High-throughput design of optoelectronic-ferroelectric heterostructure from materials to sensor–memory–computing devices

Abstract Ferroelectric-based artificial synapses have emerged as a fascinating candidate for developing intelligent sensor–memory–computing (SMC) systems, thanks to the remarkable nonvolatile properties and abundant polarization states that ferroelectrics offer. However, simultaneously modulating the ferroelectric synapse through optical and electrical excitation is challenging. Herein, we propose a high-throughput strategy for designing optoelectronic co-modulated ferroelectric synapses. This strategy involves designing a Ferroelectric field-effect transistor (FeFET) based on the Pb(Zr0.2Ti0.8)O3 (PZT)/InGaZnO (IGZO) heterostructure, which includes an IGZO homostructure, followed by high-throughput screening of IGZO materials that enable both optical and electrical modulation. The transistors with optoelectronic co-modulated synaptic functionalities are subsequently screened from a set of high-throughput FeFETs. Based on these optoelectronic co-modulated ferroelectric synapses, an artificial SMC system which can simultaneously sense and recognize images is constructed, achieving a high recognition accuracy of 88.42%, and this SMC system simultaneously exhibits the advantages of reduced hardware overhead, fast speed, and low power consumption. Our work introduces a novel strategy for designing multifunctional artificial synapses from materials to devices, which may represent a new paradigm in the development of high-performance SMC systems.