Some general trends on how the intrinsic properties of M/COF affect the 3D monolith have been discovered. The size of M/COFs in the paste has less obvious influence on the viscosity, but a higher amount of M/COFs will cause an increase in viscosity due to the granular feature of the M/COF particles. It has been found that the viscosity increases by more than twofold when the concentration of M/COF is increased by a factor of 1.5 (Fig. 3c). In DIW process, the high viscosity, which causes non-continuous flow, will then lead to a higher pressure during printing and difficulty in replicating the designed model. As shown in Fig. 3a, d discontinuity of extrusion occurs during 3D-printing when the concentration of the MOF increasing from 31 to 51 wt% with a 1.5 fold increases in the viscosity [13]. Moreover, due to the non-adhesive nature of M/COFs, high content of M/COFs will cause a lack of a crosslinked network, which greatly reduces the mechanical strength of the 3D-printed monoliths and makes the printed macrostructure difficult to maintain [31]. The variation in M/COFs types can also cause significant differences in the viscosity. Generally, anisotropic M/COFs, such as 2D M/COFs, are easier to print because during printing, anisotropic materials become oriented in the flow direction under high shear rates, which would minimize the intrinsic viscosity and increases the maximum packing density [13].

The common additives used in the mixed M/COF pastes include organic compounds [13] and ceramic clay [35], such as boehmite, bentonite, PVA, cellulose, F127, and gelatin, while most bonded M/COF pastes only utilize certain polymers, such as cellulose and PA12. This is mainly because the polymer surface has more freedom to be modified for better crosslinking between additives and M/COFs. In general, due to the high flexibility of the polymer, 3D monoliths built using polymer additives normally exhibit large flexural modulus, such as the as-printed 3D ZIF-8/Semiflex [36]. The tensile strength of the 3D-printed M/COF monoliths can also be adjusted through modification on the polymer additives. As shown in Fig. 3e, the 3D-printed MIL-53 monolith with various acrylate-based organic additives delivers different tensile strength [37]. Although the paste with ceramic clay cannot offer high flexibility, its usually has a large compressive strength. In Belmonte et al.’s work, for example, the MOF/boehmite has a compressive strength of around 1 MPa [35]. If changing the boehmite to a mixture of bentonite and PVA, the compressive strength then decreases to 0.48 and 0.56 MPa for the 3D-printed MOF-74(Ni) and UTSA-16 [38]. Normally, the compressive strength can be enhanced by raising the amount of additives. However, a high amount of additives in the monolith will inevitably cause a decrease in the surface area due to their intrinsic low surface area [38,39,40]. The problem is more serious for polymer additives, which are more likely to cover the surface of M/COFs and block their pores than the ceramic clay, leading to destruction of the intrinsic pore structure [41]. For example, the work using a high PVA content of around 20 wt% can achieve 1.7 MPa compressive strength, but it causes a drop in surface area of around 26% [40]. Besides the difference in mechanical properties, the pastes using polymer additives and ceramic additives also have other distinct properties. Firstly, the amount of polymer additives in a printable paste is normally higher, which leads to a lower content of active M/COFs. The M/COFs content in the paste with single polymer additive is often less than 30 wt%, while by using bentonite ceramic clay, the MOF (ZIF-8) content can reach 66.7 wt% [42]. Secondly, in some cases, such as 3D printing ZIF-7, the PVA binder cannot be used to form a printable paste, but using silica as the additive can lead to a good printability [40]. In addition to pastes with specific degree of viscosity, materials exhibiting sharp is situ transition from the fluid to gel state are also commonly used for extrusion. This type of transition is normally based on either chemical reaction between two components or intrinsic physical properties, such as temperature-triggered gelation [22]. M/COFs with additives that has thermal response rheological properties can be easily printed, such as F127, polyurethane-gelatin (PUG), and the thermoplastic binders [25,43]. The sol-gel transition of PUG with varying temperature is shown in Fig. 3f. However, in order to maintain the transition behavior, the M/COF content cannot be high. For example, the maximum loading of the MOF in ABS is only 10 wt%; otherwise, the filament will become too brittle to maintain the structure integrity [44].

