EGaIn tube memristors offer reliable switching on a biological time scale
🤖 AI Summary
Existing memristors suffer from substantial experimental variability and temporal dynamics orders of magnitude slower than biological neurons (millisecond-scale), hindering their deployment in neuromorphic computing and in-memory processing. To address this, we propose a tubular memristor based on the liquid-metal eutectic gallium–indium (EGaIn), wherein reversible growth/dissolution of a surface oxide layer is electrochemically modulated by applied voltage, enabling bulk resistive switching in aqueous electrolytes. The device operates robustly within biologically relevant timescales (tens of milliseconds), exhibiting exceptional cycle-to-cycle consistency and low device-to-device variability. Furthermore, we report, for the first time, its fractional-order dynamical response—a hallmark of memory-dependent, non-Markovian behavior. Functional demonstration includes reconfigurable logic gate operation. This work establishes a new paradigm for energy-efficient, highly robust memristive hardware tailored to bio-inspired computing.
Application Category
📝 Abstract
Memristive devices have been considered promising candidates for nature-inspired computing and in-memory information processing. However, experimental devices developed to date typically show significant variability and function at different time scales than biological neurons and synapses. This study presents a new kind of memristive device comprised of liquid-metal eutectic gallium indium (EGaIn) contained within a mm-scale tube that operates via a bulk, voltage-dependent switching mechanism and exhibits distinct unipolar resistive switching characteristics that occur on a biological time scale (tens of milliseconds). The switching mechanism involves voltage-controlled growth and dissolution of an oxide layer on the surface of the liquid metal in contact with an aqueous electrolyte. Through comprehensive measurements on many devices, we observed remarkably consistent cycle-to-cycle behavior and uniformity in the voltage-controlled memristance. We present our findings, which also include an experimental demonstration of logic gates utilizing EGaIn tube memristors. Furthermore, we observe both accelerated and decelerated switching behaviors and identify signatures indicative of a fractional dynamic response.
Problem
Research questions and friction points this paper is trying to address.
Develops reliable memristors for biological time scale computing
Addresses device variability in nature-inspired computing systems
Demonstrates liquid-metal memristor with consistent switching behavior
Innovation
Methods, ideas, or system contributions that make the work stand out.
Liquid-metal EGaIn tube memristor design
Voltage-controlled oxide layer switching mechanism
Biological-time-scale unipolar resistive switching
🔎 Similar Papers
No similar papers found.
