1. Introduction

Covalent organic frameworks (COFs) are an emerging category of two-dimensional (2D) or three-dimensional (3D) crystalline porous materials composed of light elements (C, O, H, N, B, and others) with organic building blocks connected by reversible covalent bonds [1]. COFs present low density, permanent porosity, highly ordered structures, large specific surface area and relatively good physicochemical stability. Especially, their structures can be pre-designed to achieve functional customization [2], which endows COFs with great potential for use in many fields such as gas separation and storage, energy storage and conversion, heterogeneous catalysis, drug delivery, and chemical sensing [3-8].

The scalable and sustainable production of high-quality COFs holds the key to their practical application. During the formation of COFs, covalent bond formation and breakage is repeated continuously due to reversible reactions, ultimately reaching the most stable equilibrium state [9]. Solvothermal synthesis is usually adopted in order to facilitate the error correction process for the formation of these extended periodic structures. Nevertheless, this method has to be performed under severe conditions such as high temperature and high pressure, with toxic and harmful solvents, and also requires a reaction time of several days or even longer, thereby not conducive to the large-scale preparation of COFs [10]. In this respect, researchers have developed a series of new schemes for synthesizing COFs, which provides theoretical basis and technical support for possible mass production. Although there have been several reviews summarizing the synthesis strategies of COFs or targeting a particular aspect of application, commercializing and producing COFs also requires considering more including costs, operation complexity, production efficiency. A comprehensive review on the scalable and sustainable synthesis of COFs is necessary to be supplemented in this field.

Herein, we offer a systematic review on studies related to efficient, green and ambient synthetic strategies for COFs (Fig. 1). The aim is to provide an in-depth understanding of the different synthetic strategies by highlighting their pros and cons, providing key aspects to improve their structures and properties so as to increase the potential in practical applications. Finally, we outline the challenges of large-scale production and give an outlook on their future development.

2. Solvothermal method

Solvothermal method is a common way to synthesize COF, and since the first case of COF reported by Yaghi and colleagues in 2005 [1], a variety of COFs have been synthesized successively in the same way (Table 1). Due to the dependence on the solubility and reactivity of the monomers and reaction reversibility, reaction conditions including pressures, times, temperatures, solvents and solvent ratios, as well as the type and concentration of the catalyst have important effects on the morphology and structure of the prepared COFs. Usually the reaction is maintained in a Pyrex tube and a temperature of 80–120 ℃. The sealing reaction is carried out for 2–9 days under the high temperature and high pressure to promote the reaction, ensuring the long-range order of the COFs. Feng et al. synthesized TD-COF with the largest pore size of 10.0 nm, demonstrating the excellent synthetic universality of the method [11]. However, the synthesis process requires a reaction of up to 5 days under high temperature, and the post-processing operation is complex. It requires further treatment of the solid with supercritical CO₂ in a Tousimis Samdri PVT-3D critical point dryer. Loh group obtained two structural variants TPE-COF through different mixed solvents o-dichlorobenzene/n-butanol or 1, 4-dioxane [12]. Ning's group combined coordination and dynamic covalent chemistry to prepare three 2D copper(Ⅰ) cyclic trinuclear units (Cu(Ⅰ)-CTUs) based covalent metal-organic frameworks (CMOFs) through a one-pot reaction. Among them, JNM-15 has the highest catalytic activity and high adsorption capacity for CO₂ [13]. Feng et al. found that when the volume ratio of mesitylene and 1, 4-dioxane was 4:1, the obtained ZnP-COF was lack of long-range molecular order, but the relative high crystallinity was obtained at the volume ratio of 9:1 and 19:1 [14]. Therefore, it takes more time and effort to explore the optimal reaction conditions of COF with high crystallinity and ideal structure. In addition, the crystallinity of COF depends on the strength of the chemical bonds, reversibility and rotatability, and reaction rate. Therefore, catalysts including acetic acid, Sc(OTf)₃ and PTSA are widely introduced in various solvents to promote the reversible reaction and accelerate the reaction rate, which can make the synthesized COFs have a rigid covalent bond (i.e., C═N bond), better crystallinity and more regular pore systems. However, the prolonged high temperature and high pressure are still a major stumbling block hindering the scalable production, and disposal of toxic and hazardous solutions is also a tiresome burden [15-23]. From the above studies, the solvothermal method has some defects, and it is urgent to develop new approaches for COF synthesis.

3. Green synthesis strategies

In order to overcome the shortcomings of usage of large amount of deleterious organic solvents during the solvothermal method, researchers have explored various green strategies for synthesizing COFs in recent years. According to the difference of reaction medium, such strategies could be divided into the following categories: Aqueous phase synthesis, low toxic organic solvent phase synthesis, ionic liquid phase synthesis, solvent-free synthesis and deep eutectic solvent synthesis.

3.1. Aqueous-phase synthesis

Water, as a universal green solvent is generally preferred for COFs synthesis, because it not only reduces production cost and safety problems, but also does no harm to the environment. In 2016, Thote's group first studied the formation of keto-enamine crystalline and porous polymers in water [24]. Based on the principle of dynamic covalent chemistry, a series of COFs, such as TpPa-1 and TpPa-2, were synthesized via Schiff-base condensation of 1, 3, 5-triformylphloroglucinol (Tp) and corresponding amines in water and acetic acid medium. Due to the hydrogen bonding environment provided by acetic acid-water medium, which is conducive to supporting reversible covalent bonding, these COFs are generally preserved in terms of crystallinity and porosity. Zhang's group prepared COF-LZU1 from 1, 3, 5-trimethylbenzene and p-phenylenediamine at room temperature by using CO₂-dissolved aqueous solution as the solvent [25]. X-ray diffraction pattern demonstrates that increasing pressure can help to enhance crystallinity. Moreover, the specific surface area of COF synthesized under 4.5 MPa is much higher than that of the COF synthesized by solvothermal method.

By using water as solvent and acetic acid as catalyst at relatively low temperature 80 ℃, Jesús Á et al. prepared three structurally and chemically different imine-based COFs with high crystallinity and porosity: TAPB-BTCA-COF, TZ-BTCA-COF and HZ-BTCA-COF [26]. This method has advantages of high yield, avoiding harmful organic solvent and high temperature. The imine-based COFs can be generated at 35 ℃ in water, and with the aid of microwave heating, the reaction time can be shortened from a few days to 5 h. However, there are still obstacles in synthesizing COFs on a large scale. Recently, Zhou et al. used amphiphilic amino acid derivatives with long hydrophobic chains to self-assemble into micelles, preventing materials from precipitating from monomer polymerization through weak interactions (hydrogen bonding, electrostatic interaction, etc.) in water, thereby regulating the polymerization and crystallization process (Fig. 2A) [27]. Disordered polyimines are obtained in the micelle, and these polyimines are then gradually converted into crystals. The COF connected by single crystal imide was prepared in aqueous solution. At the same time, the method is versatile, and the researchers obtained five different 3D COFs and one 2D COF on the gram scale. In the same year, Xu et al. used acetic acid to preactivate aldehyde monomer to improve the reactivity of aldehyde monomer in water, and then the amine monomer and preactivated aldehyde monomer aqueous solution was stirred in room temperature to rapidly synthesize 16 kinds of imine-linked COFs (Figs. 2B and C) [28]. Compared with traditional solvothermal synthesis, the method can form highly crystalline COFs with large surface area in a few minutes with relatively high yield.

