Transferring Porosity Across Physical States Using Metal–Organic Cages: Porous Liquids, Glasses, Rubbers, and More

Abstract

The development of porous materials that retain tailored porosity across different physical states beyond crystalline solids (e.g., liquid or glassy) could yield new functional materials for diverse applications, yet it remains challenging. To address this, researchers have turned to discrete porous cages such as metal–organic cages (MOCs) and metal–organic polyhedra (MOPs). The organization and physical state of these materials are governed by inter‐cage interactions that can be modulated without altering the intrinsic porosity of the individual cages. In this minireview, we highlight how the peripheral functionality of such cages governs their interactions and physical state and explain how it can be harnessed to preserve and transfer porosity across distinct physical states, including liquids, glasses, and rubbers. We conclude by outlining emerging properties and potential applications for the resultant unique porous states.

Discrete porous metal–organic cages provide a platform to extend porosity beyond crystalline solids. Their peripheral functionality governs inter‐cage interactions and physical state, enabling preservation of intrinsic porosity across liquids, glasses, and rubbers. This minireview explores how such control unlocks unique porous states with emerging properties and applications.

1. Introduction

Permanent porosity is traditionally associated with crystalline solids. In extended materials such as metal–organic frameworks (MOFs), pore formation is driven by ordered lattices, which enables precise tuning of properties and the establishment of well‐defined structure–property relationships. However, transferring these desirable features into more processable states of matter, such as amorphous solids, glasses, or liquids, is not trivial. For instance, although MOF glasses and liquids have been successfully synthesized, the amorphization, vitrification, or melting of MOFs via mechanical or thermal treatment often disrupts their coordination networks, leading to a loss of structural integrity and of porosity.^{[ 1} , ^{2 ]} This structural disruption arises because the molecular mobility required to access these alternative phases compromises the coordination bonds, resulting in unpredictable transformations.

In the above context, discrete porous units, such as metal–organic cages (MOCs) or polyhedra (MOPs)—which we hereafter refer to collectively as cages, for simplicity—emerge as promising alternatives. Unlike extended frameworks, these systems decouple porosity from lattice stabilization. Cages are macromolecules stabilized by strong intramolecular coordination bonds, whereas their physical arrangement is governed by weak intermolecular interactions. Crucially, the physical state of these cages depends on their molecular motion, which is controlled by intermolecular forces that can be tuned without affecting the integrity of the internal cavity. Thus, cages are uniquely positioned to deliver designed, permanent porosity across various physical states, from crystalline solids to amorphous glasses, rubbers, and even liquids.

Despite their promise, achieving precise control over the physical state of cages remains challenging due to the complex and non‐directional nature of inter‐cage interactions. In this minireview, we identify surface chemistry as the crucial parameter that governs the physical state of cages and present selected examples that reveal how subtle changes in surface functionality can dramatically influence phase behavior and therefore be used to transfer porosity across different physical states. Accordingly, we focus exclusively on materials composed solely of monomeric cages, excluding systems formed through linkers, co‐crystallization agents, or other additives, which have been extensively reviewed elsewhere.^{[ 3} , ⁴ , ⁵ , ^{6 ]}

We begin by examining the structural factors that drive crystalline packing of cages in the solid state. We then explore how surface functionalization can modulate the interactions between cages, enabling phase transitions and stabilizing amorphous or dynamic phases such as glasses, rubbers, and liquids. In each case, we assess the extent to which porosity is retained across different physical states. Finally, we offer our perspective on how design principles rooted in cage surface‐chemistry can enable the development of next‐generation porous materials with permanent, tuneable porosity across exchangeable physical states to open new avenues for processing, functionality, and applications.

2. Amorphous versus Crystalline Solids

Cages are macromolecules whose dimensions surpass the sub‐nanometric range. At this scale, surface energy increases, and intermolecular dispersion forces are exacerbated,^{[ 7 ]} promoting the coalescence of cages. Typically, the non‐directional nature of these interactions, combined with fast nucleation kinetics, would lead to precipitation of cages as amorphous, metastable phases. However, solvent molecules can mediate this process by stabilizing ordered, crystalline arrangements. Thus, as‐made crystalline cages generally contain solvent molecules, which are crucial not only for maintaining structural integrity but also for directing the final packing phase.^{[ 8 ]} Consequently, solvatomorphism is commonly observed in cage systems.^{[ 8 ]}

Although solvent molecules aid crystallization, they simultaneously occupy the intrinsic pores of the cages and coordinate to the axial sites of the metal centers, thereby blocking porosity. Achieving permanent porosity in solid‐state cages thus requires removal of these solvent molecules, typically by heating and/or by applying vacuum. Upon desolvation, the resulting solids are dominated entirely by intermolecular cage–cage interactions, which often lack the directionality and strength required to preserve the long‐range order or rearrange in a new crystalline network. Consequently, upon desolvation, most cages undergo amorphization.^{[ 5 ]} While the intrinsic porosity of cages is often retained, the accessible porosity in the amorphous state may be lower than expected due to hindered diffusion pathways that affect both the kinetics and overall gas uptake.

