High‐Performance GdNiO<sub>3</sub> Epitaxial Film Memristor for Neuralactivity Analysis
Bing Bai, Gongjie Liu, Yong Sun, Pan Liu, Zhen Zhao, Zhenqiang Guo, Xiaobing Yan
Abstract
Abstract Perovskite‐type rare earth nickelates based memristor have recently attracted extensive attention in the field of novel storage computing due to their special electronic structure and exotic physical properties. However, there is still a shortage of memristors with ultra‐high stability performance, which will provide a solid foundation for future neural network computing with high accuracy recognition rates. Here, a GdNiO 3 ‐based interfacial memristor is presented, which possesses ultra‐high stable performance, such as electroforming‐free, low device‐to‐device variation, reliable cyclic switching, high on/off ratio (≈10 4 ) and stable pulse modulation of conduction. Combined with the comprehensive microstructure results, this behavior is ascribed to the interface Schottky barrier variation caused by the 1D oxygen vacancy channel conduction according to the transmission electron microscopy results. In particular, based on the device's stable pulse modulation plasticity performance, the study also succeeds in achieving highly accurate neural firing pattern recognition up to ≈99.75% accuracy and monitoring of pattern transitions by implementing a reservoir computing system based on the device. This research advances the progress of nickelates in novel storage computing and paves the way for future efficient memristor‐based reservoir computing systems to handle more complex temporal tasks.