In 2023, Guo et al. adopted a two-step dissolution-precipitation (DP) strategy to synthesize a variety of ketoamine and imine-linked COFs in aqueous phase at room temperature in 5 min [29]. In the process, imidazole was used to improve the solubility of organic monomers. After the reaction, acetic acid was added to obtain a large amount of COFs precipitation immediately. The method is simple and low cost, which provides a new idea for industrial synthesis and application of COFs. In the same year, Zhang and co-workers condensed the phthalic acid group of NiPc and the aromatic amine group of 2HPor into crystalline NiPc-2HPor COF by hydrothermal synthesis in a pure water without using any catalyst [30]. NiPc-NiPor COF was further synthesized by post-synthesis coordination reaction. The formed polyimides-linked COFs (PI-COFs) showed high chemical stability and activity for the electrocatalysis of methanol oxidation reaction (MOR) coupled with electrochemical CO₂ reduction (ECR). Unfortunately, this method requires a high temperature of 210 ℃ and two days of reaction.

Compared with solvothermal method, aqueous-phase synthesis can reduce the use of organic solvents and reduce the harm to the environment, but a large amount of water may also cause the hydrolysis of COFs, which has an impact on the reaction process. At the same time, the low imide formation rate and high imide fracture rate in aqueous solution make the synthesis of COFs fully self-correcting, which is conducive to controlling the reaction process and improving the crystallinity of the product.

3.2. Ionic liquid phase synthesis

Ionic liquids (ILs) are a kind of new green solvent composed of organic cations and inorganic anions which has a very low vapor pressure and high thermal stability, as well as good solubility and designability. Ma et al. reported a simple and green method to synthesize β-ketoenamine-linked COF in water at room temperature using [BmimN(CN)₂] as catalyst (Fig. 3A) [31]. The synthesized COF has good crystallinity, high porosity and excellent stability, and can be easily produced in the aqueous phase. In addition to using ILs as catalysts for the synthesis of COFs, researchers have developed a number of preparation methods using ILs as solvents. Guan et al. reported to construct 3D-COF by ionic thermal synthesis at ambient temperature and pressure in an open system, and a series of 3D-IL-COFs with multifold interpenetrated diamondoid (dia) nets were successfully prepared (Fig. 3B) [32]. These 3D-IL-COFs have high crystallinity, high specific surface area and good affinity for CO₂, which enabled an excellent selectivity in separating CO₂ from air (N₂) and natural gas (CH₄).

Dong and co-workers synthesized a series of 2D keto-enamine COFs by ion thermal method using room temperature imidazolium IL [BSMIm]HSO₄ as the solvent [33]. Ionic thermal synthesis can also be performed in an open container under atmospheric pressure, eliminating mixed organic solvents used in the complex process of solvothermal synthesis. In addition, imidazolium-based ILs can be recycled without losing activity. By using the same IL under similar conditions, Zhao et al. synthesized keto-enamine-linked 2D COFs (TFP-EB and TFP-DAAQ) and polyimide COFs (PMDA-TAPA and PMDA-TAPB) [34]. The preparation process is environmentally friendly, and the prepared COFs have good crystallinity and high thermal stability. Gao et al. synthesized 2D COFs with high yield by polycondensation reaction in non-volatile IL [35]. This method is efficient and simple, and can generate COFs with high crystallinity and porosity, and can produce high-quality microcrystals with regular macroscopic shapes that are difficult to be obtained by solvothermal method. In 2020, Qiu's group developed a novel, environmentally-friendly, template-free method for the synthesis of HP-COFs in ILs [C_nmim][BF₄] (n = 4, 6, 10) [36]. The HP-COFs have high crystalline network, good stability and hierarchical pore structure, and their porosities can be easily adjusted by changing the alkyl chain length of the IL. These HP-COFs showed significant catalytic activity for C—C bond formation, especially for C—C coupling reactions of macromolecules, compared with single-pore COFs synthesized by solvothermal method (Fig. 3C). Gao et al. developed an adjustable IL-H₂O double diffusion control platform for the synthesis of independent COF membranes from Schiff-base reactions [37]. A set of ILs were selected to form an incompatible interface with water. Both amines and catalysts were dissolved in water, while the acetaldehyde was dissolved in IL for polymerization reaction and membrane growth. The IL-H₂O interface can be easily regulated by changing the alkyl chain length of the cations in ILs, providing a controllable reaction zone. The prepared COF membranes have high permeability of both water and organic solvents, and have good dyes rejection.

Chen's group used IL-mediated dynamic polymerization strategy to synthesize enzyme@COF biocatalysts with high crystallinity in one step in aqueous-phase. In this method, only 2 µL imidazolyl green IL was required to catalyze and accelerate the reaction, and the huge enzyme molecules were quickly fused into amorphous polymer phases [38]. The formed amorphous enzyme-polymer composites undergo spontaneous crystalline phase "repair" with the help of the same ionic liquid to form highly crystalline enzyme@COF. This green strategy is applicable to different COFs, and the crystallinity of all COF "exoskeleton" is comparable to that synthesized by traditional solvothermal methods.

These reports lead us to believe that ILs as solvent to synthesize COFs has the characteristics of simple process, high synthesis efficiency, mild reaction conditions, and reusable solvent. The synthesized COFs have high crystallinity, good thermal stability and regular pore structure. However, due to the relatively high cost of ILs, current research is still limited to a small-scale synthesis, and there are still challenges in expanding the production scale.

3.3. Deep eutectic solvent phase synthesis

Deep eutectic solvents (DES) are another kind of green solvent formed by strong hydrogen bond interaction between hydrogen bond donors (HBDs) and hydrogen bond acceptors (HBAs). It has the advantages of simple preparation process, low vapor pressure, high thermal stability and non-flammability. In the open system, Qiu and co-workers synthesized 2D and 3D COFs within 2 h by using DESs as green functional medium, which have high crystallinity and high specific surface area [39]. There was no significant loss of DESs activity after 3 times of recycling. Moreover, a new type of azo-based 3D COF with hierarchical stomatal structure was successfully synthesized, which cannot be obtained by solvothermal method.