Given the aforementioned limitations, researchers have devoted major efforts to obtaining desolvated cages that retain both the permanent porosity and the crystallinity of the corresponding solvated cages (Figure 1). Current strategies converge on the idea of enhancing directional and robust cage–cage interactions in the solvated state, such that the packing is preserved upon solvent removal.^{[ 9 ]} In cages lacking external functional groups, stabilization is primarily dictated by van der Waals forces between aromatic moieties of the ligands. Importantly, these van der Waals forces primarily dominate once cage‐solvent interactions have been sufficiently weakened. Thus, despite ordering retention depends on the nature and desolvation kinetics of the solvent, as well as of ligand flexibility (slower rates enable gradual reorganization, while flexible ligands confer dynamic tolerance to desolvation),^{[ 10 ]} the efficiency of these interactions is largely governed by the polyhedral shape of the cage as certain shapes expose aromatic faces more effectively, promoting tighter packings.^{[ 11 ]} For instance, Lantern^{[ 8} , ^{12 ]} tetrahedral‐,^{[ 13 ]} and octahedral‐type^{[ 14 ]} cages—characterized by low‐faceted polyhedral geometries and surface‐exposed aromatic panels—typically display higher crystallinity after activation compared to more complex shapes, such as cuboctahedra, where tilted aromatic panels hinder directional inter‐cage interactions.

Figure 1.

Overview of the factors governing crystallinity retention in cages after solvent removal. Crystallinity can be retained either by minimizing cage‐solvent interactions (through solvent exchange or thermal equilibration) or by enhancing cage–cage interactions or via mutual coordination.

The above principle has been elegantly demonstrated by Lee and co‐workers through solvent exchange protocols, using apolar, non‐coordinating solvents such as mesitylene and p‐xylene (Figure 2a).^{[ 15 ]} In a family of Cu(II)‐paddlewheel‐based cages functionalized with ─OH and ─NO₂ groups, solvent exchange gradually promoted denser packing, thereby reinforcing cage–cage interactions. Consequently, after desolvation, crystallinity was retained, which translated into a significant increase in porosity. For instance, the BET surface area (S_BET) of [Cu₂₄(IPA)₂₄] (IPA: isophthalate) increased from 123 to 1346 m²·g⁻¹ after solvent exchange and activation, a dramatic improvement attributed to the preservation of diffusion pathways that would otherwise collapse due to amorphization (Figure 2b).

Figure 2.

Methodologies employed to increase inter‐cage interactions and obtain robust crystalline cage arrangements with permanent porosity. a) Solvent‐exchange using non‐coordinating solvents followed by activation (right) versus conventional activation (left). b) N₂ adsorption isotherms at 77 K of treated versus untreated cages. c) Maximization of π–π interactions between triangular cage faces (right) by employing naphthalenediimide linkers (left). d) Representative N₂ adsorption isotherm at 77 K of the crystalline solid versus the amorphous solid, revealing the superior porosity of the former. Adapted from Refs. [15] (a,b) and [16] (c,d).

Building on the aforementioned concept, Tokuda and Furukawa have developed an open‐framework structure based on octahedral M(II) paddlewheel‐based cages having the general formula [M₁₂(NDI)₁₂(solvent)₁₂, where NDI denotes naphthalene‐diimide dicarboxylate linker, and M is either Cu or Rh (Figure 2c).^{[ 16 ]} They exploited the large, exposed aromatic panels of the triangular faces of the octahedral cages to drive the assembly through van der Waals interactions, with energies reaching up to 433 kJ·mol⁻¹, comparable to coordination bonds, per triangular faces pair interaction that rise from the synergistic interaction through to 180 atoms. The resulting diamond‐like networks featured significant extrinsic porosity, accounting for ca. 35% of the total crystal volume. Notably, these van der Waals interactions were sufficiently strong to stabilize external void spaces within the framework, in contrast to most crystalline cage packings, which tend to adopt dense packings with minimal extrinsic porosity. Accordingly, the open frameworks formed by this strategy exhibited the highest S_BET reported for cage‐based assemblies, reaching 1992 m²·g⁻¹. The emergence of such van der Waals‐driven porous architectures—dominated by face‐to‐face interactions—strongly resembles the solid‐state packing of porous organic cages, which are also typically crystalline.^{[ 17 ]} This analogy highlights the potential of deliberately exploiting face‐to‐face aromatic interactions as a predictive design strategy for constructing robust, highly‐porous, crystalline cage assemblies (Figure 2d).

