Designing metal halide perovskite nanoparticle-based resistive random-access memory devices through chemical treatments

Abstract

Metal-halide perovskite (CsPbX3; X = Cl, Br, and I)-based resistive random-access memory (ReRAM) devices are emerging as promising candidates for neuromorphic systems owing to their low operation voltage and distinct ‘true’ and ‘false’ signals. However, conventional ReRAM devices exploit bulk or a few hundred micrometer-sized perovskites, which are unsuitable for achieving brain-inspired memory devices. This is because nanometer-sized perovskites exhibit a high density of defects on their surfaces, thus resulting in poor or negligible resistive switching characteristics. Herein, defects are engineered on (zero dimensional; 0D) CsPbX3 perovskite nanoparticle (PNP) surfaces, their chemical compositions are tuned by water treatment, and the role of the defects in achieving ReRAM devices is investigated. Theoretical estimations (density functional theory and molecular dynamics) and empirical analyses reveal that H2O molecules reduce the defect density and generate clean surfaces. Through the water- and water-halide treatments on CsPbBr3 PNPs, we successfully demonstrated 0D (18.7±2.4 nm) perovskite-based ReRAM devices and achieve a high ION/IOFF ratio (>105).