Insights into Desalination Performance of Spacing-Tunable Laminated MXene Membranes via Molecular Dynamics and Machine Learning

Abstract

Developing desalination membranes that combine high water permeance with high ion rejection remains important for freshwater production. Laminated MXene membranes are promising because their surface chemistry and channel dimensions are tunable, yet the respective roles of sheet-edge openings and interlayer spacing in staggered channels remain unclear. We performed molecular dynamics (MD) simulations of 242 O- and F-terminated Ti3C2 MXene membrane systems, varying the sheet-edge gap and interlayer spacing, d, from 6 to 16 Å. Machine learning (ML) analysis identified gap as the main structural variable governing water permeance and the extent of ion penetration, while d further regulated transport within the interlayer channels. Water and ion density distributions showed that gap primarily controlled entry into the laminated membrane, whereas d affected postentry transport and ion hydration. Na–Ow radial distribution functions and hydration-loss analyses indicated that narrow interlayer channels mainly perturbed the second hydration shell of Na+. Under matched geometries, surface termination produced weaker and more localized changes in water and ion distributions. These results distinguish entry control by the sheet-edge gap from postentry regulation by interlayer spacing and provide molecular guidance for designing spacing-tunable laminated MXene membranes.