Another approach to strengthen inter‐cage interactions involves the promotion of strong and directional supramolecular interactions; particularly, metal coordination and hydrogen bonding. In 2012, Hong‐Cai Zhou and colleagues demonstrated that Cu(II) paddlewheel units, which are ubiquitous building blocks in many cages, can engage in mutual coordination via Cu–O bridges formed by carboxylate groups in a tridentate‐bridging mode.^{[ 18 ]} Using a cuboctahedral Cu‐cage decorated with 24 hydrophobic triisopropylsilyl groups, the authors showed that in polar solvents such as dimethylformamide (DMF), cages packed into one‐dimensional chains were stabilized by inter‐cage metal‐ligand coordination. The resulting assembly was resistant to desolvation, and its crystalline nature translated into significantly enhanced porosity: gas sorption revealed an S_BET of 739 m²·g⁻¹, in stark contrast to the negligible surface area of the amorphous analogue.

Coordination between neighboring cages can also be promoted by decorating organic ligands with surface‐accessible coordinating groups. This concept was first demonstrated by Zaworotko et al. in 2004, who showed that Cu‐cages functionalized with weakly coordinating groups (e.g., methoxy and sulphate) could assemble into extended networks via axial coordination to Cu(II) centers.^{[ 19 ]} Although the authors did not report on the stability or gas‐sorption of these assemblies, their work did lay the foundation for coordination‐driven cage networks. Extending this idea, Niu et al. used hydroxyl‐functionalized ligands and combined hydrogen bonding with coordination to form two‐dimensional arrays that, upon thermal treatment, transformed into three‐dimensional networks. Both assemblies remained crystalline upon desolvation, with progressively greater gas‐uptake: thus, the S_BET values (from CO₂ adsorption at 195 K) increased from 88 m²·g⁻¹ (amorphous) to 244 m²·g⁻¹ (2D) to 471 m²·g⁻¹ (3D).^{[ 20 ]} Further refinement of this strategy came from Rh‐based cages bearing sulphonic acid groups, which underwent pH‐responsive self‐polymerization. Upon deprotonation, the sulphonate moieties coordinated to Rh(II) axial sites, producing desolvation‐resistant crystalline networks. The self‐polymerized material exhibited enhanced porosity, with an S_BET of 500 m²·g⁻¹ compared to 313 m²·g⁻¹ for its amorphous counterpart.^{[ 21 ]}

Hydrogen‐bonding interactions have also been explored to stabilize crystalline cage assemblies against desolvation. Most reported examples involve cationic Zr(IV)‐based cages incorporating trinuclear zirconocene nodes decorated with three outward‐facing hydroxyl groups. These hydroxyl groups can form hydrogen bonds with counterions (typically, chloride anions), which act as supramolecular linkers. However, in most cases, these interactions are not sufficiently strong to preserve crystallinity upon solvent removal.^{[ 22 ]} This limitation stems from the competitive hydrogen bonding between the hydroxyl groups and polar solvents, which weakens the overall network connectivity. Yuan and co‐workers addressed this challenge by replacing DMF with a less‐polar solvent, dimethylacetamide (DMA), thereby strengthening the hydrogen bonding within the tetrahedral Zr(IV)‐cages. The resulting frameworks retained crystallinity after desolvation, with a dramatic improvement in gas sorption performance: S_BET increased from 527 to 1107 m²·g⁻¹.^{[ 23 ]} In a related study, the same group assembled hydrogen‐bonded frameworks from cigar‐shaped Zr(IV)‐cages in dioxane, producing a network stabilized by a 12‐membered chlorine cluster surrounded by cages. This resulting assembly combined high crystallinity, high porosity (S_BET: 879 m²·g⁻¹) and remarkable hydrolytic stability, enabling practical applications in water‐adsorption (250 cm³·g⁻¹ at 90% RH) and proton‐conduction (6.80 × 10⁻³ S·cm⁻¹).^{[ 24 ]}

