In nature, biological cell membranes possess selective permeability, dynamically transporting water and essential nutrients to maintain cell viability. This remarkable ability arises from their intricate architecture, which includes smart and highly selective protein channels like aquaporin 1, arranged within an ultrathin lipid bilayer less than 10 nm thick. These qualities make the cell membrane a benchmark, inspiring the design of separation membranes with profound implications for practical applications. In recent decades, the rise of nanotechnology has brought together an abundance of multifunctional nanochannels that are primed to revolutionize separation membranes, offering performance levels that surpass those achieved by state-of-the-art polymeric membranes (Nat. Rev. Mater. 6, 294–312 (2021)). However, building highly interconnected and adaptable pore channels across large-area membranes, comparable in thickness to cell membranes, has encountered challenges related to aggregation, pore channel orientation, and scalability.

Porous organic cages (POCs), as an intriguing type of 0D nanoporous molecule, can form interconnected channel networks independent of the molecule's orientation. These networks demonstrate water channel behavior, offering ultrafast guest permeance and selective separation (Nat. Commun. 11, 4927 (2020)). More importantly, POCs possess good solubility, which provide option for scalable solution-processing. Our group has focused on the cross-scale synthesis of POC-based materials and the development of their biomimetic applications (Nat. Commun. 13, 1471 (2022)). Very recently, we have established a series of methods to regulate the physicochemical structure of the cage skeletons, enabling controllable access for guests to the cage cavity. For example, we design an ionic organic cage-encapsulated metal clusters for switchable catalysis, where the associated anion transferred their designable responsiveness to the active cluster sites (Cell Rep. Phys. Sci. 2, 100546, (2021)). Given these qualities, POC appears to be an ideal candidate for crafting bioinspired membranes.

As part of our long-term goal to fabricate POC-based biomimetic materials, our group has recently developed a free-interface-confined self-assembly & crosslinking (FISC) method in 2024 to fabricate a cell membrane-inspired sub-8 nm networked cage nanofilm (Nat. Commun. 15, 2478 (2024) ). A sharp oil/water free interface was introduced to suppress the intrinsic van der Waals packing and direct their 2D self-assembly; the preorganized cage layers were then in-situ crosslinked into continuous ultrathin networked cage nanofilms within the confined 2D space. Transferring these networked cage nanofilms onto porous supports affords composite membranes with ultrahigh water permeance up to 360 L m^-2 h^-1 bar^-1 and excellent molecular sieving performance well surpasses that of most current membranes, as well as long-term operational stability. By comparing the water permeability of the fragment-based membrane with that of the crystalline membrane, we confirmed that the nanofilms inherited the nanofluidic channels from the POCs. They exhibited exceptional intrinsic water permeability (P_w) at a scale of 10^-5 cm² s^-1, surpassing conventional polymeric membranes by 1-2 orders of magnitude and comparable to that of other nanofluidic membranes such as MOFs, COFs, and GO membranes. Molecular dynamics simulations revealed 1D water chains within the networked cage, mirroring features found in biological water channels.

Furthermore, we explored the adaptivity of the networked cage nanofilm by introducing counteranions with varying hydrophilicity and conformation-responsive to the cage window. This leveraged our established facile counteranion exchange strategy, allowing for controllable separation performance. The calculated  P_w value of the cage membrane paired with hydrophobic counterions TFSI⁻ is ~1.2 times higher compared to the membranes carrying a hydrophilic Cl⁻. Impressively, a light-controlled graded molecular sieving system was established by associating light-responsive anions (azobenzoate) to the POC membrane, in which the continuous separation of three organic dyes in a  single-membrane process was exemplified. The adaptivity of cut-off rejection was attributed to the light-controlled window opening sizes, supported by rejection experiments and MD simulation.

These networked cage nanofilms present an avenue for developing bio-inspired ultrathin membranes for smart separation, more importantly, suggesting that porous organic cages may be potential building blocks for mimicking the separation behavior of cell membranes, given their flexibility in skeleton design, self-assembly, and solution processing.