An Ultra‐Robust Memristor Based on Vertically Aligned Nanocomposite with Highly Defective Vertical Channels for Neuromorphic Computing

作者：Zedong Hu, Hongyi Dou, Jeremy B. Gan, Cun Li, Xuanyu Sheng, Juanjuan Lu, Abhijeet Choudhury, Jialong Huang, Yizhi Zhang, Zhongxia Shang, Ye Cao, X. Zhang, Haiyan Wang · 发表于：Advanced Functional Materials · 年份：2025 · DOI：10.1002/adfm.202512719 · 被引用次数：3 · 研究领域：Advanced Memory and Neural Computing、Photoreceptor and optogenetics research、Neuroscience and Neural Engineering

Abstract Oxide‐based memristors are considered promising candidates for next‐generation memory and neuromorphic computing applications owing to their intrinsic resistive switching properties, non‐volatile storage, excellent device plasticity, and compatibility with current complementary metal‐oxide‐semiconductor (CMOS) technology. For the physical implementation of neuromorphic microchips, memristors with high power efficiency, reasonable on/off ratios, excellent device reliability, and enhanced plasticity are essential. In this work, a novel memristor design based on SrTiO 3 ‐CeO 2 (S‐C) vertically aligned nanocomposite (VAN) is proposed, utilizing the highly defective vertical interfaces as migration channels for oxygen vacancies (). More interestingly the interface density ( σ int ) and strain can be effectively tuned by the deposition parameters. Comprehensive analyses, including microstructural analysis, electrical characterization, and finite element modeling, reveal that S‐C VAN memristors with high interface density (HID) exhibit superior power efficiency and reduced device‐to‐device (D2D) variation. Additionally, the S‐C HID memristor demonstrates endurance up to 10 12 cycles and retains its resistance states for up to 10 5 s at room temperature. The improved plasticity of the S‐C HID memristor enables better linear modulation of conductance with minimal fluctuation over long cycles and across different devices. Furthermore, artificial neural networks (ANNs) incorpor...