Hydrogen‐bonding interactions can also be imparted to cages through the introduction of functional groups in their organic ligands, although this approach remains underexplored due to synthetic challenges. Strongly coordinating groups such as carboxylic or phosphonic acids tend to bind metal centers during cage assembly, limiting their availability for intermolecular hydrogen bonding. Gong and colleagues have overcome this challenge by employing calixarene‐capped M₄ clusters (M = Ni(II) or Co(II)) based on p‐tert‐butylsulphonylcalix[4]arene, combined with 4′,6,6′‐tetrakis(4‐benzoic acid)‐1,1′‐spirophosphonate linkers. The resulting octahedral cages incorporated the linker in a 3‐c coordination mode, leaving pairs of carboxylic and phosphonic acid groups exposed at each window. These pendant groups established directional hydrogen bonds between neighboring cages, yielding a robust crystalline framework with permanent porosity (S_BET: 1239 m²·g⁻¹). Beyond confirming the structural stability of this framework, the authors demonstrated its functional utility in heterogeneous Brønsted‐acid catalysis, including in asymmetric [3 + 2] cycloadditions of indoles with quinone monoimine and in Friedel–Crafts alkylations of indoles with aryl aldimines.^{[ 25 ]}

3. Amorphous States in Polymer‐Grafted Cages: From Rubbery to Glassy Materials

Although phase transitions such as melting are common in small (sub‐nanoscale) molecules, they are rare in large (nanoscale) molecules. This is because the number of potential interactions with neighboring molecules increases with molecular size, leading to stronger overall attractive forces and a deeper potential‐energy well. This in turn results in large energy penalties for non‐equilibrium intermolecular distances, hindering transitions to less‐ordered phases such as liquids or glasses.

In the case of cages, which are typically stabilized by short‐range intermolecular forces (as we described above), these local interactions create assemblies that cannot efficiently dissipate thermal energy through molecular motion. Consequently, the input of thermal energy often leads not to melting or vitrification, but rather to bond‐cleavage and subsequent decomposition (Figure 3a). However, in seminal theoretical work, Doye et al. proposed that the high stabilization energy of nanoscale molecules and particles could be reduced by extending the range of attractive intermolecular interactions to dimensions comparable to the size of the molecular or particle core.^{[ 26 ]} This concept was experimentally validated by Mann and co‐workers, who achieved low‐temperature melting of globular proteins (diameter: 3 to 13 nm) by functionalizing their reactive surfaces with polymeric surfactant coronas. These coronas served to spatially separate the protein molecules, thereby reducing and broadening the range of intermolecular attractions. The result was an increase in molecular mobility, which lowered the energy barrier for melting and facilitated the phase‐transition.^{[ 27 ]}

Figure 3.

a) Morse potential‐energy curves for small molecules (orange), cages as representative nanometric entities (turquoise), and polymer‐grafted cages (green). Key: d: distance between the entities. d₀: size of each entity. b) AFM image showing the polymeric corona surrounding the central cage. Adapted from Refs. [26] (a) and [29] (b).

A similar approach to the protein‐melting one above could be envisioned for cages, many of which are similar in size to globular proteins (2 to 5 nm) and exhibit a quasi‐spherical architecture. Tailoring of their peripheral functionality (e.g., through flexible chains, bulky substituents, or dynamic motifs), can weaken and delocalize inter‐cage interactions, potentially enabling thermally‐induced phase‐transitions such as melting or vitrification without any structural collapse. Although polymer‐grafted cages have been synthesized both directly and via post‐synthetic methods since 2011,^{[ 28 ]} the first systematic study on the influence of polymer chain‐length on the physical state of cages was only published in 2016. In this pioneering work, Hosono, Kitagawa et al. functionalized cuboctahedral Cu‐based cages with butyl acrylate or tert‐butyl acrylate polymers using two complementary strategies: a divergent one, grafting polymers onto pre‐formed cages; and a convergent one, introducing polymeric ligands prior to cage self‐assembly. These methods provided access to a broad range of polymer molecular weights, thus furnishing a platform with which to explore the relationship between chain‐length and cage properties (assembly, dynamics, and bulk).^{[ 29 ]} Atomic force microscopy (AFM) confirmed that each ∼2.5 nm cage‐core was enveloped by a soft polymer shell (Figure 3b). Small‐angle X‐ray scattering (SAXS) analysis, performed in the bulk (solvent‐free) state, revealed that inter‐cage spacing increased proportionally with polymer chain‐length, regardless of polymer composition. For example, cages functionalized with butyl acrylate or tert‐butyl acrylate having lengths of 2.0 or 60 kDa showed inter‐core distances of 2 or 6 nm, respectively.