Gao et al. successfully prepared COF-DES from 1, 3, 5-tris(4-aminophenyl)benzene and 2, 5-dihydroxyterephthalaldehyde, which have good dispersibility, large specific surface area (444.56 m²/g) and suitable pore size (25.9 Å) [40]. Due to the unique structure and functional group in COF-DES, the interference of alkaloids and other compounds can be avoided, the COF-DES therefore has excellent selective adsorption performance for flavonoids. Chen's group prepared a layered COF based on imine-linkage using DES as green medium, which has high thermal and chemical stability [41]. The synthesized COF was modified with catalytic copper, then the COF based catalyst was obtained. The catalyst has good crystallinity and high porosity, which shows excellent photocatalytic performance in visible-light-driven coupling reactions of terminal alkyne with H-phosphonates. In addition, the photocatalyst can be recycled and reused up to 8 times without significant loss of reactivity. Moreover, they also used DES as a solvent to prepare a simple azine-linked COF. The as-synthesized COF was used as a heterogeneous ligand to anchor Pd^Ⅱ. The obtained Pd-supported COF (defined as Pd/TFPT-azine-COF) as heterogeneous catalyst is very stable in practical applications, and Pd^Ⅱ does not leach or lose activity after 7 consecutive cycles [42].

Compared with ILs, DESs has the advantages of low price, non-toxicity and good biodegradability. Using DESs as a solvent for the synthesis of COFs not only has mild reaction conditions, but also can be recycled. However, at present, there are few researches related to this technology, and the synthesis process is not mature, which needs further development.

3.4. Solvent-free synthesis

Solvent-free synthesis of COFs means that only monomers are needed to react in the reaction process without the help of solvents. This method has low cost, simple synthesis process and post-treatment process, and avoids the use of organic solvents, which has great development potential in industrial practical applications.

Li's group proposed an efficient synthesis method of azine-linked COF (ACOF) without solvent and catalyst. Without inert gas protection or mechanical grinding, only monomer TFB and HZ were added to a sealed vial and stored at 80 ℃ for three days to obtain deep red ACOF (Fig. 4) [43]. The synthesized ACOF has high crystallinity, good physical and chemical stability, and displays good adsorption performance for both radioactive heavy metal U(Ⅳ) and common heavy metal Hg(Ⅱ).

Wang et al. synthesized an olefin-linked NKCOF-10 through benzoic-anhydride-catalyzed Aldol reaction between the activated methyl groups of 2, 5-dimethylpyrazine (PZ) and 1, 3, 5-trimethylbenzene (TFB) under solvent-free condition [44]. The prepared COF has a layered honeycomb crystalline structure with high specific surface area and excellent stability to harsh conditions including strong acid/base. Additionally, the proton carrier H₃PO₄ was anchored in the hole of NKCOF-10 by pyrazine functional group to prepare the H₃PO₄@NKCOF-10. This material exhibits extremely high proton conductivity under practical fuel cell operating conditions. Yang and co-workers adopted a solid-phase synthesis strategy to make the self-polymerized 1, 2, 4, 5-benzenetetranitrile into 2D phthalocyanine conjugated COFs (MPC-2D-cCOFs) materials under a metal template without the use of organic solvents and catalysts [45]. MPC-2D-cCOFs have good crystallization performance and ordered pore structure, which offer channels for ion transport and is conducive to efficient utilization of active sites. Moreover, this series of COFs has a solid 2D completely conjugated structure, endowing it with good stability and electron transport performance. Jiang et al. reported the thermal polycondensation reaction of a ternary monomer and NH₃, and the synthesis of 2D aza-bridged bis(phenanthroline) macrocycle-linked COF (ABBPM-COF) in the absence of solvent and catalyst [46]. The crystalline structure of ABBPM-COF was identified as ABC stacking mode, which is tolerant to 9 mol/L KOH aqueous solution and concentrated HCl solution. Moreover, the insulating ABBM-COF becomes a semiconductor material when exposed to iodine vapor, with a conductivity of 2.6 × 10⁻⁴ S/cm at room temperature, indicating the application potential in battery, electrochemical catalysis, electron and proton conduction.

Recently, Pan's team successfully prepared TFPDQ-COF/Graphene (TFPDQGO) nanocomposites by in-situ growing 2D redox active COF (TFPDQ-COF) with abundant redox active sites on graphene conductive substrate under solvent-free conditions [47]. The TFPDQ-COF particles are tightly fixed on the graphene surface through the molecular interaction and amide bond, thus reducing the accumulation/agglomeration of TFPDQ-COF. Further, the presence of graphene allows TFPDQGO nanohybrid materials to have a larger specific surface area and more porous structure. The nano-composite has large specific capacitance and long cycle life, therefore has wide application prospects in supercapacitors and hybrid capacitive deionization (HCDI).

Although the solvent-free synthesis avoids the use of organic solvents, it is similar to solvothermal synthesis, in most cases still requires a higher reaction temperature and long reaction time. There is still a lot of room for research progress in energy saving.

3.5. Low toxic organic solvent phase synthesis

The synthesis of COFs with low toxic organic solvent has the advantages of simple operation and usage of green solvent at room temperature. Wang's group synthesized for the first time β-CD COF using heptakis(6-amino-6-deoxy)-β-CD (Am7CD) and terephthalaldehyde (TPA) through acid catalyzed aldimine condensation in mixed solvent of water and ethanol at room temperature [48]. The formed pale-yellow precipitate was proved to be COFs with the desired chemical bond by PXRD. Weaving β-CD into the skeleton by mesh chemistry allows the integration of a large number of β-CD units (50 mol%), much higher than the β-CD polymer. Compared with amorphous β-CD polymers prepared with the same reagent, β-CD COF have larger specific surface area, more uniform pore size and higher thermal stability. Zhang and co-workers adopted a new set of reaction conditions, organic solvent/water/MOH (M = Na, K, Cs), which is different from typical organic solvent and acid catalyst system [49]. In the presence of MOH, phenol-based COFs could be easily synthesized by refluxing the organic solvent/water solution of monomers. More recently, Qiu et al. used 1, 3, 5-tris(4-aminophenyl)benzene (TAPB) and 2, 5-dimethoxy-benzene-1, 4-dicarboxaldehyde (DMTP) as the ligands to react in a green reaction medium composed of polyethylene glycol 400 (PEG-400) and water by adjusting the proportion and concentration of solvents and adding modifiers. Four different forms of COFs were successfully synthesized (Fig. 5) [50]. The reaction was carried out in a green reaction medium consisting of polyethylene glycol 400 (PEG-400) and H₂O. The mechanism of adjusting COFs morphology by PEG-based aqueous solution was preliminarily explored as well. Although the reaction conditions of the above studies are mild, in order to preserve the crystallinity and specific surface area of COFs, there is a defect that the reaction takes a long time in the synthesis process.

Through the examples summarized above, the synthetic mechanism of these methods and the application prospect of the products can be well comprehended. However, there are few reports on the large-scale synthesis of COFs by these green routes. It is still urgent to develop more efficient and green synthesis methods of COFs.

4. Efficient synthesis strategies

Solvothermal synthesis method is time-consuming, with reaction times typically up to several days, researchers are therefore pursuing strategies that can synthesis high-quality COF in hours or in minutes. With the assistance of microwave action, sonochemical action, and mechanical action, the synthesis time is greatly reduced. Moreover, spraying method, coating method, and other techniques are suitable for the rapid preparation of large areas of COF.

