Monovalent Ion Selectivity Incomparability of Graphene Oxide Membrane in Different Separation Systems: A Microscopic Understanding

Abstract

Precise monovalent ion sieving is essential for water treatment and resource recovery. However, ion selectivity reported from diffusion-based and electro-driven membrane processes is often inconsistent and difficult to reconcile due to unclear mechanistic understanding. Herein, combining experiments with molecular dynamics (MD) simulations, we reveal divergent microscopic mechanisms governing monovalent ion selectivity in graphene oxide (GO) nanochannels under different driving forces. Under concentration gradient-driven diffusion, ion selectivity is predominantly regulated by hydration-mediated effects (i.e., partial ion dehydration and dehydration-induced ion–membrane interactions) and cation–anion interaction. Specifically, in simulated defect-free GO nanochannel, Li+, with higher hydration energy and charge density than Na+, possesses a more intact hydration shell and stronger cation–anion interaction, which limit partitioning but reduce transport resistance, resulting in a Li+/Na+ selectivity of 3.3. Conversely, under electric field-driven electromigration, a reversed selectivity (Li+/Na+ = 0.4) occurs due to enhanced cation–anion associations, where stronger Li+–Cl– pairing suppresses both cation partitioning and mobility, generating ion association-dominated selectivity. These results demonstrate that ion selectivity is not an intrinsic membrane property but is strongly governed by the driving-force-dependent transport mechanism. This study resolves the long-standing incomparability in ion selectivity across different separation systems and provides mechanistic guidance for designing next-generation membranes with tailored ion selectivity.