Scholarlink

<p>Silicon on insulator technology requires precise control of buried oxide layers and the associated heat transfer across interfaces. Current approaches struggle to predict oxygen distribution and layer thickness after implantation and annealing, and they face challenges in computing interfacial thermal resistance under complex conditions. Here we show a computational framework that integrates machine-learned interatomic potentials with molecular dynamics to resolve oxygen diffusion kinetics and the evolution of layer thickness. The framework predicts depth-dependent oxygen redistribution during annealing and reproduces the measured dimensions of fabricated samples with high accuracy. It also enables reliable determination of interfacial thermal resistance, surpassing conventional molecular dynamics method in agreement with experimental benchmarks. By linking atomic scale oxygen migration with macroscopic heat transport, this work establishes quantitative relationships between structure, process, and properties. The framework provides guidance for optimizing implantation energy, dose, and annealing protocols, and represents a step toward rational design of high-performance silicon on insulator technologies.</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

Deep learning modeling of oxygen redistribution and thermal transport in silicon on insulator and buried oxide layers

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年3月18日

查看全文
  • Xiuwen Yan,
  • Mingzhu Liu,
  • Wenjin Cheng,
  • Yong Sun,
  • Yiwen Su,
  • Xinzhong Wang,
  • Zhuo Tang,
  • Libo Peng,
  • Guangping Zheng,
  • Jiashu Chen

摘要

Silicon on insulator technology requires precise control of buried oxide layers and the associated heat transfer across interfaces. Current approaches struggle to predict oxygen distribution and layer thickness after implantation and annealing, and they face challenges in computing interfacial thermal resistance under complex conditions. Here we show a computational framework that integrates machine-learned interatomic potentials with molecular dynamics to resolve oxygen diffusion kinetics and the evolution of layer thickness. The framework predicts depth-dependent oxygen redistribution during annealing and reproduces the measured dimensions of fabricated samples with high accuracy. It also enables reliable determination of interfacial thermal resistance, surpassing conventional molecular dynamics method in agreement with experimental benchmarks. By linking atomic scale oxygen migration with macroscopic heat transport, this work establishes quantitative relationships between structure, process, and properties. The framework provides guidance for optimizing implantation energy, dose, and annealing protocols, and represents a step toward rational design of high-performance silicon on insulator technologies.