4.1. Microwave method

Microwave method has already been applied to the synthesis of nano-porous materials such as silicon materials [51], metal organic frameworks (MOFs) [52], metal oxides [53] and carbon materials [54]. Different from traditional solvothermal method relying on heat conduction, microwave heating renders an efficient and uniform increase of the temperature [55]. Rapid heating greatly reduces the reaction time, having less loss of heat. In 2009, microwave heating was for the first time used for the synthesis of COFs (Fig. 6). Microwave heating synthesis of 2D COF-5 and 3D COF-102 took only 20 min separately, which is more than 200 times faster than the traditional solvothermal method. Surprisingly, the Brunauer-Emmett-Teller (BET) surface areas of COF-5 synthesized is 2019 m²/g, significantly higher than the surface area of 1590 m²/g obtained by the solvothermal method [56].

In a similar manner, viologen-based COFs have been synthesized under microwave condition in 2017 [57]. Researchers choose 1, 1′-bis(2, 4-dinitrophenyl)-[4, 4′-bipyridine]−1, 1′-diium dichloride (BDB) and 1, 3, 5-tris(4-aminophenyl)benzene (TAPB) as building monomers, then dissolved them in an ethanol-water mixed solvent (4:1, v: v) to synthesize highly stable viologen based COFs by microwave heating and solvothermal method, respectively. Amorphous hollow spherical products were yielded, but the traditional method took 3 days, while the microwave method took only 2 h. Interestingly, when dioxane was used as the solvent, self-templated hollow tube was produced by whatever methods, which indicates that viologen-based material is solvent-dependent (Fig. 6B). Whereas, when ethanol: water (1:1, v: v) was used as the solvent, crystalline covalent organic gel framework (COGF) was obtained, which is stacked into several layers-thick nanosheets. Rigorous test has proven that these COGFs remain stable in both water, acid and base solutions. Furthermore, their abilities to capture iodine from solution and vapor phases have been investigated, the hollow tube structure performed the best.

Ionic COFs show great potential for applications in ion conduction, separation, catalysis and so on. Wei's team prepared three kinds of COFs (SJTU-COF-Cl, SJTU-COF-AcO and SJTU-COF-CF₃SO₃) with the same cationic framework but different counter anions by microwave-assisted method. The initial SJTU-COF-Br was prepared by microwave heating method in 3 h (Fig. 6C), then a rapid microwave-assisted strategy was applied to exchange Br⁻ to other anions (Fig. 6D). The characterization shows that the main cationic skeletons of SJTU-COFs were stable during microwave-assisted anion exchange process. Among these cations, due to the interaction between acetic anion and CO₂, the CO₂ absorption capacity of SJTU-COF-AcO is significantly enhanced, which is 1.7 times that of the pristine COF SJTU-COF-Br. This shows that it is an effective new method to improve the performance of COFs for gas storage and separation by adjusting the pores of COFs through anion exchange [58].

A key difference between microwave-assisted synthesis and conventional synthesis is the heating mode. The conventional heating method is heat conduction or heat convection, which is slow, non-uniform and non-selective. Microwave heating relies on thermal radiation, which can directly heat the reactants [55, 59, 60]. The microwave method, by contrast, has the following advantages: (1) High synthesis rate and short reaction time. More rapid and uniform heat transfer, microwave equipment and closed vessel combination can be achieved in conditions above the boiling point of the reaction. The microwave method reduces the activation energy barrier, shortening the reaction time from hours to minutes (Table 2) [61-67] (2) More uniform products; (3) Better product properties in terms of higher crystallinity, smaller particle size, narrower pore size distribution, higher purity, and improved physical and chemical properties; (4) Reduced by-product generation thereby increasing the yield; (5) Selective heating, good repeatability, easy scale-up, low energy consumption, and environmental friendly.

4.2. Sonochemical method

In the sonication process, the phenomenon of bubble generation and rupture in solution called acoustic cavitation occurs, which causes extremely high local temperature (> 5000 K) and pressure (> 1000 bar), a special phenomenon capable of accelerating the crystallization rate is achieved [68, 69]. The first application of sonochemical method to COFs synthesis appeared in 2012, and the classical COF-1 and COF-5 were synthesized by the method in 1 h (Fig. 7A) [68]. The synthesis rate was dramatically increased compared to the reaction of 72 h by the solvothermal method. As the particle size of COF-1 prepared by sonochemical method was 400 times smaller (Fig. 7B), the BET area of COF-5 reached 2122 m²/g, which was much larger than that obtained by the solvothermal method of 1590 m²/g and similar to the BET area of 2019 m²/g synthesized by microwave method. As shown in Fig. 7C, the mean crystal size of COF-5 was ca. 250 nm, which was approximately 100 times smaller than the sample prepared by solvothermal method. This may explain the larger BET area of the COF prepared by the sonochemical method. Notably, the sonochemical method has the potential for large-scale production, with a space-time yield of 45 kg m⁻³ day⁻¹ for COF-5, which is 9 times larger than that of the solvothermal method. Apart from the synthesis of COF powders, the sonochemical method could also be applied to the preparation of COF films, where COF-5 particles formed in the reaction vessel were deposited onto the alumina surface by a fluid microjet formed by the rupture of bubbles near the solid surface, resulting in a uniform thin layer.

In 2019, Duan and co-workers prepared COF-5 with high structural porosity and strong adsorption capacity for CO₂ [70]. Then COF-5 was dispersed in Pebax-1657 to prepare a mixed matrix membrane with good CO₂ permeability. When the loading of COF-5 in the mixed matrix membrane was 0.4 wt%, the membrane reached the maximum CO₂ permeability and the selectivity for CO₂/N₂ increased from 31.3 to 49.3. The stability of the membrane was high, and the separation performance in the 120 h separating test remained stable.

To demonstrate the universality of this method, Cooper's group prepared a series of imine-linked COFs (Fig. 8) [71]. The reaction time was very short (< 1 h) and all these COFs exhibited good crystallinity and high porosity, even better than that of the COFs obtained by traditional strategy. Unlike previous sonochemical preparation method, this study used aqueous AcOH as the solvent, which is more environmentally friendly, compared to toxic organic solvents like 1, 4-dioxine. Subsequently, the COFs synthesized by sonochemical method was used as photocatalyst for the sacrificial hydrogen evolution from water, and its catalytic performance was more durable than that synthesized by solvothermal method. Beside COF powder, sonochemical method can also be utilized for the preparation of COF nanosheets which is usually obtained from the bulk COF by solvothermal method [72, 73].

Compared with the conventional solvothermal method, the COFs prepared by sonochemical method has a larger specific surface area, high crystallinity and high porosity, with shorter synthesis time. In comparison with the microwave method, ultrasonication is a more economical manner since no expensive equipment is required [74].

