Ion‐Permselective Porous Organic Cage Membranes

ABSTRACT

Porous organic cages (POCs) have emerged as a distinct class of molecular porous materials featuring precisely tunable pore architectures and exceptional solution processability, positioning them as promising platforms for membrane‐based ion separation. Their unique ability to synergistically combine structural confinement with tunable ion–channel interactions enables the selective separation of ions with closely similar sizes and physicochemical properties. This review provides a systematic overview of recent advances in POC‐based ion separation membranes, with an emphasis on molecular design, membrane fabrication, interfacial engineering, and the underlying separation mechanisms. Furthermore, it offers critical insights into emerging design principles and key challenges for achieving high‐performance membranes, thereby establishing a conceptual framework for the development of next‐generation ion separation systems.

This review highlights POC‐based ion separation membranes, covering molecular design, fabrication strategies, and transport mechanisms. It emphasizes how cavity–window structures enable precise ion regulation via coupled size sieving and ion–channel interactions. Representative membrane approaches and separation mechanisms are discussed, with design principles and future directions outlined for high‐performance ion separation.

1. Introduction

Efficient and selective ion separation is of great importance to a wide range of applications, including water purification, resource recovery, and emerging energy technologies [1, 2]. In particular, achieving precise discrimination between ions with similar hydrated radii and physicochemical properties, such as Li⁺/Mg^{2 +}, remains a long‐standing challenge in membrane science [3, 4]. Advanced porous materials such as covalent organic frameworks (COFs), metal–organic frameworks (MOFs), porous organic cages (POCs), and hydrogen‐bonded organic frameworks (HOFs) have been extensively explored as potential alternatives to conventional polymeric membranes due to their well‐defined and ordered channels [5, 6, 7, 8, 9]. Among these, POCs, as a class of discrete and solution‐processable molecular porous materials with intrinsically tunable cavities, have recently emerged as promising building blocks for next‐generation ion separation membranes [10].

POCs are constructed from discrete molecular cages that pack in the solid state to form intrinsic and extrinsic porosity, offering unique opportunities for precise pore engineering at the molecular level [11, 12]. Their structural diversity and tunability enable the introduction of functional groups that interact specifically with target ions through chemical interactions such as electrostatic attraction and hydrogen bonding, thereby offering additional control over ion transport behavior [13, 14]. In addition, their solution processability allows for homogeneous dispersion within polymer matrices, facilitating the fabrication of mixed matrix membranes with improved interfacial compatibility and reduced nonselective defects [15, 16]. Collectively, these features make POCs particularly attractive in membrane‐based ion separation. The combination of tunable pore sizes and functionalizable pore environments provides new opportunities for balancing ion selectivity and permeability in membrane systems through the synergistic effects of size sieving and ion‐channel interactions [12, 14].

Despite the aforementioned advantages of POCs for ion separation membranes, a comprehensive understanding of their design principles, fabrication strategies, and ion transport mechanisms remains limited. Although several excellent reviews have summarized the synthesis, structural design, and applications of POCs, systematic discussions specifically focused on POC‐based ion separation membranes remain relatively scarce. Existing reviews mainly focus on the fundamental chemistry, structural regulation, and broad applications of POCs, whereas a unified mechanistic understanding of ion transport within subnanometer confined POC channels is still lacking. Distinct from previous reviews, this review highlights recent advances in POC‐based ion separation membranes, covering the synthesis and structural characteristics of POCs, membrane fabrication strategies, and the underlying ion transport mechanisms. Particular attention is given to the interplay between size exclusion and interaction‐based selectivity in governing ion transport within confined channels. By integrating fundamental insights with representative application examples, this review aims to elucidate the structure–property‐performance relationships of POC‐based membranes and provide practical guidelines for the rational design of high‐performance ion separation systems. For ease of reference, the abbreviations frequently used throughout this review are summarized in Table 1.

TABLE 1.

List of Abbreviations.

Abbreviation	Full name
POCs	Porous organic cages
MOFs	Metal‐organic frameworks
COFs	Covalent‐organic frameworks
HOFs	Hydrogen‐bonded organic frameworks
MMMs	Mixed‐matrix membranes
TFN	Thin‐film nanocomposite
DCC	Dynamic covalent chemistry
IP	Interfacial polymerization
CC3	Covalent cage 3
ASPOCs	Amorphous scrambled porous organic cages
RCC3	Reduced covalent cage 3
RDF	Radial distribution function
DFT	Density functional theory
MD	Molecular dynamics
CHDA	Cyclohexanediamine
TFB	Triformylbenzene
EDA	Ethylenediamine
AAO	Anodic aluminum oxide

2. Synthesis and Features of Porous Organic Cages

The synthesis of POCs can be broadly categorized into two main approaches based on the nature of bond formation during the self‐assembly process: irreversible covalent chemistry and dynamic covalent chemistry (DCC) [17, 18, 19]. Irreversible covalent approaches involve the formation of robust, nonreversible bonds such as carbon–carbon bonds [20], nucleophilic substitution reactions [21], and Knoevenagel condensation [22]. These strategies usually afford highly stable cage structures. However, they generally lack error‐correction mechanisms during synthesis, which may limit structural precision and adaptability [23]. In contrast, DCC relies on reversible bond formation, allowing self‐correction and thermodynamically controlled assembly of well‐defined cage structures [24, 25, 26, 27]. Representative reactions include imine condensation [28], boronic ester formation [29], and alkene/alkyne metathesis [30] (Figure 1a). In addition, hybrid strategies combining multiple dynamic reactions, such as imine–boronic ester [31] or imine–alkyne metathesis [32] systems, have been developed to further expand structural diversity and functionality. This section provides an overview of the synthesis strategies, structural features, and functionalization approaches of POCs, with an emphasis on their relevance to ion separation.

FIGURE 1.

Synthetic strategies for POCs and selected examples prepared via Schiff‐base reactions. (a) Schematic representation of different DCC synthetic approaches for POCs [30]. (b) Building blocks for imine‐linked cage molecules [33]. (c) Structures and cage alignments for cages 1–3 [10]. (d) Assembly process of CPOC301, including its x‐ray crystal structure and the solid‐state molecular packing viewed along the [001] direction. Reproduced with permission [34]. Copyright 2021, Springer Nature. (e) Molecular structures of (left) CC3, (middle) CC3OH, and (right) CC3OHS (scrambled CC3OH) POCs and snapshots of CC3OH possible porewindow configurations with 0 % to 100 % OH substitution of the POC linker. Reproduced with permission [35]. Copyright 2025 EIsevier.

2.1. Molecular Design of POCs

POCs are constructed from organic molecular building blocks (BBs) that are covalently linked into cage‐like architectures and further assembled through weak intermolecular interactions (Figure 1b) [33]. The diversity of BBs enables precise control over pore structures, surface chemistry, and the overall physicochemical properties of POCs. By varying the geometry and functionality of the building blocks, not only can the shape and size of the cages can be tailored, and different packing modes can also be achieved in the solid state, leading to significant variations in pore connectivity and the topology of pore networks [36]. Among the various synthetic strategies, imine condensation, also known as Schiff‐base chemistry, is the most widely used approach for constructing POCs because of its synthetic simplicity and broad structural versatility [37].

In 2009, the group of Andrew I. Cooper first reported the CC series of imine‐based POCs through the condensation of 1,3,5‐triformylbenzene with diamines, affording tetrahedral cage structures (Figure 1c) [10]. A representative example is CC3, one of the most extensively studied imine‐based porous organic cages. It is constructed via a [4 + 6] Schiff‐base condensation reaction between four molecules of 1,3,5‐triformylbenzene and six molecules of 1,2‐diaminocyclohexane. CC3 has served as a model system in numerous fundamental studies of cage packing, pore engineering, and membrane fabrication. Subsequently, Yuan and coworkers synthesized a series of cup‐shaped polycyclic organic cages (CPOCs) using tetraformyl bowl‐shaped arenes and diamine linkers, affording diverse architectures ranging from [2 + 4] lantern‐like cages to [6 + 12] octahedral cages (Figure 1d) [34]. Building on these advances, Root and coworkers further expanded the scope of DCC by synthesizing a core‐scrambled POC derivative (CC3‐OH‐S) through Schiff‐base condensation between benzene tricarbaldehyde (or its hydroxylated analogue, 2‐hydroxybenzene tricarbaldehyde) and (R,R)‐cyclohexyldiamine (Figure 1e) [35]. This strategy enables the controlled incorporation of mixed hydroxyl functionalities into the cage core, thereby allowing fine‐tuning of the internal chemical environment. These studies collectively highlight the versatility of DCC in constructing structurally diverse and functionally tunable cage architectures. Moreover, the reversibility of imine bonds imparts structural adaptability and error‐correction capability to POCs, facilitating their use as modular building blocks for the construction of extended framework materials [38]. While imine‐linked POCs, such as CC3, are widely used as model systems, their dynamic imine bonds can undergo hydrolysis or structural rearrangement under humid or highly ionic conditions, which may affect long‐term membrane performance. In contrast, the synthesis of POCs via irreversible covalent chemistry remain less explored because of greater synthetic complexity and typically lower yields. Nevertheless, irreversible linkages provide higher chemical and structural stability than dynamic bonds such as imine or boronate ester linkages, making them advantageous for applications under harsh conditions.

