Oxygen vacancies kinetics in $TaO_{2-h}$/$Ta_{2}$$O_{5-x}$ memristive interfaces

TL;DR

This study combines experiments and simulations to understand oxygen vacancy dynamics in TaO-based memristive devices, demonstrating how OV profiles influence RESET times and enabling optimization for neuromorphic applications.

Contribution

It introduces a method to tune RESET times by selecting specific OV profiles and estimates activation energies for OV electromigration in TaO memristors.

Findings

OV profiles affect RESET time-scale

Gradual RESETs reveal activation energies

OV engineering can optimize switching performance

Abstract

Oxygen vacancies (OV) are pervasive in metal oxides and play a pivotal role in the switching behaviour of oxide-based memristive devices. In this work we address, through a combination of experiments and theoretical simulations, OV dynamics in $Pt/TaO_{2-h}/Ta_{2}O_{5-x}/TaO_${2-y}$/Pt$ devices. In particular, we focus on the RESET transition (from low to high resistance), induced by the application of electrical pulse(s), by choosing different initial OV profiles and studying their kinetics during the mentioned process. We demonstrate that by selecting specific OV profiles it is possible to tune the characteristic time-scale of the RESET. Finally, we show that the implementation of gradual RESETs, induced by applying many (small) successive pulses, allows estimating the activation energies involved in the OV electromigration process. Our results help paving the way for OV engineering aiming at optimizing key memristive figures such as switching speed or power consumption, which are highly relevant for neuromorphic or in-memory computing implementations.