Bismuth-Induced Modulation of Vacancy Migration Enables Linear Synaptic Plasticity in 2D Perovskite Memristors

Abstract

Memristors based on 2D organic–inorganic halide perovskites are attractive for neuromorphic computing because bulky organic cations help regulate switching by constraining ion migration, enabling gradual and analog conductance modulation. However, uncontrolled ion transport can cause abrupt switching and poor conductance linearity that are detrimental to neuromorphic performance. Here, we report Bi-incorporated 2D PEA2PbI4 (Bi:PEA2PbI4) memristors designed to enhance potentiation/depression (P/D) linearity and neuromorphic synaptic characteristics. Electrical measurements show more regulated vacancy migration in Bi:PEA2PbI4 compared to pristine 2D perovskite memristors. DFT-based nudged elastic band (NEB) calculations further indicate that Bi incorporation increases iodide migration barriers by 0.18–0.19 eV along the vacancy pathways, supporting more linear P/D behavior. In addition, the memristors exhibit reliable long-term synaptic functions under diverse spike conditions. Consequently, in artificial neural network simulations using the modified National Institute of Standards and Technology (MNIST) datasets, the Bi:PEA2PbI4 devices achieve recognition accuracies of 95.55% on Large MNIST and 80.27% on Fashion MNIST, demonstrating Bi incorporation as a practical strategy for high-performance neuromorphic memristors.