Moreover, the chemical nature of the polymer shell played a decisive role in dictating the macroscopic behavior of the cage assemblies. Cages grafted with tert‐butyl acrylate formed glassy solids at 25 °C, with a glass‐transition temperature (T _g) of 41.8 °C, whereas those bearing butyl acrylate behaved as rubbery materials (T _g: ‐43.7 °C). Since both systems had similar molecular weights, the authors attributed this large difference to steric effects: the bulkier tert‐butyl side chains restrict molecular motion more effectively, increasing the energy barrier for the glass‐to‐rubber transition.

In a subsequent study, Horike and co‐workers examined Cu‐cages functionalized with polyethylene glycol (PEG) chains of varying lengths (0.5 to 0.9 kDa).^{[ 30 ]} Differential scanning calorimetry (DSC) revealed that all PEG‐cage materials formed glasses which, below room temperature, transitioned into rubbery states. Interestingly, T _g increased as PEG chain length decreased: cages with PEG chains having a length of 0.5 or 0.9 kDa showed T _g values of –20.9 or –59.3 °C, respectively. These authors explained this trend according to polymer‐confinement at the cage surface: shorter chains remain close to the core, where steric crowding restricts mobility, whereas longer chains extend outward and behave more like free polymers. This confinement effect also appeared in mechanical properties, as the storage modulus (G’) decreased with increasing chain length. All PEG‐cage materials displayed thermoplastic behavior, undergoing solid‐to‐fluid transitions at elevated temperatures, which enabled their processing into grain‐boundary‐free membranes.

The above findings established a structural model in which polymer‐grafted cages behave as core‐shell assemblies, with a dense inner shell and a more mobile outer corona (Figure 4a). This model was reinforced by Yin and colleagues, who grafted cuboctahedral Cu‐cages with polystyrene chains ranging in length from 2.2 to 12.0 kDa.^{[ 31 ]} SAXS and small‐angle neutron scattering (SANS) analyses confirmed that tethered chains were more stretched than free polymers, consistent with confinement‐driven extension. SANS also confirmed that the cage interior remains inaccessible to solvent, consistent with a densely packed polymer shell surrounding the cage. However, beyond this confined region, the polymer chains are less crowded and behave more similarly to free polymers. Thus, the thermal properties followed a size‐dependent trend: polystyrene‐grafted cages with a chain‐length of 2.2 kDa exhibited a T _g of 90.2 °C, which is ∼20 °C higher than that of the free polymer. As the chain‐length increased to 6.0 kDa, the T _g decreased to 87.4 °C, approaching that of bulk polystyrene (T _g: 83.6 °C), indicating partial escape from the confined region. Beyond this point, T _g began to rise again with chain length, mirroring the behavior of free polymers. A similar pattern was observed in mechanical properties: G’ initially decreased from 2.2 to 6.0 kDa, as chains gained mobility without sufficient length for entanglement. In contrast, G’ increased monotonically for cages functionalized with polystyrene chains longer than 6.0 kDa, in which the entangled segments between cages contributed to mechanical stiffness (Figure 4b).

Figure 4.

a) Representation of two distinct areas of the polymeric corona surrounding the cage: the confined area and the entanglement area. b) The confined area hinders diffusion of molecules, whereas the entanglement area dictates thermal and crystalline features.

As researchers have gained understanding of confinement effects, they have devised new strategies for tuning the physical state and the macroscopic properties of polymer‐grafted cages. For example, Yin and co‐workers synthesized cages functionalized with flexible aliphatic polymeric chains capped with bulky isooctyl‐functionalized polyhedral oligomeric silsesquioxane (OPOSS) units.^{[ 32 ]} This architecture combined highly mobile aliphatic segments with sterically restricted end‐groups, yielding a glassy solid with a relatively high T _g of ‐10 °C, which the authors attributed to the interplay between the flexibility of the aliphatic chains and the restricted mobility imparted by the bulky isooctyl‐POSS moieties. The steric bulk of POSS prevented close packing and interpenetration of neighboring cages, as reflected by a low surface to space‐occupying ratio of 0.39. However, by applying mechanical pressure, the authors were able to induce entanglement between the cage units, thereby producing homogeneous transparent films with remarkable elasticity and a Young's modulus of ca. 17.42 MPa. This rare combination of stiffness and flexibility emerged from the dual role of interpenetrated POSS units as physical crosslinks and flexible aliphatic chains as elastic components.