Besides developing pastes with improved printability, higher M/COFs loadings, and enhanced mechanical properties, attention must be given to specific application requirements when designing the paste. For instance, in the biomedical applications, the paste should be biocompatible or biodegradable [25]. Commonly used additives for such applications include polycaprolactone (PCL) [45,46], PUG [25], poly(lactide-co-glycolide) (PLGA) [47], sericin [28], etc. Although the paste with these additives can be directly printed alone, they often exhibit rather low mechanical strength. In order to form a useable monolith, the M/COF content is generally low for the bioinks, typically less than 10 wt%. The highest recorded content of M/COF in a 3D-printed biocompatible monolith is around 13 wt% using gelatin [48]. Therefore, further improvements are needed to enhance the current M/COF bioink. Additionally, the compatibility of each component should be considered. For example, for the water sensitive M/COF, non-aqueous solvent should be used. In the work by Grande et al., for example, IPA was utilized with hydroxypropyl cellulose and boehmite as the additives to prevent the degradation of UTSA-16 MOF [49]. This, in turn, limits the selection of some conventional components. Hence, the development of a new paste formula would be necessary to achieve an ideal monolith with sufficient mechanical strength, superior printability, high surface area, and suitable for required occasions.

Using mixed additives can be a better option that can result in more advanced properties compared to using the single additive. For example, the inclusion of bentonite binder can enhance the rigidity of 3D-printed structures, while employing an organic binder allows fine-tuning of the rheology of the 3D-printable paste. When using a mixed binder of bentonite and PVA, the loading of MOF can reach as high as 85 wt% [38]. The combination of different polymer additives can also result in better printability with a reduced amount of usage. Compared to the common single-phase organic binders for 3D printing, such as PVA, Pluronic F127, PEO, and polyvinylpyrrolidone K25, a combination of 2-hydroxyethyl cellulose (HEC) and PVA as a mixed-additive was only required for less than 10 wt% to achieve a 3D-printed monolith, due to its high viscosity, which maximizes the porosity [13]. Mixed additives can also enhance mechanical properties. For instance, a combination of various biocompatible polymers, such as PLGA, PVA, PCL and collagen, can improve the mechanical properties of the current bioinks [46]. Besides, in the other work, the additional of EB in the paste has also demonstrated an increase in the storage modulus. The glass transition temperature of the paste with EB is also increased, making the printing process easier [50].

New types of paste compositions have also been developed to solve the problems of unfavorable paste rheology, limited printability, and weak mechanical properties, which common M/COF pastes normally exhibit. One work has demonstrated the use of liquid binder-polyamide(imide) (Torlon), which greatly affects the rheology of the paste by reducing the viscosity and decreasing the required pressure for printing (Fig. 3g) [39]. The additives with dual-functions, which can not only adjust the rheological properties but also help to improve the performance, have also attracts much attentions [51]. 2D Graphene is one example that can be used to facilitate ink extrusion and maintain the printed shape; at the same time, it can increase the electrical conductivity [52]. Other pastes that exhibit a promoted interactions between the components were also explored for improving the mechanical properties of 3D-printed monoliths. For example, graphene oxide has been investigated as an additive, which can form intermolecular hydrogen bonds due to the presence of several donor−acceptor hydrogen-bonding sites (− COOH, − OH, − epoxy) at the edges and on the basal planes. To improve the mechanical properties of 3D-printed monolith using gelatin paste, adding metal ions in the gel paste, such as Ca²⁺, can trigger the crosslinking and work as a filler to enhance the brittleness [48].

