Covalent organic frameworks (COFs) are a class of crystalline porous materials that allow the atomically precise integration of pure organic units to create predesigned skeletons and nanopores based on reticular chemistry [1-8]. Different from metal-organic frameworks (MOFs), which could be obtained via the coordinative assembly between metal ions and organic ligands as the building units [9], COFs are linked to form through irreversible covalent bonds, which can be formed, broken, and reformed, allowing reaction systems with "error checking" and "proof-reading" characteristics to result in the formation of the most thermodynamically stable structures [2]. In 2005, Yaghi and co-workers demonstrated the first successful examples of COF synthesis [10]. Since this landmark paper, COFs have attracted extensive interest, and have become one of the fastest-growing fields in both materials science and chemistry during the past decade. Depending on the dimension of building units, COFs can be categorized into two- (2D) and three-dimensional (3D) structures. 2D COFs are easier to be prepared than 3D COFs because of their structural simplicity. The covalently bonded planar sheets are stacked together through π-π interactions, leading to the formation of eclipsed or staggered one-dimensional (1D) columnar arrays in 2D COFs. In contrast, 3D COFs are generally constructed based on a T_d-structure building block. 3D COFs possess not only extremely high surface areas, but also extraordinarily low densities, and, furthermore, the entire material is accessible for adsorbed molecules to "see" all the atoms of the framework. Both 2D and 3D COFs have an open network structure, which results in accessible channels or nanopores with uniform sizes ranging from angstroms to nanometers. Their unique structure is therefore attractive for many applications, including gas storage and separation [11-15], drug delivery [16-18], and energy conversion and storage [19-22]. In addition, through decoration with functional moieties, COFs can also be endowed with unique optical [23-27] and optoelectronic [28-30] properties. The structural and functional versatility and diversity of the constituent organic units create immense possibilities in COF materials.

The high porosity, functional versatility, easy modification, and outstanding stability of COFs make them ideal materials for catalysis applications. In particular, the assembly of COFs can be achieved by various designable and tailorable organic building blocks, which gives innumerous structural topologies and diversified porosities. Therefore, in contrast to traditional activated carbon and zeolites, the assembly approach endows COFs with unlimited structural models from "designable" crystal engineering. Meanwhile, using COFs as heterogeneous catalysts could have the following advantages, including (1) enhanced catalyst reactivity due to the spatial separation of multiple catalytic sites in the framework, endowing COFs with cooperative catalysis character; (2) permeable channels, allowing the facile access of substrates to the catalytically active sites within the COF pores, endowing the chemical reactions with shape-, size-, chemo-, or enantio-selectivity; (3) due to the flexibility and dynamics of the framework in the solid state, external stimuli, with a either physical, chemical, or environmental method, could bring new COF applications, like transition-state recognition. Furthermore, similar to other heterogeneous catalysts, the ability to separate for cycle utilization is highly attractive in large-scale reactions, especially when separation and waste disposal are costly.

Using COFs in catalysis is a rapidly developing interdisciplinary research field. A recent review gives a short description of the preparation and functionalization of COF materials, as well as of their catalysis applications [31]. In the present paper, we present a thorough review of the recent advances in catalytic COF materials. For the sake of discussion, COFs are categorized based on the type of catalytic sites, including single functional active sites, bifunctional active sites, and metal nanoparticles (NPs) embedded in pores. We mainly summarize the deliberate or incidental preparation strategies, the stability, the heterogeneity, and the shape/size selectivity of COF catalysts. In addition, using COFs as photocatalysts and electrocatalysts is introduced.

In contrast to inorganic porous materials, COFs can be easily decorated with molecular catalysts that may acquire activities and selectivities comparable to their homogeneous analogues. Similar to MOFs, many COFs are synthesized in various fashions that easily enable straightforward incorporation of functional groups and, therefore, open a potential experimental environment for using COFs as catalysts. In general, there are two strategies that can be used for constructing catalytic COFs [31]. One methodology involves the post-synthetic integration of catalytic sites into a COF skeleton. This post-synthetic modification approach can reduce the influence of bulky catalytic sites on COF crystallinity, and the undesired effect of harsh solvothermal conditions on the catalytic sites. This is, by far, the method most used for the modification of COFs. Another is the bottom-up strategy that uses building blocks possessing catalytic sites to directly construct COFs. However, this strategy generally requires tedious solvothermal synthesis. In particular, if a bulky catalytic site is attached to building blocks, it will be difficult to obtain crystalline COFs. So far, the construction of covalently linked, yet highly active, COF catalysts still remains a synthesis challenge in the field.