2.2. Structural Features of POCs

The structural features of POCs are mainly defined by their cavity‐window architecture (Figure 1a), the modularity of BBs (Figure 1b), and their solid‐state packing modes (Figure 1c). Collectively, these features determine their porosity, transport pathways, and ultimately their ion separation performance. A key feature of POCs is their intrinsic permanent porosity, arising from well‐defined internal cavities interconnected by molecular windows. These sub‐nanometer apertures act as selective gateways, enabling size‐ and property‐dependent ion transport. In membrane applications, such cavity–window structures provide well‐defined transport pathways and are essential for achieving high selectivity through the synergistic effects of size sieving and specific ion‐channel interactions [39]. In addition, the modularity of building blocks allows precise control over cage geometry, cavity size, and chemical environment. Different packing arrangements of POCs further give rise to diverse pore connectivity and network topologies, which strongly influence ion transport behavior. A representative example illustrating these features is CC3, a prototypical imine cage with a tetrahedral cavity and four triangular windows. Its cyclohexyl vertices promote a window‐to‐window packing mode, leading to a three‐dimensional diamondoid porous network (Figure 1c) [10]. This packing arrangement generates dual transport pathways, including channels derived from the intrinsic cavities (∼6 Å) of individual POC molecules and extrinsic interstitial voids (∼3.5 Å) formed between adjacent cages, thereby creating a hierarchical transport system [9, 40]. The cage windows function as molecular sieving gates that regulate ion transport between these domains, and ultimately governing permeability and selectivity. The intrinsic cavities provide well‐defined selective pathways for ion permeation, while the extrinsic interstitial voids originate from packing‐induced intermolecular free volume and may serve as auxiliary transport domains under certain conditions. Notably, these confined intermolecular voids represent intrinsic structural features arising from the ordered assembly of POC molecules rather than nonselective void defects. Together, the intrinsic cavity channels and extrinsic interstitial voids cooperatively govern ion transport behavior in POC‐based membranes.

POCs exhibit a range of unique and adaptive structural properties arising from their discrete molecular nature and flexible packing behavior. One notable feature is the ability to undergo reversible polymorphic transformations. For example, Marc A. Little and coworkers showed that two thermodynamically favored “α” crystal packing modes can be interconverted by using specific cosolvents to direct crystal packing, enabling CC3‐R to adopt the packing mode of CC4‐R, and vice versa (Figure 2a) [41]. In addition, solvent‐directed assembly provides an effective route to reconstruct pore structures. Thomas Hasell and coworkers reported that the inclusion of 1,4‐dioxane as an organic directing solvent can shift cage packing away from the lowest‐energy polymorphs and generates isostructural three‐dimensional diamondoid pore networks (Figure 2b) [42]. Under these conditions, different cages, including CC1, CC2, and CC13 can all be directed into the same diamondoid window‐to‐window packing arrangement as CC3, highlighting the important role of external stimuli in controlling pore connectivity and topology. In addition, the dynamic nature of covalent bond formation makes it possible to introduce mixed building blocks and thereby generating cages with greater structural diversity [43, 44, 45]. For instance, amorphous scrambled POCs (ASPOCs) can be prepared by combining different diamine linkers (Figure 2c), such as ethylenediamine (EDA) and cyclohexanediamine (CHDA), in varying ratios, yielding materials such as CC1⁴CC3² and CC1¹CC3⁵ (commonly denoted as 1⁴3² and 1¹3⁵) [44].

FIGURE 2.

Structural adaptability of POCs arising from flexible packing and dynamic covalent chemistry. (a) Guest‐controlled crystal packing enables reversible interconversion between distinct polymorphs (e.g., CC3 and CC4), demonstrating the responsiveness of cage packing to external stimuli. Reproduced with permission [41]. Copyright 2014, Royal Society of Chemistry. (b) Solvent‐directed assembly modulates cage packing behavior, generating alternative three‐dimensional pore networks with distinct connectivity and topology. Reproduced with permission [42]. Copyright 2014, American Chemical Society. (c) Dynamic covalent scrambling through mixed diamine building blocks (e.g., EDA and CHDA) generates compositionally diverse cages, yielding ASPOCs with tunable structural disorder. Reproduced with permission [44]. Copyright 2011, Springer Nature.

Beyond their structural tunability, POCs also exhibit good solution processability, which is particularly advantageous for membrane fabrication. Unlike extended frameworks materials (e.g., MOFs and COFs), the discrete nature of POCs enables their dissolution in common organic solvents, allowing homogeneous dispersion within polymer matrices and facilitating the preparation of mixed matrix membranes with reduced interfacial defects. As a result, POC‐based membranes can be fabricated using conventional techniques commonly employed for polymer membranes, such as spin‐coating, dip‐coating, spray‐coating [46, 47], and polymer blending. In this respect, POCs combine the processability of molecular materials with the structural features of porous solids, although their discrete nature still requires careful control of fabrication conditions to obtain uniform and defect‐free membranes. For example, Song et al. first used the spin‐coating method to deposit a POC solution onto a porous anodic aluminum oxide (AAO) substrate for thin film fabrication, demonstrating the practical advantages of POCs in solution‐based processing [48]. However, the fabrication process is highly sensitive to operating conditions, and not only solution concentration and spin speed, but also environmental factors such as temperature and humidity, can strongly affect membrane quality. In addition, by adjusting solution conditions such as pH, some POCs can be dissolved in aqueous media, which further enables their use as molecular building blocks in interfacial polymerization [49, 50]. This behavior originates from the dynamic covalent nature of POCs, where reversible bond exchange allows ligand redistribution in solution, and may lead to the formation of new species or dynamically reorganized structures.

2.3. Functionalization of POCs

Functionalization of POCs is an important strategy for tuning their physicochemical properties and improving ion separation performance. Such regulation can be achieved either by modifying the building blocks or by functionalizing the internal cavities, thereby enabling precise control over the pore microenvironment [13, 51]. Through the incorporation of specific functional groups into the cage framework, the pore channel chemistry can be precisely tailored.

Common strategies include the introduction of coordinating groups, heteroatoms, and rigid or planar building blocks to enhance cage stability and functionality. For instance, porphyrin‐based building blocks have been employed to construct shape‐persistent cages with inherent coordination sites, which can be metallated to create Zn‐, Fe‐, or Co‐containing POCs with tailored ion affinity [52, 53, 54]. Similarly, salphen‐based cages provide rigid scaffolds for selective metal ion binding through well‐defined coordination environments [55]. Host–guest chemistry provides another important pathway for functional modulation. The intrinsic cavities of POCs can selectively encapsulate guest species, allowing ion recognition via noncovalent interactions such as hydrogen bonding or electrostatic interactions. Fluorinated cages, for instance, exhibit preferential adsorption of perfluorocarbons due to F–F interactions, demonstrating the role of cavity microenvironment in selective capture [56]. Wen and coworkers designed a helical porous organic salt cage (CPOS) that mimics the biological potassium channel KcsA, achieving high K⁺ selectivity over Li⁺ and Na⁺ through size complementarity and specific interaction sites [57]. Functionalization can also improve structural robustness. For example, introducing intramolecular hydrogen‐bonding motifs or creating additional mesopores via coassembly with ionic surfactants enhances cage–cage interactions, improves crystallinity, and imparts electrostatic properties to the pores. Song et al. demonstrated that introduction of hydrogen‐bonding enables the fabrication of self‐supporting POC membranes with high crystallinity, leading to significantly improved mechanical strength and long‐term operational stability [58]. In addition, postsynthetic modification provides a flexible route to regulate pore accessibility and separation performance by enabling conversion between porous and nonporous states [59]. Representative functionalization strategies of POCs, including both preassembly modification of building blocks and postassembly cavity functionalization, are summarized in Table 2.

