Human Memory‐Inspired MoS 2 /BST Ferroelectric Phototransistor with Synchronous Sensory, Short‐Term and Long‐Term Memories

Abstract The collaborative working mode of different memory patterns–sensory, short‐term, and long‐term memories–in the human brain is significantly efficient and economical. This work reproduces the three memory patterns synchronously with the retention of 60 s, 30 min, and 18 h, respectively, on a...

Full description

Saved in:
Bibliographic Details
Published inAdvanced functional materials
Main Authors Tan, Chao, Wu, Haijuan, Lin, Zhicheng, Liu, Jinxiu, Yang, Lei, Gao, Libin, Wang, Zegao
Format Journal Article
LanguageEnglish
Published 10.10.2024
Online AccessGet full text

Cover

Loading…
More Information
Summary:Abstract The collaborative working mode of different memory patterns–sensory, short‐term, and long‐term memories–in the human brain is significantly efficient and economical. This work reproduces the three memory patterns synchronously with the retention of 60 s, 30 min, and 18 h, respectively, on a Ba 0.6 Sr 0.4 TiO 3 ferroelectric MoS 2 phototransistor, showing the potential of vision perceptual‐like in‐memory sensing and computing ability. It also can be programmed to function as an optical multiply and accumulate (MAC) operator, in‐memory Boolean logic gates, and optical‐electrical cooperative synapses used for artificial neural network (ANN) pattern recognition. In addition, the current on/off ratio, subthreshold swing, and carrier mobility of the device is observed to reach 10 5 , 72 mV dec −1 and 31 cm 2 V −1 s −1 , respectively. Also, its photodetection range can be extended from visible to near‐infrared by gate tuning. The responsivity and specific detectivity at 400 nm wavelength are 171452 A W −1 and 1.3 × 10 14 Jones, respectively, while that at 1100 nm are 605 A W −1 and 3.2 × 10 11 Jones, respectively. This work demonstrates that the ferroelectric phototransistor is capable of photodetection, parallel computing, and synaptic plasticity based on various memory patterns, so as to realize the real in‐memory sensing and computing functionality.
ISSN:1616-301X
1616-3028
DOI:10.1002/adfm.202414186