The confined‐area structural model proposed for polymer‐grafted cages also predicts that the internal cavities remain free of polymer chains, a hypothesis supported by SAXS, which consistently shows inter‐cage distances exceeding the stretched length of the grafted polymer chains. Direct confirmation came in 2023, when Yin et al. reported gas uptake in polymer‐grafted cages.^{[ 33 ]} Cuboctahedral Cu‐cages functionalized with poly(propyleneglycol) chains of 0.3–12.8 kDa displayed a range of physical states: brittle glasses (0.31 kDa) to viscoelastic solids (0.6 and 12.8 kDa) to liquids (1.5 to 2.5 kDa). This non‐monotonic progression aligned with the confined‐area model: as chain length increased, cages transitioned from glassy to liquid‐like behavior, before entanglement restored viscoelasticity. Nitrogen adsorption at 77 K revealed time‐dependent uptake, which the authors attributed to slow diffusion through the polymer shell, a phenomenon reminiscent of solution‐diffusion transport in mixed‐matrix membranes. The role of polymer dynamics was further highlighted by a partially‐grafted cage bearing chains on only 12 of the 24 surface sites, which exhibited higher gas uptake than did the fully‐grafted cage, due to reduced confinement and faster chain dynamics.

4. Meltable and Porous Cages: Toward Porous Liquids

The examples that we highlighted in the previous sections demonstrate how the polymer surface layer influences the ability of cages to gain molecular mobility and transition between different solid states. However, as shown by Mann and co‐workers for globular proteins, the polymer layer can also be used to access a cage's liquid state. A pioneering example demonstrating the feasibility of synthesizing neat, cage‐based liquids was reported by Nitschke and colleagues in 2020.^{[ 34 ]} They functionalized the surface of a tetrahedral Zn(II)‐based coordination cage with imidazolium‐capped PEG (0.8 kDa) chains (Figure 5a). The long PEG chains conferred molecular mobility to the cage, and the terminal imidazolium groups prevented intrusion of the polymer into the cage core, thus preserving the internal cavity. Although no clear melting point was detected by thermal analysis, the Zn(II)‐based cage exhibited liquid‐like behavior, as confirmed by rheological and viscosity measurements. Furthermore, the neat liquid retained host–guest chemistry, taking up small alcohols and chlorofluorocarbons (Figure 5b).

Figure 5.

a) Assembly of a tetrahedral Zn(II) cage into a permanent porous liquid and b) its selective encapsulation of t‐butanol versus n‐butanol as a neat liquid. c) Formation of the Zn(II) barrel‐type cage and d) its CO₂‐adsorption performance up to 10 bar. Adapted from Refs. [34] (a, b) and [35] (c,d).

The first meltable cage reported to adsorb CO₂ was described by Huang and co‐workers in 2023.^{[ 35 ]} They employed a barrel‐shaped Zn(II)‐based cage assembled from p‐tert‐butyl‐sulphonylcalix[4]arene and a PEG‐imidazolium‐functionalized isophthalic derivative. In this case, the imidazolium derivative used to impart electrostatic repulsion between cages was directly grafted to the cage, whereas PEG chains decorated its surface (Figure 5c). The resulting cage, named Im‐PL‐Cage, exhibited a clear melting point (T _m) at 58 °C and liquid‐like behavior, confirmed by DSC and rheological measurements, respectively. CO₂‐sorption experiments on the solid state at 25 °C showed a maximum uptake of 1.78 mmol·g⁻¹ at 10 bar (Figure 5d).