TABLE 2.

List of representative functionalization strategies of POCs reported in the literatures.

Publication year	POC materials	Method of functional modification	Effect of functional modification	Ref.
2012	[2 + 3] externally functionalized cages	Peripheral functionalization at cage vertices	Tunable cavity size enabling size‐ and shape‐selective molecular recognition	[46]
2013	CC3	Chemical reduction (imine to amine bonds)	Modulated porosity and rigidity; CC3 enhances while redCC3 suppresses gas permeability in MMMs	[60]
2014	RCC3, AT‐RCC3, FT‐RCC3	Postsynthetic tethering modification to fix molecular conformation	Restored shape persistence and intrinsic porosity; enhanced chemical stability	[59]
2016	CC3, RCC3, FT‐RCC3	Chemical reduction of CC3 to RCC3, followed by crosslinking (FT‐RCC3)	Enhanced chemical stability (hydrolysis resistance) and altered intrinsic porosity	[61]
2017	Noria	Boc protection and esterification (Noria‐Boc, Noria‐COtBu)	Tuned pore environment; enhanced permeability (COtBu) or selectivity (parent Noria)	[62]
2018	Zn‐PB	Zinc coordination of porphyrin box and alkyl chain substitution	Constructed stable secondary building units; enabled coordination‐driven assembly with pyridine ligands	[52]
2018	Fe‐PB	Incorporation of Fe‐porphyrin units as cage building blocks followed by metallation	Enhanced active site exposure and mass transport; enabled efficient CO2‐to‐CO conversion	[53]
2019	ASPOCs	Cage scrambling via mixed amine linkers	Improved solubility and compatibility; simultaneous enhancement of permeability and selectivity	[63]
2020	Co‐PB‐1(6), Co‐rPB‐1(6)	Incorporation of Co‐porphyrin units followed by imine reduction	Enabled site isolation of active centers; enhanced selectivity toward H2O2 formation	[54]
2021	POC‐DICP	Incorporation of triphenylphosphine (PPh3) during cage construction	Enabled bifunctionality (catalysis and separation) with enhanced catalytic activity	[64]
2022	CC3	Reduction followed by quaternization/protonation	Introduced high charge density; enabled efficient anion capture via electrostatic interactions	[65]
2022	CPOS	Assembly in conical nanochannels to form double‐helical electrostatic columns	Constructed ordered electrostatic channels; enabled highly selective K+ transport via synergistic cation–π and electrostatic interactions	[57]
2022	Cage 7‐Ni and 7‐Pt	Synthesize and functionalize metal salphen units, and then assemble them as rigid vertices into cages	Constructed a three‐dimensional cage structure with ultra large metal spacing, achieving precise control of cavity size	[55]
2022	H‐cage, HF‐cage, F‐cage	Introduction of alkyl or fluorinated side chains onto cage skeleton	Tuned pore surface chemistry; enabled selective adsorption via F–F interactions (F‐cage) and preferential uptake of non‐fluorinated alkanes (H‐cage)	[56]
2024	CC2	Chemical reduction of CC2 (C═N to C─N bonds) using NaBH4	Induced structural collapse; enabled angstrom‐scale channel formation via interfacial polymerization	[66]
2024	NKPOC‐DT‐(I‐) Me	Grafting of quaternary ammonium groups via postsynthetic modification	Strengthened adsorption driving force and accelerated ion transport kinetics	[67]
2024	2,5‐DHA‐TAEA/2,3‐DHA‐TAEA	Aldehyde isomer engineering with flexible triamines	Enhanced intermolecular interactions, enabling robust membrane formation	[58]
2025	CC3	Amine grafting (PEI@CC3, DETA@CC3)	Improved CO2 selectivity and interfacial compatibility in MMMs	[68]
2025	CC3	Multipath functionalization (reduction, charge introduction, Ag nanocluster confinement)	Enabled multifunctionality including ion sieving and transport regulation in energy systems	[69]
2025	CC3	Core scrambling with hydroxyl‐functionalized cages (CC3‐OH‐S)	Enhanced gas uptake and optimized packing behavior	[35]
2025	CC3	Imine reduction to RCC3	Generated well‐defined ion transport channels for selective sieving	[70]
2025	(CC3)(CoCl2)6	Sequential amine functionalization and Co2 + coordination	Balanced reactivity and stability; enhanced CO2 affinity	[71]
2025	Cage‐NO2	Nitro‐to‐amine reduction (SnCl2/HCl)	Increased adsorption capacity and introduced strong interaction sites	[72]
2025	PTC‐2H	Metallation via Zn2 + coordination in porphyrin cavity	Modified adsorption behavior; reduced iodine uptake	[73]
2025	TC1/TC2/TC3	iEDDA‐based post‐synthetic modification (tetrazine chemistry)	Precisely tuned pore architecture and adsorption properties	[74]

2.4. Advantages of POCs for Ion Separation

Based on the structural and physicochemical characteristics discussed above, POCs offer several unique advantages for ion separation (Figure 3). First, their precisely tunable pore apertures can be matched with the hydration diameters of target ions, providing a structural basis for size‐selective transport [15]. Second, their chemically tunable pore environment allows the incorporation of specific interactions, such as electrostatic interactions, hydrogen bonding, ion–π interactions, and coordination effects, thereby enabling enhanced ion recognition beyond simple size exclusion [14]. Third, the excellent solution processability of POCs facilitates their integration into a wide range of membrane architectures, including mixed matrix membranes, thin‐film composite membranes, and self‐standing membranes, thereby supporting scalable fabrication [16, 75]. Finally, the hierarchical transport pathways arising from the assembly of discrete cage units can reduce transport resistance and promote efficient ion migration [9]. Collectively, these features position POCs as a versatile platform for overcoming the intrinsic limitations of conventional membrane materials and advancing next‐generation ion separation technologies that are described as below.

FIGURE 3.

Advantages of POCs for ion separation.

3. Fabrication Strategies for POC‐Based Membranes

Translating POCs from discrete molecular entities into continuous membrane architectures is a critical step toward their practical implementation in ion separation [76]. Benefiting from their well‐defined pore structures and exceptional solution processability, POCs can be processed into membranes via a variety of fabrication strategies, offering a unique level of structural and functional control [77, 78]. Unlike conventional porous materials, the discrete and soluble nature of POCs enables their integration into membranes through both top‐down (e.g., blending and coating) and bottom‐up approaches (e.g., interfacial assembly and in situ growth). These fabrication routes not only determine the spatial distribution and packing of POCs, but also strongly influence interfacial compatibility, transport pathways, and ultimately separation performance.

Importantly, the selection of membrane fabrication strategies is closely associated with the intrinsic molecular characteristics of POCs. For instance, highly soluble POCs are more suitable for solution‐based processing methods such as blending and spin‐coating, whereas rigid and highly ordered cage structures are advantageous for membrane fabrication approaches involving in situ growth or interfacial polymerization. In addition, pore window size and surface charge play critical roles in regulating ion transport behavior through size sieving and Donnan exclusion effects, respectively. Therefore, establishing a rational structure–property‐fabrication relationship is essential for the design of high‐performance POC‐based ion separation membranes.

At present, POC membrane fabrication can be broadly classified into three main categories: (i) blending method for the preparation of mixed‐matrix membranes (MMMs), in which POCs serve as fillers dispersed in a polymer matrix, (ii) direct fabrication approaches, including in situ growth and spin‐coating, and (iii) interfacial assembly for fabricating crystalline POC membranes on porous support or thin‐film composite membranes. Each strategy has its own advantages and limitations in terms of structural control, scalability, and separation performance. A summary of representative POC membranes prepared via these approaches is provided in Table 3. To provide a clearer comparative perspective on these fabrication strategies specifically for ion separation, their strengths, limitations, and most suitable application scenarios are summarized in Table 4. This comparison highlights that solution‐based methods such as blending are more suitable for scalable membrane production, whereas interfacial polymerization and in situ growth are particularly promising for constructing ultrathin selective layers or highly ordered channels with enhanced ion selectivity. Contra‐diffusion methods, on the other hand, are more appropriate for high‐selectivity systems, while direct growth approaches serve as ideal platforms for fundamental investigations of structure–property–transport relationships.

