A-Site Cation Functional Engineering Enables Lead-free Perovskite Photosynapse for Neuromorphic Visual Computing

View at publisher
Authors: Xinyu Ma, Zhihao Lei, Beining Dong, Tingting Mei, Tianxu Huang, Tianqi Huang, Zijian Feng, Xinwei Guan, Fandi Chen, Tianruo Guo, Mohit Naresh Shivdasani, Long Hu
+4 more Jiabao Yi, Shimul Kanti Nath, Chun‐Ho Lin, Dewei Chu
Cited by: 1
References: 54

Abstract

Optoelectronic synapses that directly couple light sensing with memory and computing functions offer a promising route toward energy-efficient neuromorphic vision systems. However, the development of high-performance perovskite synapses remains heavily reliant on lead-based perovskites, raising sustainability and toxicity concerns. Herein, we report a previously unexplored lead-free antimony halide perovskite, Cs 2 AgSb 2 I 9, and demonstrate its application as an ultralow-power optoelectronic photosynapse, enabled by A-site cation functional engineering. Specifically, incorporation of Ag into the perovskite lattice fundamentally modulates the electronic structure, suppressing excitonic confinement, enhancing carrier transport, and introducing energetically favorable Ag interstitial states that act as reversible charge trapping/detrapping centers to enable short-term memory. As a result, Cs 2 AgSb 2 I 9 devices exhibit extensive synaptic behaviors, including paired-pulse facilitation and spike-intensity-, spike-duration-, and spike-number-dependent plasticity, as well as Hebbian-like learning characteristics. Notably, the photosynapse can operate with an ultralow energy consumption of 6.18 × 10 –14 J per synaptic event, approaching the energy scale of biological synapses. Artificial neural networks based on Cs 2 AgSb 2 I 9 devices further demonstrate a high accuracy of 97.5% for image recognition. Overall, this work introduces a new lead-free perovskite platform and highlights A-site cation functional engineering as an effective strategy for realizing sustainable and low-power neuromorphic visual computing.

Fields of Study & Subjects

Perovskite Materials and Applications Advanced Memory and Neural Computing Neural Networks and Reservoir Computing

Publication details

Journal / source
ACS Applied Materials & Interfaces
Publisher
American Chemical Society (ACS)
Volume
18
Issue
16
Pages
23399–23409
Publication date
14 April 2026
Language
EN