The escalating global water scarcity, driven by climate change and pollution, necessitates advanced water purification technologies. Membrane-based separation, offering low energy consumption and high efficiency, is becoming increasingly important. However, achieving both precise ion sieving and ultrafast water flux simultaneously remains a significant challenge. Conventional hydrophilic nanofiltration (NF) membranes, such as graphene oxide (GOMs) and MXene membranes, achieve ion sieving by controlling pore size at the atomic scale, but this inevitably reduces water flux. Recent research has explored fast water flow in atomic-scale graphene capillaries, but rGO-based membranes often suffer from restacking, creating larger pores and hindering ion sieving. Thermal-membrane coupled technologies, like membrane distillation, partially address this, but their energy consumption remains high. Biological ion channels, with their precise and fast ion sieving and gating capabilities, serve as inspiration for developing NF membranes that can overcome this challenge. Ideally, such membranes should selectively gate water molecules while blocking ion diffusion; however, the similar sizes of water molecules and hydrated ions make this difficult. This research aims to create a membrane capable of achieving this 'ideal' scenario, demonstrating an anomalous gating mechanism capable of selectively and rapidly gating water molecules while effectively inhibiting ion diffusion under forward osmosis.

Existing membrane technologies for water purification face a trade-off between ion sieving efficiency and water permeability. Hydrophilic membranes, while effectively sieving ions by tuning pore size, suffer from reduced water flux due to steric hindrance and hydrogen bonding. Graphene oxide membranes (GOMs) and MXene membranes are examples of this limitation. In contrast, atomic-scale graphene capillaries offer the potential for fast, frictionless water flow, but the tendency for restacking in reduced graphene oxide (rGO) membranes limits their ion sieving capabilities. Attempts to utilize atomic-scale capillaries within GOMs using thermal-membrane coupled methods like membrane distillation and pervaporation have shown promise, but energy consumption remains an issue. Biological systems provide inspiration, with their ion channels exhibiting both precise and fast ion sieving, controlled through gating mechanisms triggered by external stimuli. The challenge is to create an analogous system for water molecules, selectively gating water while blocking ion diffusion, a seemingly impossible task due to the similar sizes of water molecules and hydrated ions.

The researchers synthesized a 2D island-on-nanosheet heterostructure by catalytically growing a metal-organic material, Ni-p-phenylenediamine (Ni-pPD), on reduced graphene oxide (rGO) nanosheets using the oxygenated functional groups on GO as catalytic sites. This resulted in Ni-pPD monolayer islands uniformly distributed on the rGO nanosheets, confirmed by energy dispersive X-ray (EDX) mapping, Fourier-transform infrared spectroscopy (FTIR), and X-ray photoelectron spectroscopy (XPS). Density functional theory (DFT) simulations supported the formation of a planarized Ni-pPD structure on rGO due to π–π stacking. The spontaneous restacking of these Ni-pPD@rGO nanosheets during the reduction process created the desired atomic-scale graphene capillaries. Atomic force microscopy (AFM) and high-resolution transmission electron microscopy (HRTEM) confirmed an interlayer spacing of 0.96 nm, corresponding to 6 Å graphene capillaries with 2.6 Å and 6.2 Å free spaces, further validated by DFT calculations. Metal-organic pillared graphene membranes (MOGMs) were fabricated by vacuum filtration and controlled evaporation. SEM and EDX confirmed the membrane structure. Water contact angle measurements showed MOGMs possessed a hydrophobic nature (84°), unlike GOMs (44.5°), indicating the presence of atomic-scale graphene capillaries. X-ray diffraction (XRD) patterns also confirmed the unique restacked structure of Ni-pPD@rGO with four distinct peaks. Pore size distribution analysis using BET measurements, applying a hybrid nonlocal DFT (NLDFT) kernel with N2 and CO2 gases, revealed the presence of sub-nano sized pores. Permeation experiments were conducted using a U-shape permeation system, employing various alkali-metal salt solutions (0.1 M KCl, NaCl, MgCl2) as draw solutions. In-situ FTIR spectroscopy investigated the water structure within the graphene capillaries under different humidity conditions. Molecular dynamics (MD) simulations were performed to investigate the permeation mechanism at the molecular level.

The fabricated MOGMs exhibited anomalous permeation behavior under osmosis. With no hydrostatic pressure difference, water flux was negligible, indicating significant suppression of water diffusion. However, applying a small hydrostatic pressure (as low as 7.8 × 10⁻³ bar) triggered a dramatic increase in water flux (up to 2.5 L m⁻² h⁻¹), significantly faster than GOMs under similar conditions. The water flux exhibited a reversible on/off switching behavior controlled by hydrostatic pressure. Simultaneously, ion permeation rates were significantly suppressed (≈10⁻¹⁰ mol m⁻² h⁻¹), three orders of magnitude lower than GOMs. In-situ FTIR spectroscopy revealed the formation of a highly ordered, hydrogen-bonded water network within the 6 Å graphene capillaries under low humidity conditions. The ordered water structure exhibits stronger hydrogen bonding than bulk liquid water and ice. DFT calculations confirmed a close-to-hexagonal arrangement of water molecules within the capillaries, with a water density consistent with measurements in GO membranes. The high capillary pressure (~1000 bar) generated within the capillaries is thought to compact the water molecules and induce the ordered structure. Molecular dynamics (MD) simulations supported these findings showing that under low hydrostatic pressure, water flows through as a bulk liquid and not via individual diffusion. The ordered water structure hinders diffusion under osmosis, while facilitating fast bulk flow of water. The observed results showcase an anomalous gating mechanism which achieves simultaneously high water flux and significant ion rejection.

The findings demonstrate a novel approach to water purification, overcoming the traditional trade-off between ion sieving and water flux. The anomalous gating behavior is attributed to the unique properties of water confined within the atomic-scale graphene capillaries, forming a highly ordered structure that inhibits diffusion but allows for ultrafast bulk flow. This liquid-solid-liquid phase change mechanism enables precise and ultrafast molecular sieving. The observed results have significant implications for designing next-generation water purification membranes, potentially applicable to various fields beyond desalination, including industrial water treatment and other molecular separation processes. The reversible on/off switching capability is also highly advantageous for applications requiring controlled water flow. The mechanism also suggests that similar phenomena could be found in other nano-confined systems such as metal-organic frameworks (MOFs) and covalent-organic frameworks (COFs).

This study successfully fabricated metal-organic pillared graphene membranes (MOGMs) with atomic-scale graphene capillaries exhibiting anomalous water molecular gating. This achieved both ultrafast water flux and negligible ion permeation, controlled by hydrostatic pressure. The underlying mechanism involves a highly ordered, hydrogen-bonded water network within the capillaries. This liquid-solid-liquid phase change approach opens new avenues for designing advanced membrane-based separation technologies. Future research could explore different pillar materials and membrane structures to further optimize performance and investigate the scalability and long-term stability of MOGMs.

While the MOGMs demonstrate exceptional performance, several limitations warrant consideration. The fabrication process, involving multiple steps, might not be easily scalable for large-scale production. The long-term stability of the membrane under various operating conditions needs further investigation. The current study primarily focuses on alkali-metal ions; further research is required to assess its effectiveness with other types of ions and molecules. Finally, The effect of defects on the membrane performance needs to be further investigated.