Scholarlink

<p>Fast and efficient lithium-ion transport is crucial for high-performance rechargeable batteries but remains poorly understood in complex alloy anodes due to the coexistence of multiple competing diffusion mechanisms. A key challenge is disentangling how defect type, local coordination, and structural connectivity collectively govern migration barriers, which has so far limited the development of generalizable design rules. Here, we develop neuroevolution potentials (NEPs), a machine-learning framework trained to near-density functional theory (DFT) accuracy, providing both reliability and transferability for large-scale simulations of Li diffusion in Li-In and Li-Sn alloys. Using NEPs, we identify three primary factors controlling Li mobility: the dominant defect carrier (Li vacancy, Li interstitial, or metal vacancy), the connectivity of low-barrier diffusion pathways, and the local atomic environment, with migration distance and charge redistribution as the most critical determinants of diffusion barriers. Li diffusion shows strong structural inheritance across intermetallics, where the crystal backbone dictates long-range connectivity and chemical composition fine-tunes local behavior. The favorable diffusion structural frameworks revealed by NEPs can be extended to other Li-rich intermetallics, enabling predictive screening of fast-diffusing phases. This work establishes a predictive framework for designing next-generation alloy anodes with high ionic conductivity.</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

Structural insights and predictive screening of ion transport in Li-rich alloys via neuroevolution potentials

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年2月19日

查看全文
  • Di Jin,
  • Shicong Ding,
  • Hailong Qiu,
  • Yang Li,
  • Xinyang Li,
  • Guochun Yang

摘要

Fast and efficient lithium-ion transport is crucial for high-performance rechargeable batteries but remains poorly understood in complex alloy anodes due to the coexistence of multiple competing diffusion mechanisms. A key challenge is disentangling how defect type, local coordination, and structural connectivity collectively govern migration barriers, which has so far limited the development of generalizable design rules. Here, we develop neuroevolution potentials (NEPs), a machine-learning framework trained to near-density functional theory (DFT) accuracy, providing both reliability and transferability for large-scale simulations of Li diffusion in Li-In and Li-Sn alloys. Using NEPs, we identify three primary factors controlling Li mobility: the dominant defect carrier (Li vacancy, Li interstitial, or metal vacancy), the connectivity of low-barrier diffusion pathways, and the local atomic environment, with migration distance and charge redistribution as the most critical determinants of diffusion barriers. Li diffusion shows strong structural inheritance across intermetallics, where the crystal backbone dictates long-range connectivity and chemical composition fine-tunes local behavior. The favorable diffusion structural frameworks revealed by NEPs can be extended to other Li-rich intermetallics, enabling predictive screening of fast-diffusing phases. This work establishes a predictive framework for designing next-generation alloy anodes with high ionic conductivity.