EUVIMEDEUVIMED entdecken
Crossref

Primary radiation damage and defect clustering in molybdenum: a high-fidelity benchmark study based on the NEP-ZBL machine learning potential

Guocai Lv, Qingyuan Kong, Hao Zhang, Ping Qian

Modelling and Simulation in Materials Science and Engineering · 15.09.2026 · 34 (6) · 065002

DOI: 10.1088/1361-651x/ae8563 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Modelling and Simulation in Materials Science and Engineering, online erschienen am 04.08.2026. Verfasst von Guocai Lv und 3 weiteren Autor:innen. Die Arbeit nennt 37 Literaturangaben. Gefördert von: Ministry of Science and Technology, National Natural Science Foundation of China.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Large-scale machine-learning molecular dynamics simulations are employed to establish a high-fidelity benchmark for primary irradiation damage in molybdenum. Interatomic interactions are described by a neuroevolution potential augmented with Ziegler–Biersack–Littmark repulsion, which captures point-defect energetics and surface properties with near-density-functional theory accuracy. Utilizing an expanded statistical ensemble of 20 independent cascade simulations per data point to ensure convergence, we systematically quantify Frenkel pair (FP) generation and defect clustering at 100 K, 300 K, and 1000 K across a primary knock-on atom energy range of 1–50 keV. Our results demonstrate that surviving FP counts follow a sublinear power-law dependence on energy but decrease at elevated temperatures due to enhanced recombination kinetics. Crucially, we reveal a distinct morphological crossover: vacancies predominantly exist as isolated monomers, whereas self-interstitials exhibit a markedly higher clustering propensity, with their population exceeding that of vacancy clusters for all configurations of size ⩾ 2 . This mechanism leads to significantly higher interstitial-cluster fractions (∼0.6) compared to vacancy clusters (∼0.2), resolving quantitative discrepancies in prior empirical models. Beyond confirming several qualitative trends known in body-centered-cubic metals, this work provides a Mo-specific, quantitatively benchmarked primary-damage database with statistically robust defect survival, cluster fractions, and cluster-size distributions. These data help resolve discrepancies among earlier Mo simulations based on empirical potentials and provide physically grounded inputs for kinetic Monte Carlo and cluster-dynamics modeling of radiation effects in Mo.

Förderung

  • Ministry of Science and Technology
  • National Natural Science Foundation of China

Zitieren

APA

Lv, G., Kong, Q., Zhang, H., & Qian, P. (2026). Primary radiation damage and defect clustering in molybdenum: a high-fidelity benchmark study based on the NEP-ZBL machine learning potential. Modelling and Simulation in Materials Science and Engineering, 34(6), 065002. https://doi.org/10.1088/1361-651x/ae8563

Vancouver

Lv G, Kong Q, Zhang H, Qian P. Primary radiation damage and defect clustering in molybdenum: a high-fidelity benchmark study based on the NEP-ZBL machine learning potential. Modelling and Simulation in Materials Science and Engineering. 2026;34(6):065002. doi:10.1088/1361-651x/ae8563

BibTeX
@article{Lv2026,
  author = {Lv, Guocai and Kong, Qingyuan and Zhang, Hao and Qian, Ping},
  title = {Primary radiation damage and defect clustering in molybdenum: a high-fidelity benchmark study based on the NEP-ZBL machine learning potential},
  journal = {Modelling and Simulation in Materials Science and Engineering},
  year = {2026},
  volume = {34},
  number = {6},
  pages = {065002},
  doi = {10.1088/1361-651x/ae8563}
}
Kostenlos starten →