Water scarcity is a pressing global issue, and membrane-based water treatment offers an energy-efficient and sustainable solution. Two-dimensional (2D) materials, with their sub-nanometer channels, show promise for precise molecular and ionic sieving in desalination membranes. However, many 2D materials (e.g., graphene oxide, MXene) suffer from swelling or disintegration in hydrated states, hindering their application. This study investigates polymeric carbon nitride (CN), a chemically and thermally stable 2D material, as a potential solution. While CN membranes have been explored for filtration, their random stacking and structural defects limit their effectiveness in reverse osmosis and forward osmosis applications. This paper aims to address this limitation by activating and stabilizing CN's sub-nanometer transport channels, creating a lamellar membrane structure with controlled permeation behavior for enhanced water desalination.

Existing literature highlights the potential of 2D materials for water desalination. Graphene oxide and MXene have been investigated, but their hydrophilic groups lead to swelling and instability in aqueous solutions. Polymeric carbon nitride (CN) has emerged as a stable alternative, with various fabrication methods explored. However, previous studies primarily focused on filtration and separation of large-sized dyes or nanoparticles, while its application in reverse osmosis or forward osmosis, requiring precise sub-nanometer sieving, has been limited due to random stacking and structural defects. The inherent ~3.2 Å stacking distance in CN, theoretically suitable for water transport while blocking salt ions, has yielded counterintuitive experimental results, attributed to these defects.

The authors developed a novel approach to activate and stabilize the transport channels of CN at the sub-nanometer level. They used a Keggin cluster polycation, [Al₃₀O₄(OH)₅₆(H₂O)₂₄]¹⁸⁺ (Al₃₀), as a pillaring agent to interact with the electron-rich nitrogen sites in the CN framework. The Al₃₀ intercalation between CN layers widens the interlayer spacing, creating a regular transport passage. The resulting Al₃₀-CN composite (ACN) was characterized using X-ray diffraction (XRD), transmission electron microscopy (TEM), scanning electron microscopy (SEM), atomic force microscopy (AFM), Fourier-transform infrared spectroscopy (FTIR), and zeta potential measurements. Membrane fabrication involved vacuum filtration of the ACN suspension onto a polydopamine-coated polyether sulfone (PDA-PES) filter, followed by washing to remove excess Al species. The effect of drying temperature on membrane structure and stability was investigated. Ion permeation tests were conducted using a customized H-shaped cell with various salt solutions (NaCl, KCl, LiCl, CaCl₂, MgCl₂). Water flux and salt rejection were evaluated in a forward osmosis setup using 0.1 M NaCl feed and 2 M sucrose draw solutions. The pH-dependent permeation behavior was studied by adjusting the pH of the feed solution, and ¹H CP/MAS SSNMR was used to investigate the microenvironmental changes in the ACN nanostructure under different pH conditions. Density functional theory (DFT) calculations were performed to understand the interaction between Al₃₀ and CN.

The Al₃₀ pillaring effectively widened the interlayer spacing of CN from ~3.2 Å to ~9.8 Å, creating subnanochannels of approximately 6.6 Å in width. The ACN membrane showed excellent anti-swelling properties with minimal interlayer spacing fluctuation in dry and hydrated states (5.6–5.9 Å). The lamellar structure of the ACN membrane contrasted sharply with the randomly stacked CN membrane. In forward osmosis, the ACN membrane achieved a high water flux of 6 L m⁻² h⁻¹ and a salt rejection rate exceeding 99.5%, significantly outperforming conventional CN membranes and many reported 2D membranes. The ACN membrane exhibited tunable permeation behavior, with higher water flux in alkaline conditions (pH 12) than in acidic conditions (pH 2), a phenomenon explained by ¹H CP/MAS SSNMR analysis showing changes in hydrogen bonding and water mobility within the subnanochannels. The ion permeation behavior in ACN membranes followed a size-exclusion effect, with lower permeation rates for divalent ions compared to monovalent ions due to higher dehydration energy barriers. Long-term stability tests showed minimal performance degradation after multiple drying-wetting cycles and prolonged operation in acidic and alkaline environments. The DFT calculations supported the experimental findings by showing higher adsorption energy of Al30 onto CN compared to Al13.

The findings demonstrate the successful fabrication of a highly efficient and stable water desalination membrane based on a pillared lamellar structure of carbon nitride. The use of Al₃₀ as a pillaring agent effectively addresses the limitations of conventional CN membranes by creating a well-defined, stable sub-nanometer channel structure. The high water flux and salt rejection performance breaks the permeability-selectivity trade-off, offering significant advantages over existing technologies. The tunable permeation behavior opens up possibilities for advanced water treatment applications requiring controlled water gating. The results contribute significantly to the development of next-generation 2D membranes for water purification.

This study successfully demonstrates the creation of a high-performance water desalination membrane by activating and stabilizing the transport channels of carbon nitride using aluminum polycations as pillars. The resulting lamellar structure shows excellent swelling resistance, high water flux, and high salt rejection, outperforming many existing 2D membranes. The tunable water permeation behavior in different pH environments further enhances its potential. Future research could explore the integration of the membrane's semiconductor properties to create membrane reactors that combine separation and catalytic capabilities.

The study focused primarily on NaCl and a limited set of other ions. Further investigation is needed to assess the membrane's performance with a wider range of ions and complex water matrices. While the membrane showed excellent long-term stability under various conditions tested, long-term field trials would be beneficial to fully assess its long-term durability and fouling resistance under realistic operating conditions. The DFT calculations used a simplified model of the CN structure, and more complex models could provide further insights.