Scholarlink

<p>The traditional von Neumann architecture faces challenges such as high energy consumption and latency stemming from the separation of memory and computation. In contrast, memristors, which integrate non-volatile storage with biological synaptic plasticity, show significant potential in neuromorphic devices and memory systems. This study utilizes MoO<sub>3</sub> thin films as the primary material, achieving excellent resistive switching characteristics through the regulation of electrode materials. The Au/MoO<sub>3</sub>/Au device demonstrated no drift in the high resistance state (HRS) and the low resistance state (LRS) after 10<sup>5</sup> pulses, with a retention time of up to 10<sup>4</sup>&#xa0;s. In comparison to Au electrodes, the Ag/MoO<sub>3</sub>/Ag device exhibited typical bipolar resistive switching behavior, with switching voltages of 1.03&#xa0;V and -0.47&#xa0;V, respectively, and enabled multilevel storage and biological synaptic functions. Importantly, the conduction mechanisms of the devices with different electrodes were elucidated using the trap-assisted tunneling (TAT) model and the space-charge-limited current (SCLC) model, respectively. Thus, MoO<sub>3</sub> memristors with varying electrodes provide essential device support for low-power neuromorphic computing and high-density storage.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Modulation of synaptic behavior and performance in MoO3 memristors by electrode engineering

  • 期刊论文

  • 发表日期：

    2026年4月14日

查看全文
  • Yan Liu,
  • Feng Liu,
  • Ronghong Miao,
  • Wenkang Zhang,
  • Bo Cheng

摘要

The traditional von Neumann architecture faces challenges such as high energy consumption and latency stemming from the separation of memory and computation. In contrast, memristors, which integrate non-volatile storage with biological synaptic plasticity, show significant potential in neuromorphic devices and memory systems. This study utilizes MoO3 thin films as the primary material, achieving excellent resistive switching characteristics through the regulation of electrode materials. The Au/MoO3/Au device demonstrated no drift in the high resistance state (HRS) and the low resistance state (LRS) after 105 pulses, with a retention time of up to 104 s. In comparison to Au electrodes, the Ag/MoO3/Ag device exhibited typical bipolar resistive switching behavior, with switching voltages of 1.03 V and -0.47 V, respectively, and enabled multilevel storage and biological synaptic functions. Importantly, the conduction mechanisms of the devices with different electrodes were elucidated using the trap-assisted tunneling (TAT) model and the space-charge-limited current (SCLC) model, respectively. Thus, MoO3 memristors with varying electrodes provide essential device support for low-power neuromorphic computing and high-density storage.