Scholarlink

<p>This study investigates memristive and synaptic properties of parylene (PPX)-based memristors in crossbar geometry for brain-inspired neuromorphic computing applications. We demonstrate Cu/PPX/Au memristors exhibiting stable resistive switching (<i>R</i><sub>on</sub>/<i>R</i><sub>off</sub> &gt; 7 × 10<sup>5</sup>), multilevel operation (16 resistive states), and synaptic plasticity emulation capabilities. The devices successfully implement biosimilar learning rules, demonstrating spike-timing-dependent plasticity (STDP) with up to 1400% conductance modulation for potentiation, along with long-term potentiation/depression (LTP/LTD) featuring precise analog tuning. The developed dual-channel measurement approach effectively minimizes sneak currents in passive crossbar arrays. These results confirm the promising potential of PPX memristors for energy-efficient neuromorphic systems, with prospective applications in spiking neural networks.</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

Parylene-Based Memristors in Crossbar Architecture: From Stable Resistive Switching to Synaptic Plasticity

  • 期刊论文

  • 发表日期：

    2025年6月11日

查看全文
  • Georgii A. Yuklyaevskikh,
  • Andrey V. Emelyanov,
  • Aleksandr I. Iliasov,
  • Yulia V. Grishchenko,
  • Aleksandr A. Nesmelov,
  • Timofey D. Patsaev,
  • Boris S. Shvetsov

摘要

This study investigates memristive and synaptic properties of parylene (PPX)-based memristors in crossbar geometry for brain-inspired neuromorphic computing applications. We demonstrate Cu/PPX/Au memristors exhibiting stable resistive switching (Ron/Roff > 7 × 105), multilevel operation (16 resistive states), and synaptic plasticity emulation capabilities. The devices successfully implement biosimilar learning rules, demonstrating spike-timing-dependent plasticity (STDP) with up to 1400% conductance modulation for potentiation, along with long-term potentiation/depression (LTP/LTD) featuring precise analog tuning. The developed dual-channel measurement approach effectively minimizes sneak currents in passive crossbar arrays. These results confirm the promising potential of PPX memristors for energy-efficient neuromorphic systems, with prospective applications in spiking neural networks.