Ionic species programmable synaptic plasticity in multimodal nanofluidic devices

TL;DR

Researchers developed a nanofluidic device that can be programmed to mimic synaptic plasticity by changing ionic species and concentration, enabling tunable circuits for neuromorphic computing.

Contribution

A concentration-dependent transition between capacitive and inductive hysteresis is demonstrated, enabling programmable synaptic plasticity without structural changes.

Findings

A concentration-dependent transition between capacitive and inductive hysteresis was observed in gold-nanoparticle-stacked nanochannels.

Programmable synaptic plasticity (facilitation and depression) is achieved by altering ionic species without structural reconfiguration.

A tunable high-pass filter circuit was implemented using two identical nanofluidic devices.

Abstract

Nanofluidic devices have been widely utilized to simulate electronic functionalities recently, due to their unique ion transport behaviors, such as non-linear ion transport, selectivity etc. However, the correlation between the ion transport behavior and the transitions among various nanofluidic capacitive and inductive hysteresis still remains poorly understood, which impedes the development of nanofluidic systems. Here, we report a concentration-dependent transition between capacitive and inductive hysteresis in gold-nanoparticle-stacked nanochannels. Quantitative analysis reveals that this transition is governed by the interionic distance relative to the Bjerrum length, establishing a universal mechanism for ion transport modulation. Notably, our system enables unidirectional plasticity (both facilitation and depression) by simply altering the ionic species, demonstrating programmable plasticity without structural reconfiguration. Additionally, a high-pass filter (HPF) circuit with tunable cut-off frequency is implemented through two identical nanofluidic devices. These findings establish a new paradigm for multifunctional nanofluidic devices and provide a rational foundation for the design of aqueous-phase neuromorphic computing circuits.

A single nanofluidic device made of gold-nanoparticle-stacked channels can be reprogrammed with the electrolyte: concentration flips its electrical response, and ion species selects synapse-like facilitation or depression for tunable ionic high pass filter circuits.

Figures