Molecular dynamics simulations of the defect evolution in tungsten on successive collision cascades

TL;DR

This study uses molecular dynamics simulations with different inter-atomic potentials to analyze defect evolution in tungsten under successive collision cascades, comparing results with experimental data.

Contribution

It introduces a detailed defect analysis in tungsten using both EAM and machine learning potentials, highlighting differences in defect predictions and agreement with experiments.

Findings

SNAP potential aligns better with experimental swelling data

SNAP predicts more small sessile defects affecting microstructure

Defect evolution varies with PKA energy and inter-atomic potential

Abstract

Molecular dynamics (MD) simulations of successive collision cascades within the same simulation domain were performed using two different inter-atomic potentials (IAP) in tungsten, one EAM based and the other a `quantum accurate' machine learning potential, SNAP. The micro-structural changes are analyzed as a function of displacements per atom (dpa) for primary knock-on atom (PKA) energies of 20 keV and 50 keV, reaching up-to irradiation dose of 0.1 and 0.2 dpa, respectively. Five sample simulations are carried out for each case for observing stochastic differences in the evolution of damage. A detailed defect analysis is carried out to observe changes in different parameters such as the number of surface defects, defect density, defect morphology and size distribution etc., as a function of dpa. We explore the properties that are sensitive to the IAP used and those that are sensitive to the PKA energy and note their similarities with experimental results at various dpa values. The SNAP potential shows better agreement with the experiments for swelling and number of surface defects. However, it also predicts presence of high number of small sessile defects which may have definite affect on the processes of microstructural evolution and material properties.