Self-Powered Halide Perovskite Optoelectronic Synaptic Memristors for Reconfigurable Logic and Reservoir Computing Applications.

Cui, Dongsheng; Guo, Pusheng; Xu, Yumeng; Gao, Xiangxiang; Guo, Xing; Wei, Wei; Lin, Zhenhua; Zhang, Jincheng et al. · Nano Lett · 2025

basic_science · Level V

Where this comes from

Abstract

The development of low-power neuromorphic systems requires the integration of sensing, logic operations, neuromorphic computing, and energy autonomy. Herein, we present a triple-cation and triple-anion perovskite-based (p-type/intrinsic/n-type) p-i-n optoelectronic memristor array that synergistically combines these functions. The device's inherent photovoltaic effect (∼0.8 V) enables self-powered optical synaptic plasticity at 520 nm, eliminating external biasing for near-zero power consumption. By coupling this intrinsic photovoltaic bias with tunable external voltages, we demonstrate four reconfigurable Boolean logic operations (NOT, XOR, NAND, IMPLY). Furthermore, a reservoir computing (RC) system for neuromorphic pattern recognition is implemented by leveraging the plasticity of the perovskite memristor, achieving classification accuracies of 97.94% for 1-bit and 90.73% for 4-bit handwritten digit recognition. The self-powered memristor integrating optical synapses, digital logic, and neuromorphic functionalities provide new paradigms for developing next-generation low-power, high-density integrated circuits with hybrid digital-analog architectures.