To meet the requirement of the de-binding process, additives that can be eliminated without damaging the M/COF are also widely used. If the additive has higher solubility in a certain solution in which M/COF can maintain its structure or a melting temperature lower than the M/COF decomposition temperature, it can be removed in the post-treatment. For example, by treating the 3D-printed ZIF-8/PVDF monolith in hot acetone, PVDF binder can be dissolved with limited destruction on the ZIF-8. The ZIF-8 particles will then appear intertwined through the remaining polymeric web at the surface and are loosely packed in the interior, allowing sufficient exposure. The surface area of the post-treated monolith can reach 706 m² g⁻¹ [36]. The use of thermally stable UiO-66 MOF and a commercial polymer binder can also achieve a BET surface area of 633 m² g⁻¹ after the removal of binder by heating. However, the de-binding can cause a fall in the mechanical strength as well. After decomposing the binder by heating, the mechanical strength of 3D-printed UiO-66 is reduced from 22.4 to 4.9 MPa [53]. One other example of the COF-GO monolith containing removable PTSA additive also demonstrated that the original hard and heavy monolith became soft and lightweight, once the PTSA is removed [31].

In addition to the post-treatment, adjusting the paste composition to achieve a pure M/COF paste can also reach the goal of eliminating the additives. Developing an additive-free 3D-printed M/COFs without compromising their intrinsic properties eliminates potential blockage or dead mass in the monolith, allowing a perfect combination of the advantages of designed macrostructure and the synthesized microstructure. For example, Wang et al. have developed a facile method to synthesize pure MOF gel based on fast nucleation and slow growth of MOF particles using a high concentration of MOF precursors. The MOF gel paste can be easily extruded and shaped into different architectures due to the strong inter-particle forces that maintain the structure. The as-synthesized MOF gel exhibits a much larger surface area than the gel synthesized by traditional methods, but it still cannot match the surface area of MOF powder [54]. Inspired by the self-crosslinking of M/COFs, the precursors of M/COFs can be made into the paste instead of using the M/COFs themselves and the crosslinking is triggered after 3D printing. Kim et al. demonstrated the use of a liquid ink with 1.1 mM Cu(NO₃)₂·3H₂O and 0.6 mM trimesic acid in 1 mL of dimethyl sulfoxide (Fig. 3h). The solvent quickly evaporates once the filament is extruded from the micro-sized tip, and the MOFs crystallization occurs. The HKUST-1 made using solution-mediated crystallization without the use of additives shows 1192 m² g⁻¹ BET surface area, which is superior or comparable to the values obtained using other 3D printing approaches [55]. In another work, Ke et al. mixed the COF precursors with F127 at the initial stage. The imine- and β-ketoenamine-based COFs were polymerized in the presence of F127 template. The polymerization degree of the COF is purposely limited to prevent the formation of large COF particles and hindrance of the 3D printing. The amorphous COF is then heated for further crystallization, and the F127 is removed after the 3D printing, resulting in a binder-free 3D COF monolith (Fig. 3i) [56].

In addition to the paste composition, the 3D printing process can also significantly affect the printability, mechanical strength, and texture properties. To maximize the loading and intrinsic properties of M/COFs, considerable efforts have been made to improve the current printing methods in the pre-printing, in-printing, and post-printing process.

One major improvement direction in the pre-printing modification is to functionalize the M/COFs or additives to enhance their interaction, which has been well studied for preparing bonded M/COF pastes. Promoting the bonding between M/COFs and additives in the paste will significantly enhance the dispersion of M/COFs, preventing M/COF particles from aggregating together. Several methods have been tried to functionalize the M/COFs to achieve bonding. For example, MIL53(Al)-NH₂ was functionalized with methacrylic moiety to bond with the commercial photopolymerizable acrylic oligomer and create a 3D printable paste (Fig. 4a) [37]. This strategy can enable a 12 wt% solid content of MOFs in the final 3D-printed monolith. Additionally, additives with specific functional groups can be chosen or synthesized to attract the M/COFs or their precursors for M/COF growth. Carboxylic and hydroxyl groups are well-studied examples of functional groups that can coordinate with the Zn²⁺, ensuring the crystal formation of ZIF-L on the organic compounds (Fig. 4b). The 3D-printed ZIF monolith using CelloZIF-L inks prepared by this method achieved a high ZIF content of 84 wt% [57]. The other work of 3D-printed CelloZIF-8 exhibits a high specific surface area of 900 m² g⁻¹ [34]. For SLS 3D printing, since the M/COF is distributed throughout the whole filament, the amount of exposed active M/COF will be limited. Weidner et al. fabricated a MOF-ABS filament for FDM 3D printing. The overall volume fraction of HKUST-1(Cu) is 1%; however, the volume fraction of Cu elements with contact to the outer surface was only 0.036%. This indicates that only a small amount of the HKUST-1 is located on the external polymer surface [58]. Considering this problem, a more advanced modification on the MOF-PA12 paste has been investigated by growing the ZIF-67 on the surface of PA12 particles (Fig. 4c). Although the loading of ZIF-67 in the monolith is only 2.6 wt%, it enables a better dispersion of ZIF-67 crystal and sufficient exposure [59].