4.3. Mechanical method

Mechanical method is commonly used in industry, as mechanical force and transient frictional heating can drive or accelerate chemical reactions [75, 76]. It is simple and has the merits of easy operating, short synthesis time, while ensuring high quality of products. It is widely used in organic and inorganic synthesis, and has long been used in synthesis of porous materials such as zeolites, MOFs. Mechanical chemistry is recognized as one of the most suitable synthetic tools for the formation of covalent bonds, especially Schiff base condensation, with the potential to synthesize COFs. In 2013, Banerjee's group synthesized TpPa-1, TpPa-2 and TpBD for the first time by solvent-free grinding method at room temperature [77]. This process is simple and conducted by simply placing the reaction monomer in a mortar and grinding for 40 min (Fig. 9A). The resultant products show good thermal and chemical stability, and are stable in boiling water, acids, and bases (9 mol/L HCl and 3 mol/L NaOH). The crystallinity and porosity are slightly lower, and the morphological differences are obvious, as the COFs synthesized by the mechanical method have a graphene-like layered structure. During the grinding process, mechanical exfoliation of 2D COFs was also found, which has never been reported in COF materials before.

The solvent-free synthesis method mentioned above does have many advantages, such as simple operation, high product purity and without need for complex purification operations, which is greener and environmentally friendly. However, for COFs, crystallinity and porosity are important indicators affecting their applications. To overcome these shortcomings, the liquid-assisted grinding (LAG) method is explored by adding a trace amount of liquid to the mortar to promote the diffusion rate of reactants. LAG can not only speed up the reaction rate, but also improve the crystallinity of COFs.

In 2014, Banerjee's group used the LAG method to prepare COFs, and C₂+C₃ and C₂+C₄ were chosen to synthesize 2D COFs with a predetermined topological design. The hexagonal COFs LZU-1 (Fig. 9B), TpPa-1 and TpTh and the quadrilateral COF DhaTph were obtained [78]. This improvement accelerated the reaction rate and further enhanced crystallinity of the products. LAG method can also be extended to synthesize hydrazone-bonded COFs, and the synthesized COFs containing porphyrin building blocks with moderate crystallinity and porosity. Given the high yield and ease of operation, LAG is a promising strategy for industrial production.

Since the solvent-free grinding method is not generally universal. Liu and colleagues tried to use vapor-assisted grinding method for the synthesis of imine-linked COFs [79]. The monomer mixture was ground at room temperature for 1 h, and then placed in an organic solvent atmosphere for 48 h at 120 ℃ to obtain a red-brown product in 85% yield (Fig. 9C). The synthesized COF had similar physicochemical properties to those obtained by solvothermal method, and had a larger specific surface area than that prepared by direct grinding method. It is noteworthy that the monomer and solvent were physically separated and the whole transformation occurred in the solid phase during evaporation, which greatly reduced the contamination of solvent.

Different grinding methods has their own advantages. Solvent-free method is greener but limited to the synthesis of small amounts of COF. LAG has a stronger universality, but the usage of solvent reduces purity of the product. While vapor-assistance avoids exposure of the product to solvent, but it needs to be carried out under high temperature conditions, and some materials with poor thermal stability cannot be prepared by such method. Mechanical method enables the synthesis steps of COFs to be simpler, the most suitable strategy should be chosen according to the practical needs.

The COFs prepared by mechanical method usually have poor crystallinity and porosity, small specific surface area. Banerjee's group synthesized the bipyridine COF TpBpy MC and used it as a highly efficient solid electrolyte in fuel cells, pioneered its application (Fig. 10A) [80]. Compared with the same type of COF (TpBpy ST) synthesized by solvothermal method, TpBpy MC has better performance than TpBpy ST in actual operation conditions because of its low porosity, which effectively inhibits the fuel crossing in proton exchange membrane fuel cells.

Mechanical method is very suitable for industrial applications because of its simple operation, high yield, and excellent product properties. In 2017, Borchardt and colleagues synthesized a series of covalent triazine frameworks (CTFs) using solvent-free ball milling method through Friedel–Crafts alkylation reaction [81]. The reaction occurs between carbazole and melamine by adding AlCl₃ as activator and ZnCl₂ as filler in the ball mill (Fig. 10B) for grinding of 1 h. The obtained porous CTF has a yield of more than 90% and a maximum specific surface area of 570 m²/g. The versatility of this method was demonstrated by investigating other monomers with different geometries and different sizes, among which anthracene could reach high yields of 96%. This process is solvent-free, efficient and easy to control the reaction size.

Because of the efficient energy storage and charge transport properties of 2D COFs, exfoliating them into monolayer or multilayer covalent organic nanosheets (CONs) has attracted much attention [82]. In 2013, Banerjee et al. adopted a simple, and environmentally friendly grinding method to transform synthesized COFs (TpPa, TpBD, etc.) into CONs (Fig. 10C) [83]. PXRD and (flourier transform infrared) FTIR analysis confirmed that these CONs maintain their structural integrity throughout the delamination process and remain stable in water, acidic and basic media (e.g., parent COF). Moreover, these exfoliated CONs have a graphene-like layered morphology. The mechanical delamination method is more efficient than ultrasonic or layer growth methods because it does not require ultra-pure solvents or expensive support materials.

The mechanical method is able to significantly reduce the time required for COFs synthesis. However, due to the low crystallinity and poor porosity of the products as compared to microwave and sonochemical methods, researchers are yet to further expand the application of this strategy [84, 85]. Mechanical method allows for very simple operation, with more conventional devices and high economic efficiency, avoiding the use of large amount of hazardous solvent, making it suitable for large-scale production [82, 86].

4.4. Spraying method

Hao et al. prepared COF-LZU1/PAN composite membrane by spraying layer-by-layer self-assembly method (Figs. 11A and B) [87]. Firstly, sprayed p-phenylenediamine (PDA) and 1, 3, 5-benzenetrialdehyde (TFB) alternately deposited on the modified PAN membrane in stoichiometric ratios, and reacted with each other in the presence of acetic acid to form dense crystalline COF-LZU-1 film. Such strategy was efficient, taking only 12 h to prepare a 3-layer COF membrane with a thickness of about 300 nm and a BET area of 645.51 m²/g. The membrane was applied to the separation of dyes in wastewater and achieved 99.43% retention of aniline blue with a flux of 264.94 L m⁻² h⁻¹ MPa⁻¹. Over 90% retention was also achieved for other dyes such as acid orange red, chromium black T, and MB (molecular weight 350–800). The COF-LZU-1/PAN membrane has good solvent resistance, reusability and excellent long-term operational stability, proving the effective strategy of spraying layer-by-layer self-assembly for membrane preparation.