A pioneer work using COFs as catalysts was reported by Wang and co-workers [32]. They constructed an imine-based COF material, COF-LZU1, and its 2D layered-sheet structure provides an exactly suitable space to coordinate with Pd(OAc)₂ because the distance (~3.7 Å) of eclipsed nitrogen atoms of imine in adjacent layers falls into the ideal requirement for strong coordination of Pd(OAc)₂ (Fig. 1). Metal Pd²⁺ catalytic sites can be successfully introduced via a simple post-treatment, and the as-synthesized Pd/COF-LZU1 demonstrated excellent catalytic activity for the Suzuki-Miyaura coupling reaction. The superior utility of the hybrid in catalysis was further elucidated by the broad scope of the reactants and the excellent yields (96%-98%) of the products. High stability and easy recyclability were also exhibited for the Pd/COF-LZU1 catalyst.

In addition to Pd(OAc)₂, Chen et al. found that copper ions (Cu²⁺) can also be efficiently incorporated into the layer spaces of 2D COFs via the coordination interaction with hydroxyl groups combining with imine linkers [33]. The obtained copper-containing COFs (Cu-COFHX and Cu-COFDMF) were employed in the selective oxidation of styrene to benzaldehyde, and exhibited excellent catalytic performance and recyclability. We [34] and others [35] have reported that Pd²⁺ can be coordinated with both imine groups of layers and building units (bipyridyl or porphyrin group) in one COF material. The Pd hybrid COF exhibits good catalytic activity in the Heck reaction and the Suzuki coupling reaction. In a similar way, a molybdenum-doped COF (Mo-COF) catalyst was also obtained from a two-step bottom-up approach. The resultant catalysts are active for selective oxidation reactions, such as epoxidation of alkenes (cyclohexene, cyclooctene, styrene, and isopropenylbenzene) [36]. In addition, Mn²⁺ and Co²⁺ have been successfully coordinated within the porphyrinic framework of CPF-2 [37]. Although X-ray diffraction (XRD) measurements show that CPF-2 is a crystalline material, the crystal structure is not successfully resolved. The Co(Ⅱ)-doped material Co-CPF-2 shows high efficiency in heterogeneous catalysis of the aerobic epoxidation of olefins under mild conditions. A high conversion rate (> 99%) and good epoxide selectivity (93%), with a turnover number (TON) of 29215 and a 3434 h⁻¹ turnover frequency (TOF), have been obtained for the aerobic epoxidation of styrene.

Jiang and co-workers carried out a series of experiments that introduced organocatalytic sites into COFs via a pore surface engineering strategy. This strategy allows molecular design of COF skeletons, and control over the density and composition of the functional groups on the pore walls, thus offering a general principle for designing catalytic COFs. In 2014, they first designed and prepared a series of [HC≡C]x-H₂P-COFs with different alkynyl densities [38]. Catalytic active pyrrolidine units were then integrated into the pore walls of the [HC≡C]X-H₂P-COFs through click reaction. It has been found that the COF catalyst combines several striking features, such as enhanced activity, broad applicability, good recyclability, and high capability, to perform catalytic transformation under continuous-flow conditions. Using the same strategy, they subsequently reported crystalline porous COFs, [HC≡C]x-TPB-DMTP-COFs with adjustable acetylene density [Fig. 2(a)] that are very stable against water, organic solvents, and even strong acids and strong bases [Fig. 2(b)] [39]. Anchoring chiral centers and organocatalytic sites onto the channel walls of [HC≡C]x-TPB-DMTP-COFs was then realized through click reaction [Fig. 2(a)]. The intermediate [HC≡C]x-TPB-DMTP-COFs were converted into chiral organocatalytic [(S)-Py]x-TPB-DMTP-COFs. It is shown that the high catalytic activity of [(S)-Py]x-TPB-DMTPCOFs allows Michael reactions to be conducted in neat water at 25 ℃ and 1 bar, and the insolubility of the COFs gives rise to heterogeneous systems that have superior catalytic activities. A 100% conversion with a high product yield of > 90%, and with high enantioselectivity and diastereoselectivity values, was obtained for the addition reaction between cyclohexanone and β-nitrostyrene. This outstanding catalytic performance was also observed for different β-nitrostyrene compounds, illustrating the generality of the catalyst [Fig. 2(c)].

