The synergistic effect of neutron irradiation on the tensile properties of Fe-0.74 wt.% Ni alloy: A combined study of machine-learning and molecular dynamics

The aim of this study is to elucidate the mechanisms of irradiation damage to reactor pressure vessel (RPV) steel using a machine learning algorithms and high-throughput calculations. Various Fe-Ni alloy structures were generated based on structural enumeration for high-throughput first-principles calculations, with the Fe-Ni interatomic potential trained using a Gaussian approximation function. Simulations were conducted using LAMMPS software to investigate the effects of neutron irradiation on the tensile properties of Fe-0.74 wt.% Ni alloy, utilizing the well-established Fe-Ni interatomic potential. The irradiation dose significantly impacts defects in the Fe-Ni alloy. Synergistic effects of alloy solute element content and temperature with irradiation defects reveal that defect numbers at irradiation points increase linearly with MD-dpa and PKA energy. During irradiation, Ni atoms diffuse via the exchange with vacancy, synergizing with other Ni elements. Notably, Ni content inversely affects yield stress, resulting in lower yield stress in irradiated materials compared to pre-irradiation levels. While temperature inversely affects yield stress, its synergistic effect with defects increases yield stresses post-irradiation, known as irradiation hardening. Post-irradiation, the yield strain increases, and a flat plateau stress region is observed in Fe-Ni alloys. Ni atoms act as a buffer during the stretching process, contributing to a relatively gentle slope stress region despite increasing stress. The distribution of Ni atoms significantly influences the stress-strain curve, in which the aggregated Ni atoms decrease yield strength, whereas uniform distribution increases it, highlighting Ni atoms as buffering role during stretching process. These simulations yield valuable insights for exploring scalability and enhancing the development of irradiation hardening and embrittlement models.