In 2024, the Furukawa's group and our group each demonstrated, in two independent studies, the potential of melting for shaping porous cages into macroscopic films and membranes. Both studies employed Rh(II)‐based cuboctahedral cages, chosen for their high chemical and structural stability and versatile surface chemistry.^{[ 36} , ^{37 ]} Furukawa et al. exploited the 12 exohedral coordination sites on Rh(II) to attach imidazole‐terminated PEG chains (1.9 kDa) bearing bulky tert‐octylphenyl groups at the opposite end.^{[ 38 ]} These bulky groups acted as steric barriers, preventing polymer chain intrusion into the cage cavity. The resulting polymer‐grafted cage, named MOP‐1A, exhibited phase‐change behavior, transitioning between a liquid and a semicrystalline material, with a crystallization temperature (T _c) of ‐24 °C and a T _m of 25 °C. Melt‐quenching led to the formation of a glass with a T _g of ‐56 °C. This was the first example of a cage exhibiting three distinct physical states: semicrystalline, liquid, and glassy. Importantly, the phase‐transformable nature of MOP‐1A enabled investigation of gas‐sorption behavior across different states (Figure 6). CO₂‐sorption measurements revealed enhanced uptake in the liquid state compared to the crystalline form at the same temperature. For example, at 0 °C the supercooled liquid state exhibited a CO₂ uptake of 0.014 mmol·g⁻¹, whereas the crystalline state exhibited a value of only 0.005 mmol·g⁻¹. Even more strikingly, CO₂ uptake increased with temperature in the liquid phase, reaching 0.044 mmol·g⁻¹ at 30 °C, an unusual trend for porous materials. These findings reinforce the hypothesis that polymer‐chain mobility governs gas uptake in polymer‐grafted cages, as amorphous states and elevated temperatures enhance polymer flexibility, thereby facilitating gas diffusion.

Figure 6.

Schematic of the functionalization of cage surfaces with polymer chains and the resultant materials, via the coordination route (left) or the covalent route (right). Adapted from Refs. [38] (left) and [39] (right).

In our approach, we leveraged the covalent reactivity of Rh‐cages rather than their coordination chemistry to graft 24 PEG chains.^{[ 39 ]} Specifically, we coupled a carboxylic acid‐functionalized Rh‐cage with amine‐terminated PEG (2.0 kDa) via amide formation, yielding a polymer‐grafted cage that we named BCN‐93. Unlike MOP‐1A, the PEG chains in BCN‐93 lacked bulky end‐groups as we hypothesized that the surface confinement effect alone would be sufficient to prevent polymer intrusion. This strategy produced a meltable cage with T _m and T _c of 47 °C and 1 °C, respectively. BCN‐93 exhibited a CO₂ uptake of approximately 0.2 mmol·g⁻¹ at 25 °C and 1 bar in both the supercooled liquid state and the semicrystalline state. This performance exceeds that of most state‐of‐the‐art meltable materials. We attributed it to the presence of open metal sites as well as to the absence of bulky substituents, which can hinder polymer dynamics and gas diffusion. More importantly, we envisioned a novel application for the liquid state of cages as solvents for “porous solutions.” Thus, we dissolved or dispersed additional cages and MOFs within the liquid phase of BCN‐93. After mixing, the material was shaped and then cooled to access the supercooled state, yielding a mixed‐matrix membrane in which both the continuous phase (BCN‐93) and the dispersed filler (cage or MOF) retained their porosity (Figure 6). Gas‐sorption measurements confirmed that the intrinsic porosity of the fillers was preserved, demonstrating that liquid BCN‐93 acts as a sterically hindered solvent that does not block the porosity of the incorporated cages and MOFs.

Therefore, the reported studies reveal two main strategies that have been developed to impart permanent porosity to liquids: i) the creation of confined spaces through dense surface grafting of the cages, which sterically hinders the flexible chain ends from penetrating the cage cavity; and ii) the incorporation of bulky or charged terminal groups that prevent polymer intrusion through steric or electrostatic repulsion. Both approaches, whether applied independently or in combination, have proven effective in preventing self‐intrusion of the polymeric arms.

5. Perspectives

Traditionally, the challenge of controlling the long‐range structure of porous cages in their neat solid‐state has been viewed as a limitation relative to working with more‐ordered porous materials such as MOFs and COFs. This perceived limitation stems from the inherently weak and non‐directional intermolecular forces that dictate their assembly. However, as we have reviewed in this article, researchers have been developing strategies to transform this drawback into an opportunity. By carefully tuning the surface chemistry of cages, scientists can now access a broad spectrum of architectures, from robust crystalline assemblies to liquid melts, while preserving porosity.