Very recently, Cui and co-workers reported a multivariate strategy to prepare chiral COFs with controlled crystallinity and stability for asymmetric catalysis [40], in which crystallizing mixtures of triamines with and without chiral organocatalysts and a dialdehyde produced a series of two- and three-component 2D COFs (Fig. 3). These 2D COFs were found to be efficient heterogeneous catalysts in catalyzing the asymmetric amino-oxylation reaction, aldol reaction, and Diels-Alder reaction, with the stereoselectivity and diastereoselectivity comparable to or surpassing their homogeneous analogues. Moreover, they developed a metal-directed synthesis strategy through which chiral Zn(salen)-based COFs can be obtained by the imine-condensations of enantiopure 1, 2-diaminocyclohexane and C3-symmetric trisalicylaldehydes with one or no 3-tert-butyl group (Fig. 4) [41]. The Zn(salen) modules in the COFs allow for installing multivariate metals into the skeletons by a post-synthetic metal exchange process. High crystallinity and porosity was retained, and the exchanged COFs can serve as efficient heterogeneous catalysts for asymmetric cyanation of aldehydes, Diels-Alder reaction, alkene epoxidation, epoxide ring-opening, and related sequential reactions with up to 97% e.e.

In addition to organic functional groups and metal ions, ionic liquids were also introduced into the COF framework by our group [42]. Immobilization of ionic liquids on the channel walls of Py-COFs was performed by using a post-synthetic strategy. The resulting [Et₄NBr]50%-Py-COF afforded a high CO₂ adsorption capacity of 164.6 mg g⁻¹ (1 bar, 0 ℃), and can serve as a highly effective heterogeneous catalyst for the transformation of CO₂ into value-added formamides under very mild conditions. The enrichment of CO₂ into the COF nanopores was considered to lead to its high catalytic activity.

Construction of functional COFs by a bottom-up strategy is a relatively difficult task because it must simultaneously meet the requirements for crystallinity and functionality. Even so, several attempts have been shown to be successful for the synthesis of COFs bearing catalytic sites with this method. For example, a sulfonated building block, 2, 5-diaminobenzenesulfonic acid, was used to construct a sulfonated COF together with another building block, 1, 3, 5-triformylphloroglucinol [43]. The as-synthesized TFP-DABA was found to be a highly efficient acid catalyst for fructose conversion with remarkable yields (97% for 5-hydroxymethylfurfural (HMF) and 65% for 2, 5-diformylfuran) (Table 1), good chemoselectivity, and good recyclability. Cui and co-workers [44] reported the synthesis of two 2D chiral COFs by direct imine condensations of enantiopure TADDOL-derived tetra-aldehydes with 4, 4′-diaminodiphenylmethane (4, 4′-DADPM), which, after treatment with Ti(OiPr)₄, show highly catalytic activity for the asymmetric addition of diethylzinc to aldehydes.

In a similar way, Wang et al. [45] report a facile strategy for the direct construction of chiral-functionalized COFs using chiral (S)-4, 4′-(2-(pyrrolidin-2-yl)-1H-benzo[d]imidazole-4, 7 -diyl)dianiline as a building block (Fig. 5). Two chiral COFs, LZU-72 and LZU-76, are obtained based on the building block (Fig. 6). The catalytic activity of chemically stable LZU-76 was evaluated in the asymmetric aldol reaction, and the reaction afforded the desired aldol products with excellent enantioselectivity (88.4:11.6-94.0:6.0 e.r.), comparable to the enantioselectivity (93.1:6.9) of a model catalyst, (S)-4, 7-diphenyl-2-(pyrrolidin-2-yl)-1H-benzo-[d]imidazole (DPBIP).