Luis et al. used a spraying method to prepare a 2D-COF (COF-TAPB-BTCA) (Fig. 11C) [88]. A three-fluid nozzle was used to atomize the aldehyde and amine solutions separately, and yellow powders of amorphous COF were rapidly produced. These powders were dispersed in a DIOX/mesitylene/water/HAc mixture and heated at 80 ℃ for 192 h to form hollow crystalline micro-spherical super-structures (Figs. 11D and E) with BET area of 911 m²/g and high thermal stability (up to 500 ℃). This method can also be used to synthesize other imine-linked COFs such as COF-LZU1 and COF-TAPB- PDA, illustrating a general applicability.

By simply adjusting the concentration and composition of the spray fluid, the properties and morphology of the COF product can be adjusted [89]. The spraying method is controllable and can form a uniform COF layer on various substrate surfaces, which greatly expands the application range of COF materials. Nevertheless, the binding force between COF layer and substrate prepared by the spraying method is weak, and it is necessary to build an additional protective layer.

4.5. Dip-coating method

Chen et al. developed a dip-coating strategy to synthesize imine COF on the surface of PAN nanofibers as shown in Fig. 12 [90]. The aminated nanofibers were prepared and then placed in a Teflon-lined reactor with a dioxane solution of aldehyde and reacted at 120 ℃ for 2 h to obtain PAN-Tp. Finally, the ligand solution was drop-coated onto the PAN-Tp surface to form PAN/TAPB-TPA nanofibers, and the thickness could be controlled by the concentration and the volume of monomer solution. By using the same method, HB PAN/TAPB-TPA-NH₂ nanofibers with a water contact angel of 167° were fabricated, exhibiting excellent oil-water separation properties, reusability, and recyclability.

Jun et al. also synthesized COF on the surface of zinc electrodes using the dip-coating method (Fig. 13) [91]. The polished zinc foil was immersed in a solvent dissolved with COF monomer, and a COF layer was generated on the zinc surface, which greatly alleviated the problem of easy formation of zinc dendrites on the electrode surface. Song et al. used the same method to synthesize imine-based COF on the surface of stainless-steel fibers for solid-phase microextraction [92].

The COF membrane constructed by dip-coating method grew in situ on the surface of substrate, so the binding force between the COF and the substrate is stronger. The dip-coating method is scalable, and would not be limited by the shape of the support, easy-operating and equipment maintenance cost is quite low. However, it is solvent-consuming, and the drying stage is easily affected by the environment, making it necessary to strictly control temperature, humidity, and other factors.

4.6. Casting method

Lambeth's group uses a casting method to prepare free-standing COF films with a uniform thickness directly from a solvent mixture containing monomers [93]. The monomer solution was poured onto a steel plate at 35 ℃, and the blade was set at a certain height and then the solvent is evaporated at 25% humidity to obtain a red film. This method is more efficient compared to the way of synthesizing COF powder first and then dissolving it for casting. Similar method could be used to prepare mixed matrix membranes, Wang's group prepared cationic TpEB-PAN membranes by adding the monomer pair and catalyst to the PAN solution [94]. The in situ grown TpEB COF was uniformly dispersed into the PAN solution, which was left to stand at 70 ℃ for 12 h to ensure the high crystallinity of COF. Then the solution was casted onto a PP nonwoven support using a casting knife. Finally, the film was obtained after solvent evaporation. This method also allows the production of large-size films, which has the potential for commercial application.

The casting method is scalable with high flatness and high uniformity of the produced membrane. The surface roughness of the substrate has limited influence on the product. The method has strong adaptability in viscosity and solid content of casting solution. Further, the membrane could be manufactured roll-to-roll, it is easy-operating, and cost low. The disadvantages are that it is difficult to avoid the use of a large amount of organic solvents, and the thickness of the prepared COF membrane is in micron level, which limits its application.

4.7. Other methods 4.7.1. Electron beam irradiation method

Wang's group prepared COFs by a unique electron-beam irradiation method [95], the radiation-induced synthesis of COF was carried out using high-energy (1.5 MeV) electron-beam irradiation provided by an electron accelerator (Fig. 14). 2, 4, 6-Tris(4-formylphenoxy)1, 3, 5-triazine (TPT-CHO) and 1, 3, 5-tris(4-aminophenyl)triazene (TAPT) were added to a 3 mL scintillation vial containing a mixture of 1, 2-dichlorobenzene/n-butanol/acetic acid solvent. The solution was sealed under nitrogen atmosphere and directly irradiated with an electron beam at a cumulative absorbed dose of 100 kGy for 160 s. EB-COF-1 was obtained in form of dense spherical particles with an average diameter of ~250 nm, and showed high thermochemical stability. The adsorbed doses determine the crystallinity, BET area, and thermal stability of the products. There is a poor crystallinity at low doses and a partial structural degradation at high doses. By balancing the radiation-induced activation and radiation-induced structural degradation of product formation, high-quality porous crystalline COFs can be prepared.

4.7.2. Salt-mediated crystallization method

Banerjee's group developed a novel and efficient salt-mediated crystallization method by mixing p-toluenesulfonic acid and diamine thoroughly, and then adding 1, 3, 5-tricarbonylresorcinol (Tp) and ~100 µL of water [96]. All materials were mixed completely until they were transformed into dough, which was further heated at 170 ℃ for 60 s. The resulting dark-red powder was immersed in 500 mL of hot water for 5 min to isolate the powdered COF with a separation rate of more than 90%. Researchers used the same method to synthesize 12 highly crystalline COFs in 60 s, with a specific surface area of up to 3109 m²/g, the highest of all reported 2D COFs. Interestingly, the doughy mixture of PTSA and COF monomers could be processed into various shapes (Fig. 15) and then baked to form COF sculpture with smooth and crack-free surface, which are structurally intact, having high crystallinity, high porosity as well as strong water absorption properties.

4.7.3. Linkage conversion method

Amide-bonded COFs have great potential for practical applications due to their high stability and variable structures, but direct preparation of amide-based COFs is difficult. Researchers have developed linkage conversion method to convert the easily prepared imine linkages into amide linkages, making it possible to efficiently prepare amide-based COFs.

In 2022, Zhao's group used KHSO₅ as an oxidant to convert imine linkages to amide linkages [97]. A glacial acetic acid was used as a buffer agent to convert imine-based COF to amide-based COF (AM-COF-QPTA-PDA) by mildly stirring a suspension of imine COFs and Oxone in anhydrous dimethylformamide for 5 h (Fig. 16). This synthesis strategy was simple and efficient. The resulting amide COF is more stable, further, researchers converted 6 other imine COFs to amide COFs by the same method, demonstrating its general versatility. This method could be applied to produce gram-scale COF and has the potential for mass production. However, the drawback is that the conversion leads to decrease in crystallinity and small loss of BET area.