TABLE 3.

Landscape of POC‐based membranes across gas, liquid, and ion separations, correlating cage chemistry, separation targets, and fabrication strategies.

Application	Type of POC	Separation system	Preparation method	Ref.
Gas molecule separation	C1, C2	CO2/N2	Blending	[79]
	CC3	CO2/N2		[60]
	CC3	C3H6/C3H8		[80]
	ASPOC	CO2/N2		[63]
	CC3	CO2/N2	Cosolvent casting	[81]
	Noria	CO2/CH4		[62]
	CC3	CO2/N2	Spin coating	[82]
	CC3, CC13, ASPOCs, FT‐RCC3	CO2/N2		[61]
	CC3	(He, CH4, CO2, Kr)/Xe	Secondary seeded growth	[83]
	PIL@CC3	I2/air	Evaporation and electrostatic crosslinking	[84]
	CC2	CO2/N2	Interfacial polymerization (IP)	[66]
	RCC3	CO2/N2		[85]
Liquid molecule separation	CC3	(R)‐ and (S)‐2‐amino‐1‐butanol	Blending	[86]
	ASPOCs	PhMe/TIPB		[87]
	Imine cage	benzene, toluene, xylenes, mesitylene	Spray coating	[46]
	CC1	CR	Interfacial polymerization (IP)	[88]
	CC3α	RB		[89]
	CC3‐R	R‐(+)‐Limonene		[90]
	Noria	BBG, RB		[91]
	Noria	BBG		[92]
	RCC3, Tren‐cage	RB, CR, AF, OGS, MO, MR		[49]
	2,5‐DHA, TAEA cage	VB12, RB, BBR, CR, MO		[58]
Ion separation	CC3	H+/MeOH	Blending	[93]
	CC3	Li+	Spin coating	[94]
	I‐Cage‐Cl	Cr(VI)	Vacuum assistance	[65]
	RCC2	NaCl/Na2SO4;LiCl/MgCl2	Interfacial polymerization (IP)	[95]
	RCC3	NaCl/Na2SO4		[96]
	RCC3	MEB		[50]
	CC3	(KCl, NaCl, LiCl)/MgCl2	Contra‐diffusion	[9]
	CC1, CC3, CC19	(KCl, NaCl, LiCl)/MgCl2		[97]

TABLE 4.

Comparison of POC‐based membrane fabrication strategies for ion separation, highlighting advantages, limitations, and recommended applications.

Method	Main strengths	Main limitations	Most promising application scenario
Blending	Simple processing, good scalability	Possible interfacial defects and non‐selective voids	Large‐scale ion separation membranes
Interfacial polymerization	Formation of ultrathin selective layers with high ion permeance	Dependence on cage solubility and interfacial stability	High‐flux ion separation
Contra‐diffusion	Construction of highly ordered transport channels	Low fabrication efficiency and limited scalability	High‐selectivity ion separation systems
Direct synthesis	Precise control over membrane structure and crystallinity	Complex preparation conditions	Model systems for ion transport studies

3.1. Mixed‐Matrix POC Membranes Via Blending

Blending represents the most straightforward and widely adopted strategy for incorporating POCs into membrane systems. In this approach, POCs serve as porous fillers dispersed within a continuous polymer matrix to form MMMs. Typically, membrane fabrication involves dissolving a polymer in a suitable solvent, followed by the addition of POCs and their dispersion via stirring and ultrasonication. The resulting casting solution is then used to prepare the membranes. The performance of MMMs strongly depends on the uniform dispersion of POCs and the interfacial compatibility between the filler and polymer matrix. Poor compatibility often leads to particle aggregation or the formation of nonselective interfacial voids, which deteriorate membrane selectivity. Owing to their discrete molecular nature and good solubility, POCs generally exhibit improved compatibility with polymer matrices compared to conventional fillers such as MOFs and COFs, thereby reducing interfacial defects. A representative example was reported by Jansen and coworkers, who prepared a POC‐based MMM by incorporating CC3 into the polymer of intrinsic microporosity PIM‐1 polymer (Figure 4a) [60]. Notably, the strategy generates the dispersed cage phase from a single homogeneous solution, leading to improved interfacial compatibility. The resulting membrane exhibited a remarkably high CO₂ permeability (37 400 Barrer), and still maintains a permeability of 13 000 Barrer and a CO₂/N₂ selectivity of 15 after 1 year, demonstrating enhanced long‐term stability. Jin et al. further investigated the blending process by incorporating CC3 into 6FDA‐DAM using CHCl₃ as the solvent [80]. Although CC3 could be fully dissolved in the solvent to form a clear homogeneous solution, subsequent solvent evaporation induced recrystallization of cage species into porous domains within the polymer matrix, thereby generating hierarchical diffusion pathways. Despite their favorable performance, interfacial gaps between CC3 crystallites and the polymer matrix were still observed by electron microscopy, indicating that interfacial integration remain incomplete.

FIGURE 4.

Evolution of blending strategies for POC‐based MMMs. (a) Incorporation of CC3 cages into a PIM‐1 polymer matrix from a homogeneous solution, resulting in improved dispersion and interfacial compatibility compared to conventional porous fillers. Reproduced with permission [60]. Copyright 2013, Wiley‐VCH. (b) Molecularly mixed MMMs enabled by ASPOCs, which suppress cage aggregation and promote uniform dispersion due to their disordered packing, resulting in enhanced permeability and selectivity. Reproduced with permission [63]. Copyright 2019, Wiley‐VCH. (c) Advanced cosolvent engineering for constructing defect‐free and scalable POC‐based MMMs. Reproduced with permission [98]. Copyright 2023, American Institute of Chemical Engineers.

To address these challenges, significant efforts have been devoted to improving interfacial compatibility at the molecular level. For instance, Ryan P. Lively and coworkers developed amorphous scrambled porous organic cages (ASPOCs) with mixed vertex functionalities (Figure 4b) [63]. Because of the presence of different linker types, these cages exhibit disordered packing with only short‐range order, which suppresses strong intercage interactions and prevents aggregation. As a result, ASPOCs can be uniformly dispersed within polymer matrices, forming molecularly mixed composite membranes with minimal interfacial defects. The resulting membranes exhibited significantly enhanced gas separation performance, including nearly fourfold increases in permeability for N₂, CO₂, and CH₄, along with improved selectivity (e.g., N₂/SF₆ selectivity up to 115), indicating the emergence of molecular sieving effects.

Building on these studies, further progress has been made toward eliminating interfacial defects and extending POC‐based membranes to ion separation. Xu and coworkers developed a cosolvent casting approach to achieve molecular‐level compatibility between POCs and polymers by codissolving both components, followed by rapid solvent evaporation (Figure 4c) [98]. Using this method, an interfacial defect‐free MMM was fabricated by incorporating CC3 into an HPABP polymer matrix. Benefiting from the intrinsic window–cavity structure of CC3, the resulting membrane exhibited efficient ion transport and separation performance. In particular, the HPABP–CC3 membrane with 15 wt% loading delivered a high chloride ion flux of 2.24 mol m⁻² h⁻¹ and a selectivity of 12.67. More importantly, by effectively resolving the interfacial compatibility issues, this strategy enabled the scalable fabrication of POC‐based membranes, as demonstrated by membranes area as large as 40 cm × 53 cm.

3.2. POC Membranes Via Direct Synthesis

Direct fabrication represents a straightforward and effective approach for constructing POC‐based membranes by leveraging the intrinsic solubility and processability of POCs. Unlike blending strategies that rely on polymer matrices, this approach enables the formation of continuous POC layers with reduced interfacial complexity and more direct control over pore structure and transport pathways.