In addition, COFs themselves can serve as an efficient catalyst for some reactions. Crystalline covalent triazine frameworks (CTFs) are a special class of COFs that are constructed through the reversible ionothermal trimerization of aromatic nitriles (ZnCl₂, 400 ℃). The majority of CTFs are amorphous materials, and crystalline CTFs are only limited to two nitriles. These CTFs exhibit the best thermostability among all COFs. The inherent basic feature of triazine materials makes it possible to use CTFs as base catalysts. The base catalytic properties of a series of CTFs have been evaluated for the conversion of CO₂ to organic carbonates [46]. The high number of basic nitrogen sites of the as-synthesized CTFs efficiently catalyzed the formation of cyclic carbonates via the cycloaddition of CO₂ to various starting epoxides. Jiang and co-workers [47] develop a π-electronic COF as heterogeneous catalyst that enables the use of columnar π-walls as catalytic beds to facilitate organic transformations in their 1D open channels. The π-frameworks show outstanding catalytic activity in Diels-Alder reactions under mild conditions.

The combination of two different reactive sites working in concert in one substrate to facilitate the reaction is now regarded as the state of the art in catalysis, yet there a dramatic challenge still remains. In particular, bifunctional heterogeneous catalysts are more preferred for cascade/tandem/one-pot synthesis reactions. To date, only a few heterogeneous systems of this sort have been exploited in the COF field.

In 2015, Banerjee and co-workers [48] reported the successful incorporation of bifunctional (acid/base) catalytic sites in the crystalline organocatalytic porous COF (2, 3-DhaTph). Owing to the presence of acidic (catachol) and basic (porphyrin) sites, 2, 3-DhaTph demonstrates significant selectivity, reusability, and excellent ability to perform the cascade reaction. Constructing COFs by employing two types of covalent bonds involving both boroxine and imine linkage has been realized by the Qiu group [49]. These two different linkages provide both acidic and basic sites in one 3D COF material. The double-linked COF can serve as excellent bifunctional catalysts for one-pot cascade reactions, which involves the hydrolysis of the acetal catalyzed by the acidic sites of boroxine group, followed by Knoevenagel condensation catalyzed by the basic sites of imine bonds. A high conversion rate of > 90% for final products has been observed.

In this regard, our group [50] sophisticatedly designed and synthesized a series of 2D COFs containing two different types of nitrogen ligands, namely imine and bipyridine, with controllable contents. The selective coordination of the two nitrogen ligands to two different metal complexes, chloro(1, 5-cyclooctadiene)rhodium(Ⅰ) [Rh(COD)Cl] and Pd(Ⅱ) acetate [Pd(OAc)₂], has been realized using a programmed synthetic procedure. The bimetallically docked 2D COFs show excellent catalytic activity in a one-pot addition-oxidation cascade reaction. The high surface area, controllable metal-loading content, and predesigned active sites make them ideal candidates for using as heterogeneous catalysts in a wide range of chemical reactions.

Recently, Ma and co-workers [51] introduced a new strategy for decorating functional active sites into COFs (Fig. 7). The underlying principle of this strategy is to integrate catalytic components with excellent flexibility and porous heterogeneous catalysts. The surface Lewis acid sites are anchored on the walls of a COF (COF-TpBpy) by the metalation of build blocks, namely bipyridine units, of the COF. In situ radical polymerization of ionic monomers in COF-TpBpy accommodates the linear polymers in COF channels. The excellent flexibility of linear polymers makes it possible in close proximity to surface Lewis acid active sites. Through this strategy, Lewis acid sites on the walls can cooperate well with the co-catalyst, Br⁻, of the linear polymers to efficiently catalyze the cyclo-addition of the epoxides and CO₂. The dramatic activity improvements have been achieved for the composite catalysts in relation to the individual catalytic component.

