Broad-spectrum tunable and low switching voltage memristor based on CsBi3I10 for artificial optoelectronic synapses

X Xiaohui Zhang H Haowen Qian (State Key Laboratory of Flexible Electronics (LoFE) & Institute of Advanced Materials (IAM), Nanjing University of Posts and Telecommunications (NJUPT) 3 , Nanjing 210023,) Y Yunqing Yang M Mingdong Yi (State Key Laboratory of Flexible Electronics (LoFE) & Institute of Advanced Materials (IAM), Nanjing University of Posts and Telecommunications (NJUPT) 3 , Nanjing 210023,) J Jiesheng Niu (Beijing Guo Wang Shengyuan Intelligent Terminal Technology Co. Ltd 3 ., Beijing 100080,) Y Yilin Fan (Key Laboratory of Quantum Materials Under Extreme Conditions in Shandong Province, School of Physics Science & Information Technology, Liaocheng University 1 , Liaocheng 252059,) K Kai Wang H Huiyun Zhang G Guangyu Wang C Cailong Liu (Laboratory of Quantum Materials Under Extreme Conditions in Shandong Province, School of Physics Science and Information Technology, Liaocheng University 1 , Liaocheng 252000,) W Wen Li

Abstract

Artificial optoelectronic synapses exhibiting both electrically and light-triggered synaptic behaviors have garnered much attention in recent years for applications in neuromorphic computing and artificial vision systems. However, achieving the coexistence of wide spectral responsivity and low switching voltages remains a critical challenge. Herein, we have demonstrated the fabrication of a Pt/CsBi3I10/PbS quantum dots (QDs)/FTO memristor that exhibits stable bipolar resistive switching, coupled with broad-spectral responsiveness and low switching voltages. Leveraging the localized electric field enhancement effect and anion reservoir functionality of PbS QDs, the device delivers high-performance memristive characteristics: low switching voltages (∼0.1 V/−0.25 V), high on/off ratio (∼102), high endurance (>100 cycles), and retention capability (>103 s). It effectively emulates core synaptic functions encompassing long-term potentiation/depression and spike-timing-dependent plasticity under electrical stimulation, which enables an artificial neural network constructed using this device to achieve a recognition accuracy of 93.37% after 50 training epochs. The device displays wavelength-dependent tunable memristive behaviors under 660/520/405 nm illumination, thereby simulating the adaptive behavior of biological visual systems, accompanied by recognition accuracy from 16.68% to 91%. This work is anticipated to facilitate the advancement of artificial optoelectronic synapses featuring broad-spectrum tunability and low switching voltages.

Article Details

Volume / Issue Vol. 129, Issue 1
Published July 06, 2026
ISSN 0003-6951
Publisher American Institute of Physics

Journal Info

Applied Physics Letters

American Institute of Physics

ISSN: 0003-6951 Physical Sciences

Authors (11)

X

Xiaohui Zhang

H

Haowen Qian

State Key Laboratory of Flexible Electronics (LoFE) & Institute of Advanced Materials (IAM), Nanjing University of Posts and Telecommunications (NJUPT) 3 , Nanjing 210023,

Y

Yunqing Yang

M

Mingdong Yi

State Key Laboratory of Flexible Electronics (LoFE) & Institute of Advanced Materials (IAM), Nanjing University of Posts and Telecommunications (NJUPT) 3 , Nanjing 210023,

J

Jiesheng Niu

Beijing Guo Wang Shengyuan Intelligent Terminal Technology Co. Ltd 3 ., Beijing 100080,

Y

Yilin Fan

Key Laboratory of Quantum Materials Under Extreme Conditions in Shandong Province, School of Physics Science & Information Technology, Liaocheng University 1 , Liaocheng 252059,

K

Kai Wang

H

Huiyun Zhang

G

Guangyu Wang

C

Cailong Liu

Laboratory of Quantum Materials Under Extreme Conditions in Shandong Province, School of Physics Science and Information Technology, Liaocheng University 1 , Liaocheng 252000,

W

Wen Li