During the 3D printing process, the presence of large M/COF particles can lead to favorable shear-thickening behavior, resulting in slower printing speeds. Moreover, certain solvents required for printing can also cause decomposition of some M/COF particles, necessitating lengthy solvothermal growth to regenerate them. To address these issues, the gel-print-grow method has been developed (Fig. 4d) [60]. In this approach, MOF precursors are first prepared into a gel for 3D printing, which helps avoid the problems associated with the large particle size and MOF decomposition in solvent. The MOFs can then be formed in situ by applying heat, removing the solvent, and activating the MOFs, which simplifies the optimization of printing parameter and benefits for universal 3D printing of M/COFs. The resulting activated monolith achieved 83% of the powder’s N₂ adsorption capacity (Fig. 4e). Similar insights have been applied in microfluidic 3D printing (Fig. 4f) [51]. For example, in the case of black phosphorus (BP), cobalt source, and 2-MIM, added into a microfluidic device, the MOFs can be rapidly formed on the BP during printing, enabling a smooth printing process. Furthermore, strategies to control the exposure of M/COF to increase the active area have been investigated. Wang et al. applied an acoustic field to a liquid photosensitive paste containing MIL-53(Fe). The MIL-53 was patterned into parallel lines, and the stereolithography was then used to prepare a polymer film in situ. This method allowed for better control of the dispersion of MOF particles in the polymer film (Fig. 4g) [61].

Process parameters also play a crucial role, especially in SLS 3D printing. The layer formed by the laser consists of a densely sintered top portion and a partially sintered porous bottom portion. It would be essential to strike a balance in laser power to avoid degradation of the polymer, leading to reduced mechanical strength, or insufficient power, resulting in a weakly sintered film. Therefore, carefully reducing the sintering power and time to achieve minimally sintered top portion with more voids while maintaining sufficient mechanical strength is ideal [62]. By optimizing the printing conditions in SLS, it leads to the formation of a solid, porous, powder-bed-like object, where the sintered polymer particles retain their particle-like appearance and contain accessible voids between them. The functional additive mixed with the printing matrix is attached only on the surface of the sintered particles, allowing for interactions with the fluid passing through the material [15].

Controlling the post-treatment is also important for enhancing the properties of 3D-printed M/COF monoliths. Indeed, the drying process is a critical step in the post-treatment, and it has been found that slow drying at lower temperatures, allowing for shrinkage during drying, leads to a denser crystal packing and higher mechanical strength [13]. However, a nonuniform shrinkage rate in the monolith can result in cracks and compromise structure integrity. To address this issue, Wang et al. has utilized a porous drying substrate to enable a uniform drying speed (Fig. 5a, b). This approach reduces the build-up of stress in the 3D printing monolith during the drying process, which significantly helps to maintain the structural integrity of the 3D-printed monolith with even no additives in the paste. As a result, the as-fabricated additive-free 3D monolith shows almost similar surface area, pore size distribution, and gas separation performance as the initial powder sample (Fig. 5c) [63]. Post-treatment can also be used to modify the M/COFs after shaping them into a 3D monolith. One example is post-impregnation, which introduces guest molecules into the M/COF after 3D printing. For example, Rezaei et al. impregnated TEPA and PEI into MIL-101 by immersing the 3D-printed monolith in the corresponding solution (Fig. 5d). The post-impregnated monoliths exhibited a fourfold increase in surface area and pore volume compared to their pre-impregnated counterparts, which could well be attributed to the reduced amine-pore diffusion (Fig. 5e) [64].