The inherent host-guest chemistry of COFs enables the implementation of desired substances by filling the framework pores with various guest molecules. In particular, the doping of COFs with metal NPs is of great interest for heterogeneous catalysis. Using COFs as catalyst supports could potentially have the following advantages: (1) The high surface area of COFs can provide more active sites, which generally facilitates chemical reactions; (2) COFs are pre-designable, with well-defined topological structures, which can effectively confine metal NP growth and thus the size of NPs; (3) the pore channels in COFs are well isolated, and, thus, the aggregation of entrapped NPs can be hampered to an certain extent; (4) the pure organic component of COFs makes it easy to introduce the surface functional groups that can anchor metal NPs; (5) due to the linkage of covalent bands, many COFs are chemically/thermally stable, and thus would be resistant to decomposition under harsh reaction conditions. Until now, only three catalytically active metal NPs (Au, Pd, and Pt) have been finely dispersed into the COF frameworks, and their corresponding hybrids exhibit high catalysis activities.

The first attempt to load metal NPs was carried out by Banerjee's group in 2014 [52]. They synthesized a COF-supported, highly stable Au(0)-based catalyst by using a solution infiltration method (Fig. 8). The as-synthesized Au(0)@TpPa-1 catalyst shows a higher activity towards nitrophenol reduction reaction than HAuCl₄·3H₂O, imparting the advantages of heterogeneous catalysts. The crystallinity of the Au(0)@TpPa-1 was found to remain well over a number of cycles, which can be attributed to the high stability of TpPa-1.

Soon after, Banerjee's group further report the successful incorporation of Pd(0) NPs into the pores of TpPa-1 with the same method [53]. The Pd(0)@TpPa-1 catalyst exhibits excellent catalytic activity towards the Cu-free Sonogashira, Heck, and sequential one-pot Heck-Sonogashira cross-coupling reactions under harsh conditions. Importantly, this catalyst was found to be highly stable under either acidic or basic reaction conditions, evidenced by negligible metal leaching, non-sintering behavior, and good recyclability. Moreover, the immobilization of Pd(0) NPs has also been realized by this group via a different synthetic strategy in which a pre-designed Pd²⁺-anchored building block is used to obtain a porous, crystalline Pd@TpBpy COF. Pd(0) NPs are in situ produced without using conventional reducing agents because, during the course of Pd@TpBpy formation, the breakage of the Pd-N bond of the Bpy-PdCl₂ complex results in the formation of Pd NPs [54]. The average size of Pd NPs is in the range of 12 ± 4 nm, which is much larger than the pore size of Bpy COF (2.3 nm), revealing that these NPs are distributed around the surface and the interlayer spacing of the COF. Even so, this hybrid material has been found to be an excellent reusable heterogeneous catalyst for the synthesis of 2-substituted benzofurans from 2-bromophenols and terminal alkynes via a tandem process.

Very recently, a homochiral nitrogen-rich COF, namely CCOF-MPC, has also been used as catalyst support to immobilize metal NPs (Pd and Pt) [55]. Both Pt and Pd NPs immobilized inside the cavity of the COF possess a narrow size distribution (1.7 ± 0.2 nm) with a high loading level. The synthesis strategy is shown in Fig. 9. Such a small size of metal NPs is attributed to the crystallinity of the COF support and the presence of thioether groups inside the cavities. It has been further demonstrated that the obtained Pd@CCOF-MPC hybrid is a highly active asymmetric heterogeneous catalyst that effectively promotes the Henry and reductive Heck reactions with excellent yield and stereoselectivity under mild reaction conditions.

In addition, Pd NPs have also been successfully loaded into an amphiphilic triazine-containing COF (trzn-COF) by Vaidhyanathan and co-workers [56]. The facile single-step loading of Pd NPs into COF cavities has been realized by stirring trzn-COF in an ethanolic solution of Pd(OAc)₂. Surprisingly, the reduction of Pd²⁺ into Pd(0) NPs occurs in this single step, and the authors attribute this to the richness of basic nitrogen in the COF that could play a role of reducing agent. Their calculation simulation shows that the presence of a triazine group and Schiff bond linker is favorable for the immobilization of metal NPs. This material can catalyze simultaneous multiple-site Heck couplings and C-C couplings with good recyclability. Moreover, as an oxidation catalyst, a 100% conversion of CO to CO₂, has been observed at 150 ℃ with no loss of activity with time and between cycles. The loading of metal NPs into a COF matrix with a controlled size and shape is of interest in the catalysis field. In this regard, this kind of composite has been generated by a simple one-pot method. The PVP-modified Au NPs with desired size and shape, as well as tunable contents, has been successfully loaded within the COF pores during the synthesis process of 2D TAPB-DMTP COF [57]. The resulting composite retains its thermal stability, high crystallinity, large surface area, open mesoporous structures, and good catalytic performance for reduction of 4-nitrophenol.

