Understanding the metal-to-insulator transition in La1\u2212xSrxCoO3\u2212\u03b4 and its applications for neuromorphic computing

Shenli Zhang,Giulia Galli

doi:10.1038/s41524-020-00437-w

Shenli Zhang, Giulia Galli

Open Access

https://doi.org/10.1038/s41524-020-00437-w

Copy DOI

Abstract

Transition metal oxides that exhibit a metal-to-insulator transition (MIT) as a function of oxygen vacancy concentration are promising systems to realize energy-efficient platforms for neuromorphic computing. However, the current lack of understanding of the microscopic mechanism driving the MIT hinders the realization of effective and stable devices. Here we investigate defective cobaltites and we unravel the structural, electronic, and magnetic changes responsible for the MIT when oxygen vacancies are introduced in the material. We show that, contrary to accepted views, cooperative structural distortions instead of local bonding changes are responsible for the MIT, and we describe the subtle interdependence of structural and magnetic transitions. Finally, we present a model, based on first principles, to predict the required electric bias to drive the transition, showing good agreement with available measurements and providing a paradigm to establish design rules for low-energy cost devices.

Highlights

The search for computing architectures with low-power consumption is an active field of research, and neuromorphic architectures, which aspire to mimic the human brain[1], have attracted much attention lately
Systems showing tunable resistance states, induced by an external electrical bias. These transition metal oxides (TMO) exhibit a metal-to-insulator transition (MIT) as a function of pressure, temperature, or doping, which may be designed to mimic the behavior of neurons and synapses in the presence of stimuli[5,6,7]
A similar transition was observed between SrCoO3 and SrCoO2.5 (SCO2.5)[16,17] and oxygen vacancy concentrations have been varied in multiple ways in cobaltites, e.g. by depositing oxygen-scavenging metals[15], annealing in reducing environment[16], using electric fields[17,18,19] and with epitaxial strain[20]

Summary

Introduction

The search for computing architectures with low-power consumption is an active field of research, and neuromorphic architectures, which aspire to mimic the human brain[1], have attracted much attention lately. We start by analyzing the structural distortions associated with the topotactic transformation between perovskite and BM phases in LCO and LSCO, as a function of oxygen vacancy concentrations (VO, including 4.2, 8.3, and 12.5%, where VO = δ/3 × 100%).

Results

Conclusion