Various materials can be used to fabricate the 3D scaffolds, each tailored to meet certain specific requirements, such as the high flexibility, biocompatibility, high stiffness, or high electrical conductivity [74,75,76]. However, caution is needed when selecting the materials, as the M/COF precursor solution may cause the degradation of scaffolds under certain conditions, such as in a too acidic environment or due to high water solubility of the scaffolds, resulting in a reduction of mechanical strength [75]. In the work by Zhang et al., for example, silica was chosen as the skeleton material due to its resistance to degradation in organic solvents and at high temperature, which makes it a suitable alternative to ABS or PLA [77]. Similarly, the use of kaolin-based support instead of silica and bentonite clay was preferred, as the latter materials are prone to disintegration upon exposure to the synthesis liquor [78]. By carefully selecting compatible materials for 3D scaffolds, one can ensure the stability and mechanical integrity of the final monolith.

To enhance the M/COFs loading and create more surface area in the 3D skeleton, etched skeletons have been widely employed. For example, Wang et al. etched the skeleton in the KOH solution to remove the SiO₂ nanoparticles on the surface and generated more pores in 3D SiO₂ skeleton [79]. Similar approaches have also been taken, such as removing the PLA polymer using DMF or ABS using acetone for 3D PLA scaffold [67,80]. In addition to etching the skeleton to create more growing sites, modifying the skeleton surface to provide stronger adsorption and improve the M/COFs growth is also effective to enhance the M/COFs loading. Polymer supports are particularly advantageous for improving M/COF growing sites, as their surface can be easily modified and functionalized by various methods [81]. For example, the existing carbonyl group and carboxyl group in PLA can efficiently bind with Cu atoms in Cu-MOFs through the coordination process, producing Cu-MOFs/PLA films [73]. Another approach involved forming a polydopamine (PDA) layer on a etched 3D SiO₂ skeleton to facilitate the encapsulation of MOF particles within the hierarchical structure [79]. These methods effectively provide more surface area on the skeleton and enhance the M/COFs loading, thereby optimizing the performance of the 3D-printed M/COF monoliths.

A number of research works are devoted to developing advanced M/COFs deposition methods for 3D M/COFs-covered monolith. These methods can be categorized into two types: directly coating M/COF particles and growing M/COF particles, as shown in Fig. 6a. In the direct coating approach, M/COF particles are synthesized and then loaded onto the surface of the 3D-printed scaffold separately; while the growing method entails immersing the 3D monolith in a precursor solution as well as allowing M/COFs to nucleate and grow on the surface of the scaffold. Both of these methods offer unique advantages and have been explored extensively to achieve more advanced 3D-printed M/COF monoliths.

The direct coating approach offers a simple and straightforward method for creating 3D-printed M/COFs-covered monolith. In the traditional preparation process, a M/COFs dispersion is used to create a M/COFs ink, which is similar to but much less viscous than the ink used in the preparation of 3D-printed M/COFs-mixed monoliths. The M/COFs dispersion is then either dipped onto the 3D lattice [82,83] or the lattice is immersed in the M/COFs dispersion to produce the 3D-printed M/COF-covered monolith [84,85]. The adhesion between the M/COF and the skeleton is achieved by using the additives, such as Chitosan [84,86], and PVDF [82]. However, a major challenge is the potential weak adhesion between the M/COFs and the monolith, which can lead to poor stability in applications. To improve adhesion, Lu et al. have developed a PDA-hybridized nano-sized ZIF-8 (Fig. 6b). The catechol motifs on the ZIF-8 surface enable versatile adhesiveness, which is similar to the strong adhesion between mussels and various surfaces due to the abundant amount of L-DOPA contained proteins [87]. Furthermore, surface modification of the 3D skeleton can be employed to attract M/COF more strongly. Jin et al. for example, have developed a plasma activated 3D-printed skeleton, where the surface is functionalized with hydroxyl or hydroperoxyl groups (Fig. 6c). The modified Fe-MOF with PAA shell can then be bonded to the surface through coordination bonds, significantly enhancing the loading and dispersion of Fe-MOF [81]. In another study, a combination of plasma-treated 3D-printed SBS and mussel-inspired polydopamine-decorated ZIF-8 was used (Fig. 6d). This resulted in firm deposition of ZIF-8 on the SBS surface [88]. These innovative approaches are promising for achieving strong and stable M/COFs coverage on 3D-printed monoliths.