In contrast to inorganic catalysts (such as zeolites and mesoporous silica), a special attraction exists between reactants and organic framework pores of COFs, such as hydrophobic and π-π interactions. This feature may endow COFs with special properties in catalysis. In 2014, Yan and co-workers [58] designed and prepared two 3D microporous base-functionalized COFs, termed BF-COF-1 and BF-COF-2, through the condensation reaction of a tetrahedral alkyl amine, 1, 3, 5, 7-tetraaminoadamantane (TAA), combined with 1, 3, 5-triformylbenzene (TFB) or triformylphloroglucinol (TFP) [Fig. 10(a)]. Using these two COFs as catalysts, both BF-COFs show remarkable conversion rates (96% for BF-COF-1 and 98% for BF-COF-2) and good recyclability in base-catalyzed Knoevenagel condensation reactions. Importantly, the COFs exhibit highly efficient size selectivity due to the pore size of the COFs (Fig. 10(b)). The strong interaction between reactants and the COFs may lead to a high efficiency in size selectivity. Recently, a dynamic 3D COF, LZU-301, an analogue to COF-320, has been prepared by Wang et al. [59] Knoevenagel condensation, catalyzed by LZU-301, takes place within the pores of the COF.

Porous network structures, together with robustness and molecular functionality with the possibility to load desirable species, favor the generation of highly effective photocatalysts and electrocatalysts. This is the case for new COF materials useful for visible-light-driven H₂ production with the help of Pt as a proton reduction catalyst (PRC). When the suitable organic solid was illuminated by visible light, this Pt-docked COF continuously produced H₂ from water without structural degradation [60]. This result indicates that photoactive COFs are well-defined model systems with which to carry out efficiently photocatalytic processes. CdS NPs have also been deposited on a highly stable 2D COF matrix, and the generated hybrid was used as a photocatalyst for visible-light-driven hydrogen production. In particular, the efficiency of CdS-COF hybrid with different COF contents has been studied. Upon the introduction of only 1 wt% of COF, a 10-fold increase in the overall photocatalytic activity has been observed. The hybrid with 10 wt% COF content exhibits a marked H₂ production of approximately 3700 μmol h⁻¹ g⁻¹, which is significantly higher than for bulk CdS NPs (124 μmol h⁻¹g⁻¹) [61]. A series of water- and photo-stable 2D azine-linked COFs has been also synthesized from hydrazine and triphenylarene aldehydes with varying numbers of nitrogen atoms [62]. In addition, structure-property-activity relationships in a pyridine containing azine-linked COF for photocatalytic hydrogen evolution have been elucidated [63]. The electronic and steric variations in the precursors are successfully transferred to the resulting frameworks, which results in progressively enhanced light-induced hydrogen evolution with increasing nitrogen content. A benchmark example of COF-based photocatalysts for solar fuel production from CO₂ has been realized. It is demonstrated that the visible-light-harvesting capacity, suitable band gap, and highly ordered π electron channels contribute to the excellent performance of the COF film photocatalyst [64]. In 2015, Jiang and co-workers examined the photocatalytic activity of CuP-DHPh COF by using 1, 3-diphenylisobenzofuran (DPBF) as a label for singlet oxygen generation. In comparison with other CuP derivatives, the CuP-DHPh COF is exceptionally active as a photocatalyst, exhibiting a 10-20-fold enhancement in activity [65]. Recently, a π-conjugated TpMA COF with triazine units and cyclic ketone units was artfully designed and synthesized. This COF was found to exhibit an excellent visible-light photocatalytic capacity for the decomposition of organic pollutants [66]. Chemically stable CTFs have been also used as photocatalysts for H₂ evolution in the presence of Pt under visible-light irradiation, and relatively low hydrogen evolution rates of approximately 200 μmol h⁻¹ g⁻¹ were obtained [67]. Subsequently, a fast and facile route for the optimization of CTFs for photocatalytic hydrogen production was presented by Thomas's group [68]. They found that the optimized CTF catalysts showed an average hydrogen evolution rate of 1072 μmol h⁻¹ g⁻¹ under visible light (> 420 nm). Very recently, Lotsch et al. [69] further demonstrated photocatalytic hydrogen evolution using COF photosensitizers with molecular proton reduction catalysts. With an azine-linked N2-COF photosensitizer, chloro(pyridine)cobaloxime co-catalyst, and TEOA donor, a H₂ evolution rate of 782 μmol h⁻¹ g⁻¹ and a TON of 54.4 was obtained in a water/acetonitrile mixture [69]. In addition to hydrogen evolution, a 2D COF was found to be a highly efficient, metal-free, recyclable heterogeneous photocatalyst for oxidative C-H functionalizations under visible-light irradiation using O₂ as a green oxygen source [70]. In addition, 2D COFs can act as effective type-Ⅱ photosensitizers for photodynamic inactivation of bacteria [71]. In addition to 2D COFs, 3D COFs have been attempted for photocatalytic study. Wang and co-workers [72] reported two 3D porphyrin-based COFs (3D-Por-COF and 3D-CuPor-COF) bearing photoelectric units, and found that 99% 9, 10-dimethylanthracene (DMA) degraded after photoirradiation for 90 min (Fig. 11).

