Photo‐Assisted Ferroelectric Domain Control for α‐In2Se3 Artificial Synapses Inspired by Spontaneous Internal Electric Fields

Seok‐Ju Kang,Wonzee Jung,Oh Hun Gwon,Han Seul Kim,Hye Ryung Byun,Jong Yun Kim,Seo Gyun Jang,BeomKyu Shin,Ojun Kwon,Byungjin Cho,Kanghoon Yim,Young‐Jun Yu
DOI: https://doi.org/10.1002/smll.202307346
IF: 13.3
2024-01-13
Small
Abstract:White‐light illumination into α‐In2Se3‐based ferroelectric memories generates the enhanced downward built‐in electric field (Fbi) within the α‐In2Se3 channel due to the upward shift of α‐In2Se3 Fermi level. This intensification of downward Fbi, causing preferential downward reorientation of α‐In2Se3 ferroelectric domains, improved the linearity of long‐term‐depression characteristics. White‐light‐assisted artificial neural networks have significantly improved the recognition accuracy for hand‐written digit numbers. α‐In2Se3 semiconductor crystals realize artificial synapses by tuning in‐plane and out‐of‐plane ferroelectricity with diverse avenues of electrical and optical pulses. While the electrically induced ferroelectricity of α‐In2Se3 shows synaptic memory operation, the optically assisted synaptic plasticity in α‐In2Se3 has also been preferred for polarization flipping enhancement. Here, the synaptic memory behavior of α‐In2Se3 is demonstrated by applying electrical gate voltages under white light. As a result, the induced internal electric field is identified at a polarization flipped conductance channel in α‐In2Se3/hexagonal boron nitride (hBN) heterostructure ferroelectric field effect transistors (FeFETs) under white light and discuss the contribution of this built‐in electric field on synapse characterization. The biased dipoles in α‐In2Se3 toward potentiation polarization direction by an enhanced internal built‐in electric field under illumination of white light lead to improvement of linearity for long‐term depression curves with proper electric spikes. Consequently, upon applying appropriate electric spikes to α‐In2Se3/hBN FeFETs with illuminating white light, the recognition accuracy values significantly through the artificial learning simulation is elevated for discriminating hand‐written digit number images.
materials science, multidisciplinary,chemistry, physical,physics, applied, condensed matter,nanoscience & nanotechnology
What problem does this paper attempt to address?