A covalent organic polymer-based transistor with multifunctional memory and synaptic functions

Abstract

Inspired by biological systems that are highly connected and can process massive amounts of data, neuromorphic transistors may overcome the traditional von Neumann bottleneck. In this work, an organic synaptic transistor was fabricated with a covalent organic polymer MT-TP to mimic the behavior of biological synapses. The device exhibits hole trapping/detrapping ability and outstanding stability. The common biological synaptic behaviors were successfully simulated by the MT-TP synaptic transistor. Eventually, an artificial neural network was constructed and the total accuracy for identifying numbers could reach 86.2%. The results demonstrate that the MT-TP based organic synaptic transistor is a good candidate for future computing systems.