Incorporating tunable molecular units within an extended COF facilitates the use of COFs in electrocatalysis applications. Lin et al. [73] reported modular optimization of COFs, in which cobalt/copper porphyrins were used as building units to prepare a catalytic material for aqueous electrochemical reduction of CO₂ to CO (Fig. 12). The catalysts show high Faradaic efficiency of 90% and a TON of up to 29000 with an initial time of flight of 9400 h⁻¹ at pH 7 [73]. To rationally design COF-based electrocatalysts for oxygen reduction and evolution reactions in fuel cells and metal-air batteries, a simulation calculation was recently performed to identify the direction for transition-metal-loaded COFs [74]. Furthermore, COFs have been recently transformed into carbon materials upon pyrolysis, and the resulting porous carbons exhibit excellent electrocatalytic performance for the oxygen reduction reaction [75].

Owing to the unique features of COFs and recent interesting exploitation of their properties, COFs undoubtedly demonstrate significant potential and advantages in catalysis. Therefore, COF catalysis has attracted a great deal of attention, and some progress in the field has been made in recent years. However, the investigation of catalytic COF materials is still in its infancy, and COF catalysis still faces many challenges in practical applications, e.g., how to realize the desirable design of COFs for a special purpose, and how to further improve the chemical/thermal stability of COFs so that more reactions can be carried out under very harsh conditions (e.g., strong acidity or alkalinity, high temperature, and high overpotential). The successful synthesis of crystalline COFs has long been an intractable problem, due to the lack of universal regulation to construct crystalline COFs, and significant time and effort must be spent on searching for the appropriate reaction conditions. Therefore, there is a clear need to expand the synthetic possibilities in different ways. Currently, synthesis of COFs is only limited to the laboratory, and it has been a challenge to prepare COFs on an industrial scale for prospective practical applications. Moreover, finding a new synthetic route to reducing the cost of COFs is also a difficult task, and far in the future. It is also of interest to develop 3D COFs with versatile inner cavities, which can provide active sites on the pore surface and transport reactants/products to or from inner reactive vessels. This would enable supramolecular catalysis or enzyme catalysis to take place in an appropriate space under nanoscale confinement, allowing multicomponent reactions and a wide range of substrates from small molecules to large polyaromatics and carbohydrates. In the future, the goal of COF catalysis is to develop heterogeneous catalysts with high stability, low cost, and high conversion and selectivity. New synthesis methods and process routes for large-scale synthesis should be explored simultaneously, thus laying the foundation for the commercialization and industrialization of COF catalysts.

Acknowledgments

This work was supported by the National Natural Science Foundation of China (21473196, 21406215) and the State Key Laboratory of Fine Chemicals, Dalian University of Technology (KF1415). The authors are also grateful for funding from Dalian Institute of Chemical Physics, Chinese Academy of Sciences (DICP_M201401).