Scholarlink

<p>Predicting the intrinsic corrosion tendency of biodegradable Zn-based alloys at the atomic scale remains a major challenge. In this work, a deep learning interatomic potential (Deep Potential, DP) with near first-principles accuracy is developed to investigate the mechanical properties and electrochemical corrosion mechanisms of Zn–Fe alloys. The DP model, trained via active learning, achieves root-mean-square errors of 2.99 × 10⁻<sup>3</sup>&#xa0;eV/atom for energy and 2.24 × 10⁻<sup>3</sup>&#xa0;eV/Å for forces. It accurately reproduces lattice constants, phonon dispersion, and elastic properties, demonstrating DFT-level predictive capability. By integrating DP-derived energetic descriptors with an enhanced Butler–Volmer framework, the intrinsic polarization behavior of Zn–Fe alloy surfaces is quantitatively predicted. Surface energy density is identified as the key descriptor governing corrosion susceptibility. Close-packed surfaces of Zn₉Fe₄ exhibit enhanced stability due to their lower surface energy density, while ZnFe₃ shows the lowest corrosion current density (log[icorr] ≈ − 7.13 A/cm<sup>2</sup>). These results indicate that Fe incorporation suppresses anodic dissolution by strengthening atomic bonding. This multiscale framework bridges atomistic energetics and macroscopic electrochemical response, providing a reliable theoretical basis for the design of biodegradable Zn–Fe alloys.</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 potential enabled atomistic investigation of the mechanical properties and electrochemical corrosion mechanisms of biodegradable Zn–Fe alloys

  • 期刊论文

  • 发表日期：

    2026年7月10日

查看全文
  • Xiangyang Zhang,
  • Zhijie He,
  • Keyuan Chen,
  • Dehong Lu,
  • Xiaohua Yu

摘要

Predicting the intrinsic corrosion tendency of biodegradable Zn-based alloys at the atomic scale remains a major challenge. In this work, a deep learning interatomic potential (Deep Potential, DP) with near first-principles accuracy is developed to investigate the mechanical properties and electrochemical corrosion mechanisms of Zn–Fe alloys. The DP model, trained via active learning, achieves root-mean-square errors of 2.99 × 10⁻3 eV/atom for energy and 2.24 × 10⁻3 eV/Å for forces. It accurately reproduces lattice constants, phonon dispersion, and elastic properties, demonstrating DFT-level predictive capability. By integrating DP-derived energetic descriptors with an enhanced Butler–Volmer framework, the intrinsic polarization behavior of Zn–Fe alloy surfaces is quantitatively predicted. Surface energy density is identified as the key descriptor governing corrosion susceptibility. Close-packed surfaces of Zn₉Fe₄ exhibit enhanced stability due to their lower surface energy density, while ZnFe₃ shows the lowest corrosion current density (log[icorr] ≈ − 7.13 A/cm2). These results indicate that Fe incorporation suppresses anodic dissolution by strengthening atomic bonding. This multiscale framework bridges atomistic energetics and macroscopic electrochemical response, providing a reliable theoretical basis for the design of biodegradable Zn–Fe alloys.