Realization of tunable artificial synapse through ambipolar charge trapping in organic transistor with pentacene/poly(α-methylstyrene) architecture

Abstract

Transistor-based artificial synapses are expected to tackle the inherent limitations of traditional von Neumann architecture for neuromorphic computing paradigm. Organic electronic materials are promising components of future neuromorphic systems, but mimicking the functions of biological synapses for symmetric weight update and desired variation margin still remains challenging. Here, we propose a synaptic transistor based on pentacene/poly(α-methylstyrene) (PαMS) architecture capable of exhibiting the main behavior of a biological spiking synapse. The ambipolar charge trapping of the transistor enables symmetric variation of the channel conductivity with desirable margin. Comprehensive synaptic functions, including the postsynaptic current with different pulse amplitudes, short-term to long-term plasticity transition, reversible channel conductance potentiation and depression, and repetitive and symmetrical learning processes, are emulated. The realization of essential synaptic functions based on the cumulative charge trapping of pentacene/PαMS structure provides a feasible device structure toward the future demand of neuromorphic computing.