Solvent Structure and Dynamics Controlled Memristive Ion Transport in Å-Scale Channels

Memristors are resistive elements whose conductance depends on past bias history. They are central to non-Von Neumann computing architectures, neuromorphic circuits, and memory-in-logic devices. Nanofluidic memristors extend this concept to ionic and solvent systems, where history-dependent conduction has been realized through ion accumulation, surface charge regulation, phase transitions, etc. Here we demonstrate a distinct mechanism: solvent structure and dynamics mediated memory in an Å-scale graphene channel connecting two electrolyte reservoirs with three different solvents having different hydrogen bonding relaxation time scales. Using molecular dynamics simulations, we find pronounced I-E (current-electric field) hysteresis in NaCl conduction across methanol, water, and formamide/water mixture solvents. The hysteresis originates from the coupled evolution of in-channel ion concentration and hydrogen bond network order, i.e., elevated ion concentrations in the channel at high electric fields disrupt solvent ordering and enhance ion mobility by lowering solvent drag. The magnitude of hysteresis follows the hierarchy of the hydrogen bond relaxation time scales inside the nanochannel, i.e., formamide/water > water > methanol. These results demonstrate that solvent structure can act as a tunable internal state variable for nanofluidic memristors, enabling programmable iontronic devices with memory for applications in neuromorphic computing.