4.7.4. Imine exchange method

Synthesis of imine COFs mainly relies on the condensation reaction between aldehydes and amines, achieving the transformation of COFs from amorphous to crystalline form, which often consumes more time to obtain highly crystalline framework. Zhao and coworkers developed a strategy based on dynamic imine exchange reaction for efficient synthesis of imine COFs at atmospheric pressure with scalability [98]. Firstly, a mixed solution of aldehyde and monofunctional amines (n-propylamine, n-butylamine, isobutylamine, n-hexylamine, dodecylamine or 2-aminopyridine) was stirred at high temperature for 30 min to obtain the imine precursor solution, after which the amine and catalyst solution was added to the above solution and heated to 180 ℃ under inert atmosphere and maintained for 6 h. The brown powder, that is, the product, was obtained with high crystallinity and high porosity. This method is expected to achieve a large-scale production of COFs.

Synthesis of large-size single crystal COF is helpful to the characterization of crystal structure of COF, but its synthesis is often time-consuming. Wang et al. divided the formation of COF crystal into two steps as imine formation step and imine-exchange reactions step (Fig. 17A) [99]. By using trifluoroacetic acid as catalyst and trifluoroethylamine as intermediate reactant, the obtain COF single crystals with a size up to 150 µm (Fig. 17B) in 1–2 days, which is quite efficient since it usually takes tens of days [100], and the XRD resolution of COF crystals is as high as 0.79 Å.

4.7.5. Phase switching strategy method

The Schiff-base COFs condensation reaction exhibits an unstable reversible state at low temperatures and a more stable irreversible state at high temperatures, this special property makes it possible to control the synthesis process of COFs by regulating temperature. In 2022, Jiang's group developed a phase switching strategy to synthesize imine COF membranes to realize the transformation from polymer to COF, which provides a new idea for COF synthesis [101]. The specific operation (Fig. 18) was to pour a mixed solution of amine and aldehyde monomer onto a carrier to generate a polymer film at 60 ℃. Next, the polymer film was placed into an autoclave that contained an organic solvent and catalyst but without contact between the solvent and the film. The catalyst was then volatilized at 145 ℃, resulting in a further reaction of the polymer film to form a COF film. The thin film had a high rejection (99%) of dyes (e.g., Congo red), metal ions (chromium) when subjected to nanofiltration experiments. Similarly, in 2015, Thomas et al. synthesized 2D COF by vapor-assisted conversion method [102]. High-quality COFs with tunable thickness were synthesized on different substrates by adopting different synthetic conditions, including BDT-COF and COF-5. BDT-COF films of few microns thickness exhibited mesoporous and structural porosity, while thinner BDT-COF films formed cohesive dense layers. In addition, researchers investigated the formation of COF-5 films with different solvent mixtures as the vapor-phase source. As for the above two studies, the former can enhance the order of COF materials and improve their crystallinity. While the latter research is suitable for preparing COF films on a fragile and sensitive substrate.

In terms of operation process, the phase switching method is similar to the dynamic imine exchange method, but there are essential differences in the purpose of the two methods. The dynamic imide exchange method is not to directly synthesize imide COF, but to synthesize other materials first, and then convert to imide COF through substitution reaction, with the purpose of making synthesis conditions milder and shortening reaction time. Meanwhile, the aim of phase switching method is to improve the disordered and loose structure of COF caused by concurrent polymerization and crystallization process. In practical application, above two methods could be selected considering different requirements.

4.7.6. Synthesis of COF foam

Banerjee's group developed an in situ foaming method to prepare COF foam [103], where the highly porous structure contributes to the ultrafast adsorption of various pollutants from water. Excessive PTSA was added to the monomer mixture to react with NaHCO₃ to generate bubbles, then the slurry expanded continuously during the crystallization process to induce ordered micropores in the disordered 3D structure, which was dried to obtain hierarchical porous COF foam. The appearance of the solid expanded nearly 15–20 times in volume compared to the corresponding COF powder at the same weight. In 2020, the same group used graphene oxide (GO) as a blowing agent to prepare COF foams as well [104]. GO induces the formation of disordered macropores and mesopores in COF-GO foams. The hierarchical inclusion of macropores employing GO as a foaming agent in a COF matrix, rendering crystalline and porous 3D printed COF-GO foam. The interconnected macropore matrix of the foam is decorated by fixed micropores that have excellent adsorption capacity for pollutants in water. In 2021, Huang's group synthesized self-supported monolithic COF foam using NaCl as a template [105]. PTSA and amine monomer were added to the mortar and ground thoroughly for 5 min, then a certain amount of NaCl was added for mixing. Subsequently, aldehyde monomer was added for grounding of 10 min, lastly a small amount of water was added to obtain COF paste, which was reacted at 170 ℃ for 5 min. The product was washed and dried after cooling to room temperature to obtain HHP-TpBD-X COF foam. The material is rich in macro- and mesopores with strong adsorption capacity to the sulfamerazine in aqueous.

5. Ambient synthesis strategy

This chapter presents a range of methods for the synthesis conducted at low temperature and atmospheric pressure, which reduces energy consumption, and circumvents the degradation of temperature-sensitive monomer at high temperatures. These methods are easy-operating with the potential to be applied to the large-scale production of COFs.

5.1. Novel catalysts

Generally, the synthesis of imine COFs is carried out under catalysis of acetic acid, which often requires high temperature and a long reaction time. Dichtel et al. developed a strategy to accelerate the synthesis of imine COFs at room temperature by using Lewis acidic trifluorometal salts as catalyst (Fig. 19A) [106]. XRD pattern indicates high crystallinity with a BET surface area up to 2175 m²/g. Scandium trifluoromethanesulfonate stands out among the many trifluorometallates because of its exceptional performance. Likewise, CTFs were also successfully catalyzed by trifluoromethanesulfonic acid, which also performed well [65]. Verduzco's group used transition-metal nitrates instead of acetic acid to rapidly catalyze the synthesis of imine COFs at room temperature [107]. It was found that, Fe(NO₃)₃·9H₂O has the best catalytic effect to produce the crystalline COFs in 10 min (Fig. 19B). This catalyst can also synthesize 2D imine COFs, 3D COFs and azide-bonded COFs with different structures. Hou's group uses plasma to induce polymerization of COF monomers [108]. Plasma possesses similar chemical properties to acidic catalysts and can therefore catalyze organic reactions. Liquid DBD (dielectric barrier discharge) plasma was used effectively for the catalysis of COF monomers polymerization at room temperature and atmospheric pressure. However, special plasma reactor equipment is required and only specific acid-activated organic reactions can be catalyzed. Dong et al. developed a photocatalyst to synthesize COF LZU-191 under nature-sunlight, with a scale up to gram level [109]. This approach offers a novel method for COF synthesis, and it has been demonstrated to be effective in synthesizing additional COF, like DADB-TBC—COF and DABD-PTA-COF, indicating its universality.

5.2. One-pot method

In 2015, Zamora et al. prepared RT-COF-1 by one-pot method, the reactant solution was stirred slightly for 1 min at room temperature to obtain colloid COF [110]. After washing and drying, a yellow solid product was obtained with yield of 96%. The solid-state ¹³C test, cross polarization/magic angle spinning nuclear magnetic resonance (CP-MAS NMR), FTIR and elemental analysis confirmed the production of COF-1. The BET surface area of RT-COF-1 was determined to be 369 m²/g through CO₂ adsorption and desorption experiment. The findings of this study allow the direct fabrication of micron/submicron COF patterns on rigid and flexible substrates through the aid of inkjet printing techniques, which is efficient, scalable and has great potential for application.

