Dependence of radiation damage accumulation in iron on underlying models of displacement cascades and subsequent defect migration

Groups of displacement cascades calculated independently with different simulation models and computer codes are compared on a statistical basis. The parameters used for this comparison are the number of Frenkel pairs (FP) produced, the percentages of vacancies and self-interstitial atoms (SIAs) in clusters, the spatial extent and the aspect ratio of the vacancies and the SIAs formed in each cascade. One group of cascades was generated in the binary collision approximation (BCA) and all others by full molecular dynamics (MD). The MD results differ primarily due to the empirical interatomic potentials used and, to some extent, in code strategies. Cascades were generated in simulation boxes at different initial equilibrium temperatures. Only modest differences in the predicted numbers of FP are observed, but the other cascade parameters may differ by more than 100%. The consequences of these differences on long-term cluster growth in a radiation environment are examined by means of object kinetic Monte Carlo (OKMC) simulations. These were repeated with three different parameterizations of SIA and SIA cluster mobility. The differences encompassed low to high mobility, one- and three-dimensional migration of clusters, and complete immobility of large clusters. The OKMC evolution was followed until 0.1 dpa was reached. With the range of OKMC parameters used, cluster populations after 0.1 dpa differ by orders of magnitude. Using the groups of cascades from different sources induced no difference larger than a factor of 2 in the OKMC results. No correlation could be identified between the cascade parameters considered and the number densities of vacancies and SIAs predicted by OKMC to cluster in the long term. However, use of random point defect distributions instead of those obtained for displacement cascades as input for the OKMC modeling led to significantly different results.