3.2.1. Spin Coating

Spin coating is a solution‐based deposition technique for fabricating thin POC membranes. In this process, a POC solution is deposited onto a substrate and then rapidly rotated, so that centrifugal force drives uniform spreading of the solution, followed by solvent evaporation to form a continuous thin film. This method is simple to operate and allows precise control over film thickness, making it particularly suitable for constructing ultrathin separation layers. Song et al. first used spin‐coating to deposit cage molecules onto a porous anodic aluminum oxide (AAO) support to fabricate thin POC membranes (Figure 5a) [48]. By tuning the spin speed and solution concentration, the film structure and porosity could be precisely regulated, resulting in uniform and defect‐free membranes. The resulting CC3‐based membranes exhibited high gas permeability and selectivity at 295 K, showing promising separation performance for gas pairs such as H₂/N₂ (∼30), H₂/CH₄ (∼20), CO₂/N₂ (∼19), and CO₂/CH₄ (∼10).

FIGURE 5.

Spin‐coating strategies for ultrathin POC membrane fabrication. (a) Fabrication of ultrathin CC3 membranes on porous AAO supports via spin coating. Reproduced with permission [48]. Copyright 2016, The Authors. Published by WILEY‐VCH. (b) Electrostatic‐field‐induced crystal rearrangement in BMIMBF₄@CC3 membranes [82] Copyright 2022, Wiley‐VCH GmbH. (c) Spin‐coated CC3 membrane on PP separator for selective Li⁺ transport. Reproduced with permission [94]. Copyright 2022, American Chemical Society.

Building upon this method, Qu et al. further introduced an electrostatically induced crystal rearrangement strategy to enhance membrane performance (Figure 5b) [82]. By incorporating the ionic liquid BMIMBF₄ (1‐butyl‐3‐methylimidazolium tetrafluoroborate) as an auxiliary agent, crystal rearrangement was induced on the membrane surface, which effectively eliminated nonselective intercrystalline voids and generated more ordered transport channels. This structural reorganization enabled precise tuning of pore size and transport pathways, thereby facilitating preferential CO₂ transport while suppressing the permeation of N₂ and CH₄.

POC membranes prepared by this spin coating approach also exhibit strong potential for selective ion transport. For example, a POC‐based ionic sieve has been developed as an ultrathin (∼300 nm) functional layer on a commercial polypropylene (PP) separator, achieving highly selective ion sieving (Figure 5c) [94]. Benefiting from the well‐defined cavity–window structure of CC3, whose pore apertures are smaller than lithium polysulfide species, the membrane allows preferential transport of Li⁺ while effectively blocking larger polysulfide ions. Consequently, the system demonstrates excellent cycling stability, with a capacity fading rate as low as ∼0.04% per cycle over 500 cycles.

3.2.2. In Situ Growth

In situ growth is another effective strategy for constructing POC‐based membranes by directly forming POC structures on a substrate through controlled chemical reactions., Compared with ex situ assembly methods, this approach can minimize interfacial defects and improve structural continuity. For example, Wu et al. employed an in‐situ crystallization method to fabricate a composite proton exchange membrane (PEM) composed of Nafion and CC3 (Figure 6a) [93]. The incorporation of CC3 significantly improved both water retention and proton conductivity. Its three‐dimensional interconnected pore network not only increased the water content of the membrane but also facilitated efficient proton transport. The resulting NC3‐5 membrane achieved a proton conductivity of 0.27 S·cm⁻¹ at 90°C and 95% relative humidity, markedly higher than that of a recast Nafion membrane under the same conditions.

FIGURE 6.

In situ growth strategies for POC membrane fabrication. (a) In situ crystallization of CC3 within a Nafion matrix to form a composite proton exchange membrane. Reproduced with permission [93]. Copyright 2018, American Chemical Society. (b) Secondary seeded growth of continuous CC3 membranes on porous AAO supports. Reproduced with permission [83]. Copyright 2020, American Chemical Society.

Building upon this concept, secondary growth strategies have been developed to further control membrane thickness, crystallinity, and pore connectivity, thereby enabling more precise control over ion or gas transport pathways. Carreon and coworkers synthesized continuous CC3 membranes on AAO supports via a secondary seeded growth method (Figure 6b) [83]. CC3 seed crystals were first prepared with different synthesis times of 60 and 120 h, and were then used for the direct growth of a continuous CC3 layer. This approach enabled effective separation of rare gases, such as xenon, from lighter gases including helium, carbon dioxide, krypton, and methane. The thickness of the CC3 membrane could be reduced to about 2.5 µm, while maintaining high single gas permeances and selectivities.

3.3. POC Composite Membranes Via Interfacial Assembly

Interfacial assembly has emerged as a powerful bottom‐up strategy for constructing highly ordered and ultrathin POC membranes. This method exploits a liquid–liquid or gas–liquid interface to create a confined two‐dimensional space that guides the self‐organization of POC molecules. Within this confined environment, intermolecular interactions orchestrate the orderly alignment of cages, ultimately forming a continuous and seamless membrane. In the early stages, however, interfacial polymerization still faced clear limitations. Because POC molecules are primarily held together by weak van der Waals forces, making it was difficult to form achieve stable free‐standing membranes. As a result, the membranes usually required a supporting substrate to maintain structural integrity.

3.3.1. Interfacial Polymerization

Interfacial polymerization (IP) is a widely used technique to fabricate ultrathin, cross‐linked active layers at the interface of two immiscible phases. In POC‐based systems, IP not only enables the formation of polymeric matrices, but also provides a confined reaction environment for the in situ construction and assembly of cage structures. Typically, building blocks such as aldehydes and amines are distributed in two immiscible phases, where dynamic covalent reactions (e.g., imine condensation) occur at the interface, leading to the formation of POC‐based networks or continuous cage layers. This interfacial confinement effectively regulates reaction kinetics and molecular organization, promoting the formation of highly ordered and defect‐minimized membranes. A representative example was reported by Cooper et al., who prepared a continuous, nearly defect‐free, shape‐persistent imine‐based POC membrane, denoted CC3α‐PAN, through interfacial synthesis (Figure 7a) [89]. The reaction at the water–dichloromethane interface produced highly crystalline CC3α domains, with controllable thickness and continuity. Notably, solvent modulation enabled reversible switching between CC3α and CC3γ′ polymorphs, allowing tunable molecular sieving behavior. A further breakthrough was achieved by Song et al., who constructed crystalline free‐standing POC membranes by strengthening intermolecular interactions through the introduction of intramolecular hydrogen bonding motifs (Figure 7b) [58]. Specifically, terephthaldehyde building blocks with tailored substituents were reacted with flexible triamine precursors via interfacial polymerization, enabling controlled cage assembly at the interface. Single‐crystal x‐ray analysis confirmed the formation of intramolecular hydrogen bonds, while DFT calculations revealed that these interactions significantly enhance intercage cohesion. As a result, the rigidity and ordered assembly of POC molecules were markedly improved, driving the formation of continuous, self‐supported, and highly crystalline membranes rather than discrete powders. The resulting membranes exhibited fast solvent permeance and high dye rejection in both aqueous and organic solvents, demonstrating excellent molecular sieving performance.

FIGURE 7.

Interfacial strategies for constructing ultrathin, ordered POC membranes. (a) CC3α‐PAN membranes fabricated via interfacial polymerization at the water–dichloromethane interface, producing highly crystalline, defect‐minimized domains with tunable thickness. Reproduced with permission [89]. Copyright 2022, Springer Nature. (b) Free‐standing CC3 membranes formed via hydrogen‐bond–guided interfacial assembly. Reproduced with permission [58]. Copyright 2024, Wiley‐VCH. (c) Ultrathin CC3 monolayers (∼1.4 nm) assembled at the air–water interface using the Langmuir–Blodgett method. Reproduced with permission [99]. Copyright 2025, Elsevier.

Beyond interfacial polymerization, interfacial assembly strategies based on physical interactions have also been developed to construct ultrathin POC membranes. For instance, Ding et al. reported a strategy to fabricate centimeter‐scale monolayer POC membranes at the air–water interface by exploiting strong hydrophobic interactions (Figure 7c) [99]. In this method, droplets of a chloroform solution containing CC3 molecules were deposited onto the water surface, where the hydrophobic cages became confined to the interface. Using the Langmuir–Blodgett technique, the molecules were compressed into a densely packed monolayer and subsequently transferred onto a porous substrate. Molecular dynamics simulations revealed a dynamic assembly pathway, in which individual CC3 molecules initially adopted a tilted configuration (∼80°) at the interface, followed by aggregation into small clusters (∼25 molecules), and ultimately the formation of a continuous monolayer upon compression (∼100 molecules). Despite an ultrathin thickness of only ∼1.4 nm, the resulting membranes exhibited excellent gas separation performance (e.g., N₂ permeability of 153 GPU and N₂/SF₆ selectivity of 74), while maintaining good mechanical flexibility.

