Sputtered Oxide-Based Thin-Film Transistors for Neuromorphic Applications

Abstract

Human brain, which has ~1011 neurons interconnected by ~1015 synapses, is regarded as the most intelligent natural computing system. Hardware implementation of brain-inspired computing systems, namely neuromorphic engineering, has attracted increasing attention. A key step for neuromorphic engineering is to develop devices that can realize neural or synaptic behaviors and functions. Among many different types of new-concept neuromorphic devices, electrolyte-gated transistors have been intensively studied in recent years. Electrolyte dielectrics enable conduction of ions and insulation of electrons. Under an applied gate voltage bias, mobile ions migrate to and form electric-double-layers at the interface, resulting in an effective modulation of the channel conductance. The unique ion-mediated modulation makes electrolyte-gated transistors ideal for mimicking the synaptic behaviors, where signals are transmitted via neurotransmitter-mediated ionic fluxes. In this project, indium-gallium-zinc-oxide thin-film transistors gated with sputtered electrolytes were fabricated and proposed for neuromorphic applications. In-depth understanding of the device characteristic was gained after systematically studied the dielectrics, the semiconductors, and the bias-stress behaviours. Tuneable synaptic spiking behaviours were achieved within 1 V by altering the sputtering pressure of the dielectrics, demonstrating short-term synaptic plasticity behaviours. A wide range of memory time tunability over seven orders of magnitude was achieved on the high-pressure sputtered SiO2 electrolytes, demonstrating long-term synaptic plasticity behaviours. Afterwards, the temperature dependence of both the device performance and the synaptic behaviours were studied. Lastly, modulated metaplasticity behaviours, or the second-order synaptic plasticity behaviours were also studied.

Date of Award	31 Dec 2022
Original language	English
Awarding Institution	The University of Manchester
Supervisor	Aimin Song (Supervisor) & Alex Casson (Supervisor)