With the assistance of cationic and anionic surfactants, stable crystalline imine-based TAPB-BTCA COF nanoparticle hydrocolloids can also be prepared by the one-pot method [111]. The reaction is confined in micelles, and the particle size of COF particles are less than 20 nm, breaking the limitation to achieve nano-sized imine COF particles. The colloids could be moulded into any shape like films (Fig. 20A), and the film surface is very smooth (Fig. 20B). Interestingly, it can also be shaped into 3D octahedron (Fig. 20C) or printed on the substrate as shown in Figs. 20D and E, proving its good processability.

5.3. Room temperature batch and continuous flow synthesis

In 2016, Zhao et al. synthesized the classical imine COF-LZU-1 continuously at room temperature [112]. COF-LZU-1 was synthesized using a continuous flow system (Fig. 21A) in which solutions of aldehydes and amines were first prepared separately and then injected together into a flow reactor at flow rate of 20 µL/min and residence time of 11 s. The reaction was carried out at room temperature to give a yellow product. The production rate was up to 41 mg/h and the space time yield was up to 703 kg m⁻³ day⁻¹. Although the application of continuous synthesis is limited by some prerequisites such as the solubility of the monomer, this attempt is of great importance for the large-scale production of COFs for practical applications.

A similar approach was taken to synthesize the COF that can be used in inkjet printing technology by David and coworkers. The microfluidic device used in this method consists of four input channels and leads to a main microfluidic channel where the reaction occurs (Fig. 21B) [113]. The two monomers were injected into the middle B and C channels respectively, and the catalyst passed into the A and D channels to prepare the high crystallinity MF-COF-1 composed of fibrous microstructures.

6. Conclusion and outlook

We summarized the recent studies on the synthesis of COFs under green, high efficiency, low temperature, and other newly explored approaches in recent years. Compared with the traditional solvothermal method, the use of aqueous solvents, low-toxic organic solvents, ILs, DESs or solvent-free as reaction solvents to synthesize COFs can avoid the use of toxic solvents and achieve the purpose of green synthesis. Although some of the current synthesis strategies still have drawbacks such as long reaction time and high temperature, with the assistance of microwave, ultrasound, or mechanical force, the polymerization of COF monomers is accelerated and the synthesis rate is greatly improved, among which the microwave method can ensure high crystallinity of the product. The phase switching method enables the separation of the polymerization process from the crystallization process, which improves the order and the crystallinity of the product. Furthermore, this method can be employed to prepare thin films. In the case of the reaction between temperature-sensitive monomers, the development of novel catalysts allows polymerization reactions to occur at ambient temperature, thus avoiding the damage to the monomer structure. For fragile supports, the vapor-assistant conversion method could be employed. In this approach, the gaseous monomers are deposited on the support surface then polymerize to form a COF film without destroying the support. Moreover, the thickness of the film could be controlled.

For the further development in the field of COFs synthesis, the following problems still need to be considered: (1) Even though there are many new synthesis strategies of COFs, these methods have certain limitations and can only reach satisfactory results in the synthesis of one or several types of COFs, and their versatility needs to be further improved. (2) The cost of some methods is relatively high. For example, the use of ILs as green solvents to synthesize COFs reduces environmental pollution, but the high cost of ILs is a major obstacle to the expansion of production scale. (3) The use of organic solvents in post-processing steps after COFs synthesis should be avoided as far as possible. (4) Although each strategy has its own advantages, a synthesis strategy that combines the benefits of low reaction temperature, sustainability, and efficiency is currently lacking. (5) While some strategies improved the synthesis efficiency, they could not ensure that the crystallinity, specific surface area and porosity of the synthesized COFs reached the ideal standard.

Crystallinity is one of the key factors that distinguish COFs from other common porous organic polymers. Currently, COFs with high crystallinity can be synthesized mainly by slowing down the growth rate of COFs. In the future, studies on the formation mechanism can be strengthened, and a synthesis method with both short time and high crystallinity can be explored by optimizing the reaction conditions. Recently, there are also some studies on the rapid and efficient synthesis of COFs using microwave method, sonochemical method and other strategies, which not only improves the reaction rate, but also ensures the crystallinity of COFs. However, the investigation for more common and economical methods is necessary, and how to ensure high crystallinity in large quantities of preparation is also a problem to be solved.

COFs have been in development for nearly 20 years since their emergence, but cases of commercial application are still rare. On the one hand, how to mold the powdered COFs into a specific shape to meet the industrial requirement is still challenging. On the other hand, the commercially available reaction monomers for COFs are limited, and as a result, a complex monomer preparation step is even needed before the COFs synthesis, influencing the speed of exploring new types of COFs and their physicochemical properties. The cost of COF monomers is also relatively high as compared to other porous materials such as zeolite.

COFs materials are mainly used in gas separation and storage, energy storage and conversion, catalysis, etc. These applications have a long history of development, with very abundant and mature materials. Considering the unique architectural features of COFs, it is urgent to seek different application domains. For example, COF-based smart materials with various capabilities can be achieved based on the designability and easy functionalization of the COFs, and easy hybridization with other nanomaterials such as MOFs. Moreover, compared with common 2D COFs, 3D COFs have been less studied. With more complex pore structure and abundant active sites, 3D COFs have great application potential and deserve more attention.

Pore environmental and structural regulation is still one of the research hotspots and important directions in the COFs synthetic fields. Especially, more COFs with ultra-micropores (< 0.7 nm) and larger nanopores (> 5 nm) are in urgent need of development, because they would probably show exceptional performance in practical application. In the future, machine learning or artificial intelligence can be used to assist the COF preparation, including monomer screening, synthetic condition optimization, process assessment. We believe that with the steady advances in synthetic chemistry, more simple, green, and efficient approaches for COFs fabrication will be emerged, thereby promoting the progress on the large-scale and sustainable production of high-quality COFs targeting specific scenario.

Declaration of competing interests

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence that work reported in this paper.

CRediT authorship contribution statement

Yujie Wang: Writing – original draft, Visualization, Investigation. Haoran Wang: Writing – original draft, Visualization, Investigation. Yanni Liu: Writing – original draft, Visualization, Investigation. Manhua Peng: Writing – review & editing, Supervision, Funding acquisition. Hongwei Fan: Writing – review & editing, Supervision, Funding acquisition, Conceptualization. Hong Meng: Writing – review & editing, Supervision, Funding acquisition.

Acknowledgments

The work was financially supported by the National Natural Science Foundation of China (Nos. 22322801, 22108010, 22278124), and Fundamental Research Funds for the Central Universities (No. buctrc202135).