Thin‐film nanocomposite (TFN) membranes represent a typical architecture derived from interfacial polymerization. These membranes possess an ultrathin selective layer consisting of POCs and polyamide supported on a porous substrate, effectively decoupling separation performance from mechanical stability. In such systems, the porous support offers negligible mass transfer resistance, while the POC‐based active layer governs ion transport and selectivity. This architecture enables the full utilization of the precise sieving capability of POCs, resulting in high permeability and selectivity.

In TFN systems, POCs can be incorporated as functional fillers or molecular building blocks within the interfacial polymerization process. For example, Wang et al. developed a CC3‐R/polyamide thin‐film composite membrane for chiral separation by dispersing chiral CC3‐R crystals in the aqueous phase containing m‐phenylenediamine (MPD), followed by interfacial polymerization with trimesoyl chloride (TMC) in the organic phase [90]. During this process, the polyamide thin film (∼320 nm thick) formed at the interface simultaneously encapsulated CC3‐R, resulting in an active layer with integrated cage functionality. This incorporation of POCs into the polyamide matrix enhances structural control and provides additional selective transport pathways, thereby improving separation performance. Despite these advantages, the practical implementation of POCs in interfacial polymerization is often limited by their low solubility in the aqueous phase, particularly for cages with large molecular weights and hydrophobic functionalities. This limitation hinders their direct use as reactive monomers or building units for forming uniform and defect‐free thin films.

To overcome this challenge, several strategies have been developed to improve the aqueous processability of POCs, which can be broadly categorized into two approaches. First, pH regulation provides an effective route to enhance cage solubility. Zhao et al. improved the aqueous solubility of RCC3 through protonation under acidic conditions, enabling its use as a comonomer with piperazine (PIP) in interfacial polymerization on hydrolyzed polyacrylonitrile (HPAN) substrates [50]. The incorporation of RCC3 introduced intrinsic subnanometer pores and additional transport pathways, resulting in membranes with high permeability (52.6 L m⁻² h⁻¹ bar⁻¹), excellent dye rejection (∼98.6%), and high dye/salt selectivity (52.8), along with improved antifouling properties and stability. Li and coworkers employed water‐soluble cages, such as tren‐cages and RCC3, as aqueous‐phase monomers to fabricate highly crosslinked polycage membranes via interfacial polymerization (Figure 8a) [49]. By enhancing solubility through additives (e.g., HCl or trifluoroethanol), continuous and ultrathin selective layers could be formed. Notably, cage geometry and cavity size play a decisive role in membrane performance: RCC3‐based membranes with larger cavities enable guest encapsulation, whereas tren‐cage‐based membranes exhibit tighter packing, narrower pore size distributions, and near‐complete rejection (>99%) of dyes with molecular weights above 580 g mol⁻¹. Second, supramolecular complexation has been employed to enhance dispersibility and interfacial compatibility. Hua et al. constructed meso‐/microporous colloidal assemblies via host–guest interactions between CC3 and the ionic surfactant dodecyltrimethylammonium bromide (DTAB), forming hydrophilic CC3‐DTAB complexes [75]. Building on this concept, Wang et al. utilized RCC3@DTAB as aqueous‐phase building blocks in interfacial polymerization with terephthaloyl chloride (TPC), leading to dense membranes with topologically structured supramolecular channels (Figure 8b) [96]. Benefiting from improved hydrophilicity and the synergistic effects of size sieving, Donnan exclusion, and channel transport, the resulting membranes exhibited both high permeability (13.9 L m⁻² h⁻¹ bar⁻¹) and excellent selectivity (Na₂SO₄ rejection of 98.5% and NaCl/Na₂SO₄ selectivity of 42.6). Collectively, these strategies demonstrate that regulating the solution chemistry and intermolecular interactions of POCs is critical for their successful integration into TFN architectures, thereby enabling the translation of molecular‐level structural precision into high‐performance ion and molecular separation membranes.

FIGURE 8.

Strategies to enhance aqueous processability of POCs for interfacial polymerization. (a) pH‐regulated solubilization enabling ultrathin polycage membranes with tunable pore structures. Reproduced with permission [49]. Copyright 2023, Springer Nature. (b) Supramolecular complexation enhancing aqueous dispersibility and interfacial compatibility via host–guest interactions. Reproduced with permission [96]. Copyright 2022, Elsevier.

3.3.2. Contra‐diffusion Growth

The contra‐diffusion (or reverse‐diffusion) method is an effective strategy for formation of POC layers on porous substrates. In this approach, two reactive precursors are placed on opposite sides of a support, allowing them to diffuse toward each other and react at or within the substrate. This diffusion‐controlled process enables localized cage formation and growth, facilitating the formation of continuous and uniform POC layers even on substrates with complex pore structures. A representative example was reported by Xu et al., who fabricated a pore‐ordered CC3 membrane on an anodic aluminum oxide (AAO) substrate via the contra‐diffusion growth process (Figure 9) [9]. In this system, triformylbenzene (TFB) and cyclohexanediamine (CHDA) were introduced into two separate chambers, respectively. Through counter‐diffusion and interfacial imine condensation, CC3 cages were formed and subsequently assembled into a continuous membrane layer. Benefiting from the well‐defined window–cavity structure of CC3, the resulting membrane exhibited both high ion selectivity and fast ion transport. Remarkably, the membrane demonstrated exceptional selectivity for monovalent/divalent ion separation, with K⁺/Mg^{2 +}, Na⁺/Mg^{2 +}, and Li⁺/Mg^{2 +} selectivities of 1031, 660, and 284, respectively.

FIGURE 9.

Contra‐diffusion growth of CC3 membranes. Reproduced with permission [9]. Copyright 2022, American Chemical Society. (a) Fabrication of CC3 membrane via the contra‐diffusion method. (b) Ion separation performance of the CC3 membrane.

4. Ion Separation Mechanisms of POC‐Based Membranes

The high ion selectivity of POC membranes arises from the synergistic interplay of multiple mechanisms within their subnanometer channel structure and local environment. In ion separation systems, size sieving serves as the fundamental mechanism, governing ion accessibility through the matching between pore dimensions and hydrated ion sizes. At the same time, ion‐channel interactions provide an additional level of regulation over separation performance. In this review, the separation mechanisms of POC‐based membranes are broadly categorized into two major categories: (i) size sieving, which is primarily governed by structural confinement and ion dehydration‐rehydration effects, and (ii) interaction‐based screening, which arises from interfacial interactions between ions and channel walls. Thus, the overall separation performance is defined by the coupled effects of structural confinement and ion–pore interactions. In the following sections, these two aspects and their interplay are discussed in detail.

4.1. Size‐Sieving Effect

At the structural level, pore characteristics (pore size, surface curvature, channel connectivity, and overall topology) of POC, govern ion transport pathways and migration dynamics. Among various POCs, the pore‐channel characteristics demonstrate distinct variations. For the commonly investigated imine‐type POCs, CC1 has nonconnected pores, CC2 has vertical pores, and CC3 generates an interconnected diamondoid‐pore network [10]. A number of studies have examined the influence of these structural characteristics on transport behavior. For instance, Jiang and co‐workers systematically investigated the water desalination performance of several POC membranes using simulation studies, focusing on pore size, surface curvature, and channel connectivity [100]. The CC1 membrane exhibits negligible water permeability due to its disconnected pore structure. Whereas membranes such as CC3 and CC16, which contain interconnected tetrahedral pore networks, displayed intermediate water permeability. The CC2 membrane possesses straight channels with pore diameters of 3–4 Å, enabling efficient water transport (kinetic diameter ∼2.65 Å) while effectively excluding hydrated ions such as Na⁺ and Cl⁻, achieving nearly complete salt rejection. In contrast, the CC17 membrane possesses larger pore apertures (∼10 Å), which facilitate ion permeation and result in a reduced salt rejection rate to 89%, highlighting the weakening of size sieving with increasing pore size. Yuan et al. focused their research on tetrahedral‐shaped POCs that possess 3D interconnected networks [101]. This study elucidated the structure‐performance link via experiments and simulations of water permeation. Subsequently, Xu's team systematically investigated POC membranes with different cage arrangements and channel wettabilities, revealing that the CC3 membrane with interconnected hydrophobic channels exhibits outstanding ion separation performance [97]. Notably, the pore topology of certain POCs, such as CC3, can be dynamically modulated by the surrounding environment (e.g., solvent conditions), which in turn alters the effective transport pathways and consequently influences the separation performance [89].

