Title: Principles of Hodgkin-Huxley iontronics with two-dimensional nanofluidic memristors

Abstract: New energy-efficient architectures inspired by the brain are growing as an alternative to traditional von Neumann computing. While existing hardware implementations use electrons as charge carriers, advances in nanofluidics have opened the possibilities for ion-based computation, inspired by biological neurons. We develop an analytical theory, backed up by molecular dynamics simulations, predicting neuromorphic behaviour in recently demonstrated two-dimensional electrolytes confined in a subnanometric slit. We show that under an electric field, ions assemble into micelle-like clusters, whose long-timescale dynamics result in a history-dependent conductivity. We carry out molecular simulations of two such devices, reproducing the Hodgkin-Huxley neuron model with experimentally-accessible biomimetic nanofluidic circuitry. The resulting system highlights spontaneous emission of voltage spikes trains characteristic of neuromorphic activity.