An Artificial Neuron Founded on Resistive Switching of Mott Insulators

Abstract

Narrow-gap Mott insulators exhibit under electric field a resistive switching related to the formation of a conducting filamentary path made of metastable metallic domains. When this effect is non-volatile it can be used to build up a new type of Resistive Random Access Memory called Mott memory. But we show here that when the resistive switching remains volatile it is of great interest for neuromorphic applications different than artificial synapses implemented by the so-called Memristors. Specifically, we show that under electric field the dynamics of the creation and destruction of the metastable metallic domains implement the three basic functions Leaky Integrate and Fire of artificial neurons. The central result of the present work is therefore to demonstrate that a simple two terminal device made of Mott insulator can be considered as an analogue of an artificial Leaky Integrate and Fire neuron.