Beyond geometric confinement, ion transport is typically accompanied by a coupled dehydration–rehydration process. In aqueous solution, ions are surrounded by hydration shells, and both the hydrated ion size and the energy required for dehydration determine whether an ion can pass through confined channels. For example, the hydration diameters of K⁺, Na⁺, Li⁺, and Mg^{2 +} are 6.62, 7.16, 7.64, and 8.56 Å, respectively, which are significantly larger than the typical window size of CC3 (∼5.8 Å) [102]. As a result, ions must undergo partial dehydration to enter and traverse the pores [103]. It has been reported that the number of hydrogen bonds formed by water molecules within POC channels is lower than that in bulk water, indicating that the confined environment alters the structure and dynamics of water molecules, thereby creating favorable conditions for ion dehydration [104]. This dehydration process introduces a substantial energy barrier that governs ion transport. Xu et al. conducted a comprehensive study of the transport behavior of hydrated ions in POC membranes using MD simulations (Figure 10) [9]. The results showed that ions traversing CC3 channels exhibit a distinct free energy barrier sequence: K⁺ (∼5 kcal mol⁻¹) < Na⁺ < Li⁺ < Mg^{2 +} (∼29 kcal mol⁻¹). Monovalent ions such as K⁺, with relatively lower hydration energy, can more readily undergo partial dehydration and permeate through the POC channels. In contrast, ions with higher charge density, such as Mg^{2 +}, possess more stable hydration shells and require significantly higher energy for dehydration, thereby effectively hindering their transport. These results confirm that dehydration energy is a dominant factor governing ion selectivity. The changes in the hydration state of ions during transport through CC3 window were analyzed using the water radial distribution function (RDF). The results indicate that ions lose part of their first solvation shell upon passing through the CC3 window: K⁺ loses approximately 1.5 water molecules, and Na⁺ loses about one, whereas Li⁺ and Mg^{2 +} exhibit negligible changes in their hydration shells. Consequently, K⁺ and Na⁺ can reduce the dehydration energy barrier by shedding hydration water during migration from the external to the internal cavity, enabling easier passage through the window compared to Li⁺ and Mg^{2 +}.

FIGURE 10.

Ion separation mechanism of CC3 membranes. Reproduced with permission [9]. Copyright 2022, American Chemical Society. (a) MD simulations depicting ion migration pathways within CC3 channels and probable ion positions derived from free energy surfaces at 300 K. (b) Analysis of the number of coordinated water molecules for ions in the external and internal cavities of CC3, along with free energy barriers at four representative positions along the transport pathways.

4.2. Interaction‐Oriented Selectivity

Beyond structural confinement, interaction‐oriented effects provide an additional level of selectivity control in POC‐based membranes. Among these, electrostatic interactions, arising from ionizable functional groups such as carboxyl, amine, and methoxy groups, play a central role by inducing selective attraction or repulsion toward ions, thereby modulating transport behavior. In parallel, other noncovalent interactions further regulate ion migration within confined nanochannels. For example, hydrogen‐bonding interactions can restructure the hydration shell of ions, influencing transport energetics and diffusion barriers, while host–guest interactions within intrinsic cage cavities provide well‐defined recognition sites that enable selective ion transport through combined size exclusion and chemical affinity.

A representative example is the cation‐selective TpOMe‐CDA/AAO nanofluidic membrane, where abundant methoxy (─OCH₃) groups generate negative surface charges under neutral conditions [105]. This electrostatic effect, coupled with hydrophobic internal/external cavities and subnanometer windows, enables preferential transport of cations (e.g., K⁺ and Na⁺) while excluding anions such as Cl⁻. Compared to CC19/AAO membranes with hydrophilic hydroxyl‐functionalized channels, where strong hydrogen bonding hinders ion mobility, the hydrophobic channels in TpOMe‐CDA and CC3 reduce interfacial friction and promote ion slip, resulting in lower transport resistance. Consequently, the TpOMe‐CDA/AAO membrane exhibits superior ion selectivity, higher transference numbers, and enhanced osmotic energy conversion performance, achieving a maximum power density of 81.61 W m⁻² under a 500‐fold NaCl gradient. In addition to intrinsic functional groups, the electrostatic properties of POCs can be deliberately engineered through chemical modification to achieve targeted ion selectivity. For instance, Zhao and coworkers synthesized a highly charged cage (I‐Cage‐Cl), which exhibits strong electrostatic affinity toward CrO₄ ²⁻ ions, enabling efficient capture and selective separation in aqueous environments [65]. More sophisticated control is demonstrated by Wang et al., who developed charge‐tailored POC interfacial layers based on RCC3 derivatives. Positively charged RCC3⁺, obtained via protonation of amine groups, suppresses the shuttle of polyiodide anions (I₃ ⁻/I₅ ⁻) through electrostatic repulsion, significantly improving cycling stability [69]. Conversely, negatively charged RTP‐CC3⁻ selectively anchors vanadyl cations (VO^{2 +}), mitigating cathode dissolution while maintaining rapid Zn^{2 +} transport. Importantly, these electrostatic effects are further enhanced by subnanometer confinement and coordination interactions; for example, Ag nanocrystallites confined within RCC3 pores (Ag@RCC3) enable selective transport of partially dehydrated Zn^{2 +} while excluding hydrated species, mimicking biological ion‐pump mechanisms. This multiscale synergy, which combines electrostatic interactions, size exclusion, and coordination effects, results in enhanced ion selectivity and long‐term electrochemical stability.

In addition to static charge effects, the surface charge of POC membranes can be dynamically regulated by external stimuli, particularly pH. The influence of solution pH on the ion transport behavior of POC membranes has been systematically investigated by Xu and coworkers (Figure 11) [97]. Zeta potential measurements of CC3 membranes over a pH range of 3–9 reveal an isoelectric point at approximately 5.6. Correspondingly, the ion selectivity (K⁺/Mg^{2 +}) was evaluated at different pH values (3, 5, 6, and 8), demonstrating a strong dependence on surface charge. At pH values below the isoelectric point, protonation of tertiary amine groups within the CC3 cage imparts a positive surface charge, which enhances electrostatic exclusion of divalent cations (e.g., Mg^{2 +}), thereby increasing K⁺/Mg^{2 +} selectivity. In contrast, at pH values above the isoelectric point, adsorption of hydroxide ions renders the membrane surface negatively charged. This promotes electrostatic attraction toward cations, leading to increased ion permeation rates but reduced selectivity. Overall, as the pH increases from 3 to 8, ion flux increases while selectivity decreases, highlighting the critical role of pH‐regulated surface charge in ion transport behavior.

FIGURE 11.

pH‐dependent ion transport in CC3 membranes. Reproduced with permission [97]. Copyright 2023, American Chemical Society. (a) Zeta potential measurements and ion permeation rates/selectivities in single‐ion systems (0.1 M KCl or 0.1 M MgCl₂) under pH 3–8. (b) Illustration of positively and negatively charged channel surfaces influence on cation transport.

It is also important to note that surface charge in POC‐based membranes may originate not only from the cage molecules themselves but also from the surrounding membrane matrix. For example, incorporation of CC19 into TFN membranes increases the density of negatively charged carboxyl groups through partial hydrolysis of residual acyl chloride groups, thereby enhancing Donnan exclusion effects [106]. Similarly, in HPABP‐CC3 mixed‐matrix membranes, the positively charged polymer matrix (arising from protonated amine groups) facilitates electrostatic attraction of anions such as Cl⁻ and SO₄ ²⁻, further contributing to selective ion transport [98].

Overall, these interaction‐based effects, together with size sieving, collectively determine the overall ion separation performance of POC membranes. The synergistic interplay between structural confinement and ion‐channel interactions enables simultaneous modulation of transport pathways and energy barriers, thereby governing the ion selectivity and permeability. This integrated mechanism provides fundamental insights into the structure–property relationships of POC‐based ion separation membranes and establishes guiding principles for the rational design of high‐performance, next‐generation ion separation membranes.