Numerous works have also demonstrated the immersion of 3D skeleton in M/COF precursor solutions for growing M/COFs on their surfaces, such as Zn/Co-MOF@TCP [75], ZIF-L MOF@PA [89], Co-,Cu-MOF@SiO₂ [68,77], and PCN-224 MOF@Ca₂SiO₄ [74]. By repeatedly immersing a 3D monolith in the precursor solutions, the M/COF loading can be controlled [73,90]. In addition, the concentration of the precursor can also affect the loading of the M/COF [91]. However, it should be noted that secondary growth of M/COFs on the 3D-printed M/COFs monolith can lead to a significant drop in compressive strength, imposing a limitation on the M/COFs growth [39]. To facilitate the uniform growth and achieve higher loading, the skeleton should provide sufficient nucleation sites. Some materials, like the alginate matrix, can effectively coordinate the ligand, facilitating M/COFs growth on their surface [69]. For instance, one work has prepared a mixed surface by dip-coating the 3D skeleton in a bovine serum albumin (BSA) solution. The BSA forms a bio-surface which provides both anchoring and nucleation sites to initiate ZIF-8 growth (Fig. 6e) [92].

For skeletons with inert surfaces that do not readily facilitate M/COFs growth, modifications are necessary. For instance, one work electro-oxidized the 3D-printed carbon to introduce COOH functional groups on the surface for growing COFs. This step provides activated carboxyl groups that enable functional COF materials with numerous amino groups to cross-link on the monolith surface through covalent bonding (Fig. 7a) [93]. In some other cases, a small amount of M/COFs can also be deposited as seeds on the surface to facilitate the subsequent M/COFs growth. For example, 3D-printed Ti₆AlV₄ was sonicated in a 1 wt% ZIF-8 solution and then immersed in the ZIF-8 precursor solution for hydrothermal synthesis to achieve a rather uniform ZIF-8 coating (Fig. 7b) [94]. Sometimes, it can be very challenging to form a uniform seeding on complex 3D structures. To tackle this problem, advanced deposition methods, such as atomic layer deposition (ALD) can be used to ensure a uniform distribution of M/COF on the surface of skeletons. The ALD-deposited ZnO on ABS, for example, exhibits a uniform distribution (Fig. 7c). After conversion, homogenous and continuous ZIF-8 crystal layers are present both on the outer walls and inner areas of the 3D-printed monolith [71].

Another approach involves mixing one of the M/COF precursors in the paste first and then reacting it with the other precursors to form M/COFs by immersing the 3D-printed monolith into the solution containing the other precursors. This method allows the growth step to occur in the inner part of the monolith, increasing the effective M/COFs loading and achieving more homogeneous composites. For instance, Ca-contained or Co-contained monolith 3D-printed monolith can be immersed in a linker solution [67,78], and the linker-containing 3D monolith can be immersed in lanthanide-containing solution [69]. In addition, researchers have also explored the incorporation of metal oxide as metal precursors in the paste. For example, ZnO-ABS filament was fabricated, and the MOFs crystal growth was conducted during the post-printing treatment. This approach has been proven to offer higher performance compared to directly making Zn-MOF filament for 3D printing Zn-MOFs (Fig. 7d) [72]. In the case of an in-situ grown HKUST-1 monolith using this method, the MOFs was observed to grow all over the 3D-printed skeleton and even within its interior (Fig. 7e). The HKUST-1 crystal exhibits a higher density and smaller size (approximately 5.5 μm) near the surface, while the crystals were bigger (around 10 μm) toward the interior. This difference in density and size distribution is likely due to variations in the ligand concentration and distribution of Cu²⁺ ions, as well as the limitations imposed by diffusion-limited growth in the bulk [70]. This represents an advantage over M/COFs-polymer composites prepared through direct blending methods or surface in-situ growth techniques.