis [168]. Apart from C₁ products, highly dispersed Fe-based catalysts derived from Fe-MIL-88B could catalyze CO₂ actively and selectively to valuable hydrocarbons (C₂₊) [169,170,171,172,173,174]. Ramirez et al. examined the promotion effects of various elements (Cu, Mo, Li, Na, K, Mg, Ca, Zn, Ni, Co, Mn, Fe, Pt, and Rh) for the resulting Fe@C-based composites during CO₂ hydrogenation, where the Basolite F300(Fe) was used as a template and incipient wetness impregnation was utilized to add the various promotors (Figure 14e) [56]. Among them, only K could increase the activity and enhance C₂–C₆ olefin selectivity from 0.7% to 36%.

Another property of the confinement effects is the strong interaction between the metal NPs and the encapsulating materials, which facilitates the electron transfer under some conditions and maximizes the interfacial area due to their close contact [175,176,177,178,179,180,181,182,183,184,185,186,187,188,189,190] Pustovarenko et al. demonstrated the MOF-mediated route for the preparation of highly efficient Co₃O₄-supported In₂O₃ catalyst via stepwise pyrolytic–oxidative decomposition, as shown in Figure 15a, which selectively catalyzed selective CO₂ hydrogenation to CH₃OH [191]. The stable Cu-ZnO interfacial sites were also constructed using a novel bimetallic CuZn-BTC MOF for CO₂ hydrogenation to CH₃OH [192]. Additionally, Cu-based catalysts were revealed to be structure sensitive, including facet, defect, particle size, and interface [22,193,194,195,196]. Han et al. prepared hollow Cu@ZrO₂ catalysts through pyrolysis of Cu-loaded Zr-MOF, where low-temperature pyrolysis (e.g., 300 °C) produced highly dispersed Cu nanoparticles with balanced Cu⁰/Cu⁺ sites, larger amounts of surface basic sites, and an abundant Cu-ZrO₂ interface in the hollow structure, as shown in Figure 15b, which delivered the best performance for CO₂ hydrogenation to CH₃OH with 5% CO₂ conversion and 85% CH₃OH selectivity at a reaction temperature of 220 °C [197]. Similarly, Yu et al. prepared highly dispersed Cu NPs on ZrO₂ via the calcination and reduction of ZrO₂@HKUST-1 (Figure 15c). Due to the confinement of MOFs, highly dispersed Cu on ZrO₂ was generated, which increased the efficient active sites and produced more interface between the Cu and ZrO₂. The as-prepared catalysts exhibited 5.2 times higher CH₃OH yield than that of those catalysts prepared by the conventional impregnation method [198]. Similar enhancement was also demonstrated in the PdZn alloy and In₂O₃/Pd during CO₂ hydrogenation [199,200].

Overall, the MOF-derived materials exhibited higher stability than MOFs themselves in CO₂ hydrogenation. The electronic effects of SMSI and interfacial catalysis could dramatically enhance catalytic performance. However, the reaction mechanism at the molecular level remains elusive; more advanced operando characterizations in conjunction with theoretical calculation are needed in further work to decipher the mechanism in greater detail. It is worth noting that the encapsulating layer (especially the carbon layer) seems to be consumed on stream to decrease the catalytic performance because of the involvement in CO₂ hydrogenation.

4. Conclusions and Outlook

Different from zeolite with fixed tetrahedral Si/Al coordination and pore sizes (<1 nm) [201], the uniform and tunable cavities and tailorable composite make MOFs especially attractive for heterogeneous catalysis. The confinement effects in MOFs or MOF-derived materials involving immobilization, size of active sites, encapsulation, and synergy effect significantly perturb the catalytic performance of CO₂ hydrogenation. In this review, the microstructural optimization and engineering of MOF-based catalysts for CO₂ hydrogenation were systematically summarized. The MOFs or MOF-derived catalysts can be simply divided into three categories: (1) molecular complexes, atomic or nanosized active species are confined into cavities or channels of MOFs via one-pot synthesis or post-synthetic modification; (2) active species are surrounded by MOFs or MOF membranes by epitaxial growth; (3) active sites are embedded in carbon or oxide layers derived from MOFs. The most important merits of the heterogenized molecular complexes are their dramatically enhanced recyclability and stability. The advantages of the confined NPs are the cage or channel stabilization and modification of the local reaction microenvironment. The composites derived from MOFs exhibited strong electronic transfer effects and increased interface catalysis between active sites and the encapsulating layers.

Despite all the progress achieved by researchers, many problems remain unsolved. For example, the precise synthesis with desired spatial distribution, loading, and size of active species in MOFs represent the biggest challenges in this field; the hydrothermal stability of host MOFs under reaction conditions is still limited; most products are CO, CH₄, and CH₃OH, but the high-value-added chemicals or fuels produced from CO₂ hydrogenation are limited; and the enhancement interpretation of confinement effects at the molecular level is less than well understood.

In the future, extensive efforts are required to resolve the existing challenges. The detailed directions are as follows:

• The development of sophisticated organic synthesis to design more linkers with desired functional groups is highly needed, which will make the precise tailoring of active sites highly possible. For example, some functional groups (e.g., carboxyl, sulfonic acid, or amino groups) are effective to stabilize the active sites; therefore, their controllable synthesis with desired positions and contents are essential to obtain rationally designed catalysts.
• More fundamental interpretations for MOF nucleation, growth, decomposition, and collapsing can provide a strong foundation for more stable MOFs preparations. Since the generated water in CO₂ hydrogenation is unavoidable, water management near the active sites can change the stability of MOFs and tune product distributions.
• To date, only imidazole or carboxylate-based MOFs were widely investigated for CO₂ hydrogenation; more linkers and metal nodes (e.g., Ru, Ti, Mo, and Mn) are needed to obtain a higher diversity of applicable MOFs. Apart from Cu-, Zr-, or Zn-based composites, some new catalyst systems (e.g., Ru, Mn, and Rh) are more interesting for high-value-added chemicals or fuel formation.
• Advanced characterization techniques are required to reveal the structure of the active sites; the reaction mechanisms at the molecular scale are helpful to increase the selectivity of target products.