Looking ahead, future efforts should focus on deepening our understanding of the molecular‐level organization of cages in their neat state. Such insights will be crucial to unlocking novel functionalities and to guiding the development of new classes of porous materials. For example, Furukawa and co‐workers have demonstrated that the presence of large, exposed aromatic panels in octahedral cages can direct their assembly into open‐framework structures. This finding suggests that shape complementarity between exposed surfaces could be strategically exploited to rationally design new cage‐based frameworks. Additionally, computational tools could play a major role in predicting such assemblies and guiding synthetic efforts as they have for porous organic cages.^{[ 40 ]} Indeed, Jelfs and co‐workers have recently developed computational tools to predict the assembly of octahedral cages.^{[ 41 ]} Furthermore, the rational design of complementary building blocks could enable the co‐assembly of distinct cage units into extended frameworks.^{[ 40 ]} Such architectures would combine the robustness and permanent porosity typically observed in crystalline frameworks with the processability and solubility of discrete cages. Importantly, supramolecular cage‐based frameworks might exhibit dynamic behavior such as responsiveness to external stimuli, due to the relatively weak intermolecular forces that govern their assembly. Indeed, gate‐opening effects triggered by guest adsorption have already been reported in crystalline cage systems,^{[ 12} , ^{42 ]} and such phenomena could be further enhanced by precise control over extrinsic porosity (i.e., the void between individual cages).

Significant conceptual advancements have also been made in understanding amorphous cage states (e.g., rubbers, glasses, and liquids), where the surface‐confined polymer layer model has proven effective for rationalizing their physical state, thermal properties, and mechanical behavior. This model has successfully predicted the presence of empty cavities within cages, attributed to the high surface steric hindrance caused by the densely grafted polymer chains, and highlighted the key role of polymer dynamics in the diffusion and uptake of gasses. Consequently, it provides a predictive framework for tuning the mechanical and porous characteristics of soft porous matter, thereby paving the way for establishing structure–property relationships that link mechanical and gas transport behavior. This understanding can guide the optimized processing of soft porous cages into macroscopic materials with tailored properties for targeted applications, such as flexible and selective membranes for gas separation or robust adsorbents with high uptake capacity for gas storage and conversion potentially rivalling conventional porous materials. Achieving this goal could usher in a new era in porous materials, in which high processability would no longer compromise gas‐sorption performance.^{[ 43 ]}

Furthermore, a new phase domain could become accessible for porous cages by achieving long‐range inter‐cage order while maintaining fluidity.^{[ 44} , ^{45 ]} Such porous liquid‐crystalline phases would enable the oriented arrangement of cages, allowing controlled diffusion pathways that enhance transport kinetics and selectivity, thereby offering new opportunities for gas separation applications.

Finally, the emergence of porous, phase‐changeable cages creates new opportunities for designing composite materials. These novel materials could span from liquid solutions, whereby the cage acts as the solvent for gas–liquid transformations,^{[ 35 ]} to mixed‐matrix polymer and glass composites, in which the cage forms the continuous phase.

Overall, we are confident that a deeper molecular‐level understanding of how surface chemistry impacts the organization of cages in their neat state will lead to a new generation of porous media in which pre‐designed porosity could be merged with tunable phase, mechanical, and thermal characteristics.

Author Contributions

The manuscript was written through the contributions of all authors. All authors have given approval to the final version of the manuscript.

Conflict of Interests

The authors declare no conflict of interest.

Biographies

Francisco Sánchez‐Férez earned his BSc (2016), MSc (2017), and PhD (2024) from Universitat Autònoma de Barcelona (UAB). He is currently a postdoctoral researcher in Prof. Daniel Maspoch's group under the supervision of Dr Arnau Carné at the Catalan Institute of Nanoscience and Nanotechnology (ICN2). He focuses on devising new tailoring strategies to develop interstitial materials based on metal organic polyhedra.

Arnau Carné‐Sánchez is a Ramón y Cajal (RyC) researcher at the Universitat Autònoma de Barcelona (UAB). He earned his PhD in Chemistry at the UAB in 2014 under Prof. Daniel Maspoch and Dr. Inhar Imaz. That year, he joined Prof. Susumu Kitagawa and Prof. Shuhei Furukawa at Kyoto University (iCeMS, Japan) as a JSPS fellow. In 2020, he became a LaCaixa Junior Leader at ICN2 and in 2022 returned to UAB. His research explores synthesis, reactivity, and function of molecular cages.

Daniel Maspoch is an ICREA Research Professor at the Institut Català de Nanociència i Nanotecnologia (ICN2). He received his B.S. degree at the Universitat de Girona and his PhD degree at the Universitat Autònoma de Barcelona & Institut de Ciència de Materials de Barcelona. He worked as a postdoctoral fellow at Northwestern University. His research interests include reticular materials (MOFs, COFs and cages) and delivery systems.

Sánchez‐Férez F., Carné‐Sánchez A., Maspoch D., Angew. Chem. Int. Ed. 2026, 65, e21455. 10.1002/anie.202521455

Data Availability Statement

The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.