5. Design Principles for POC‐Based Membranes in Ion Separation

Guided by the interplay between size‐sieving and interaction‐governed selectivity, the design of high‐performance POC‐based membranes for ion separation requires an integrated consideration of molecular architecture, nanochannel structure, and membrane fabrication across multiple length scales. In contrast to neutral molecule separations, ion transport in confined environments is strongly coupled with hydration structure, charge interactions, and energy barriers associated with partial dehydration. In this context, the intrinsic structural tunability of POCs at the molecular level offers a unique platform to precisely regulate subnanometer transport environments, thereby enabling selective ion transport beyond simple size exclusion. This provides a pathway to simultaneously optimize ion selectivity, transport kinetics, and operational stability.

5.1. Molecular‐Level Design and Channel Engineering

The performance of POC‐based ion separation membranes fundamentally originates from the molecular design of cage building blocks, which dictates pore size, geometry, and structural stability. Key parameters include the geometry and connectivity of organic linkers, where bond angles and linker length define cage dimensions and window apertures, while molecular rigidity governs shape persistence and resistance to structural collapse [107]. These features are particularly critical for ion separation, where subtle differences in ionic radii and hydration shells require angstrom‐level precision in pore size control. Equally important is the choice of bond‐forming chemistry. Dynamic covalent chemistry (DCC), such as imine condensation, enables error correction and the formation of well‐defined cage structures, whereas irreversible covalent bonding enhances chemical and hydrolytic stability under aqueous or harsh ionic environments. Hybrid strategies that combine reversible and irreversible bonding offer a promising approach to balance structural precision with long‐term robustness, which is essential for maintaining stable ion transport pathways.

Building upon molecular design, precise engineering of subnanometer channels is central to achieving high ion selectivity. Continuous and well‐connected transport pathways minimize transport resistance, whereas disordered or poorly interconnected pores lead to nonselective leakage. Critically, rational tuning of pore size relative to hydrated or partially dehydrated ions allows selective transport through a combination of steric exclusion and dehydration‐mediated energy barriers. Furthermore, the chemical environment within nanochannels plays a decisive role in governing ion transport behavior. Functional groups such as charged, polar, or hydrophobic moieties modulate ion–channel interactions, influencing ion hydration structure, energy barriers for dehydration, and transport kinetics. For example, charged sites can introduce electrostatic selectivity, while polar groups can stabilize partially dehydrated ions within confined channels. Therefore, the design of POC nanochannels should balance strong ion–channel interactions, which enhance selectivity, with low transport resistance to maintain high ion flux.

These structural and chemical features can be further tailored through postsynthetic modification, enabling precise regulation of channel functionality and interfacial chemistry within confined spaces.

5.2. Integration Into Membrane Architectures and Scalable Fabrication

Translating molecular‐level control into practical ion separation performance requires membrane fabrication strategies that preserve and propagate the intrinsic advantages of POCs across multiple length scales. Although pure POC membranes can be constructed via direct assembly or interfacial growth, these systems often face challenges in achieving mechanical robustness and scalable fabrication. In this context, MMMs and TFN membranes provide more practical platforms by incorporating POCs into polymer matrices. A central challenge lies in achieving molecular‐level interfacial compatibility between POCs and the surrounding polymer, as interfacial defects can create nonselective pathways that severely compromise ion selectivity. Recent advances demonstrate that interfacial functionalization, supramolecular interactions, and the development of solution‐processable cage derivatives can significantly improve dispersion and interfacial adhesion, thereby enabling defect‐free selective layers and continuous ion transport pathways. Importantly, membrane fabrication should not be viewed merely as a processing step, but as a critical means of translating molecular design into hierarchical structures that govern ion transport. Preserving pore accessibility, channel connectivity, and functional group distribution from the molecular to the membrane scale is essential for maintaining selective ion transport within subnanometer confined channels.

6. Conclusions and Perspectives

POCs possess several distinctive features, including well‐defined subnanometer cavities, tunable pore apertures, modular molecular structures, and excellent solution processability. These characteristics render them highly promising candidates for constructing advanced ion separation membranes. In this review, we establish a unified framework linking molecular design, nanochannel architecture, and membrane fabrication, and highlight how the synergistic interplay between size sieving and interaction‐based selectivity governs ion transport within confined POC channels. This mechanism enables efficient discrimination of ions with similar sizes and physicochemical properties, providing a foundation for the rational design of high‐performance ion separation membranes.

Despite these advances, the field still faces fundamental challenges in transitioning from proof‐of‐concept studies to practical applications. Key challenges include scalable fabrication, long‐term operational stability, and, more fundamentally, a lack of predictive understanding of structure–transport relationships under realistic conditions. Addressing these challenges requires a paradigm shift from empirical optimization toward mechanism‐informed and design‐oriented engineering. Future research should focus on four closely related directions.

6.1. Predictive Membrane Structure Design

Achieving precise control over subnanometer channel structure and chemical environment is critical for the development of next generation POC‐based ion separation membranes. Future studies are expected to move beyond empirical trial‐and‐error approaches toward more rational and predictive design of channel architectures. In particular, the integration of data‐driven methods, such as machine learning combined with multiscale simulations, offers powerful tools for mapping structure–transport relationships and enabling inverse design of cage structures and channel environments tailored for specific ion separations [17, 108]. Such approaches are expected to accelerate materials discovery and enable rational optimization of ion selectivity and transport kinetics.

6.2. Scalable and Controllable Fabrication

Translating POC‐based membranes from laboratory studies to practical applications requires scalable, reproducible, and cost‐effective fabrication strategies. Although POC membranes demonstrate excellent separation performance in both molecular and ion systems, their industrial implementation still faces several key challenges. Future progress will depend on better control over nonequilibrium assembly processes and interfacial structure formation to ensure uniform subnanometer transport channels and minimize defect formation during scale‐up. Processing methods that improve filler dispersion, enhance interfacial compatibility, and suppress structural heterogeneity will be particularly important for achieving consistent membrane performance. In particular, maintaining structural integrity and minimizing nonselective defects under large‐area and continuous fabrication conditions remain critical issues. Meanwhile, addressing challenges related to large‐area fabrication, batch‐to‐batch reproducibility, and long‐term stability is essential before these membranes can move closer to practical deployment.

6.3. Operando and Multimodal Characterization

Advancing the understanding of POC‐based membranes also requires characterization techniques that can probe ion transport under realistic operating conditions. Ion migration within subnanometer confined channels involves coupled effects of confinement, hydration, and ion–channel interactions, which are difficult to capture using conventional ex situ methods. The development of in situ and operando techniques, combined with multimodal spectroscopic and microscopic approaches, will enable time‐ and spatially resolved observation of ion transport pathways and interaction dynamics. Such advances will provide a stronger experimental basis for establishing reliable structure‐transport relationships.

6.4. Mechanistic Understanding

A deeper mechanistic understanding of ion transport is essential for enabling truly rational membrane design. Although ion transport in POC membranes is governed by both size exclusion and specific ion‐channel interactions, their relative contributions and coupled effects under realistic conditions are still not fully understood. This challenge is further complicated by the dynamic nature of solvation and confinement within POC channels. Developing predictive and quantitative models that connect molecular‐level interactions with macroscopic separation performance remains a central goal. Achieving this will require the integration of multiscale simulations, theoretical modeling, and data‐driven approaches. Progress in this direction is expected to move the field from empirical optimization toward mechanism‐guided and ultimately predictive design of high‐performance ion separation membranes.

The future development of POC‐based ion separation membranes will depend on the s coordinated advancement of predictive structural design, scalable fabrication, advanced characterization, and mechanistic understanding. Beyond their role in membrane separation, POCs also provide a valuable model platform for studying ion transport in confined environments, with broad relevance to energy, environmental, and resource‐related applications. With continued progress in these areas, POC‐based membranes are expected to become an increasingly versatile and powerful platform for next‐generation ion separation technologies.

Author Contributions

Tingting Xu: writing – original draft, writing – review and editing, funding acquisition, supervision. Zheng Liu: visualization, formal analysis. Shuhong Zhao: formal analysis, visualization. Yubin He: writing – review and editing. Xingya Li: funding acquisition, writing – review and editing, project administration, supervision. Tongwen Xu: funding acquisition, writing – review and editing, project administration, supervision.

Conflicts of Interest

The authors declare no conflicts of interest.

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.