Multi-scale Modeling of Radiation Damage: Large Scale Data Analysis

Abstract

Abstract. Modification of materials in nuclear reactors due to neutron irradiation is a multiscale problem. These neutrons pass through materials creating several energetic primary knock-on atoms (PKA) which cause localized collision cascades creating damage tracks, defects (interstitials and vacancies) and defect clusters depending on the energy of the PKA. These defects diffuse and recombine throughout the whole duration of operation of the reactor, thereby changing the micro-structure of the material and its properties. It is therefore desirable to develop predictive computational tools to simulate the micro-structural changes of irradiated materials. In this paper we describe how statistical averages of the collision cascades from thousands of MD simulations are used to provide inputs to Kinetic Monte Carlo (KMC) simulations which can handle larger sizes, more defects and longer time durations. Use of unsupervised learning and graph optimization in handling and analyzing large scale MD data will be highlighted.

Inclusion and validation of electronic stopping in the open source LAMMPS code

“…The number of defects as a function of energy for both Fe and W are calculated using the methods described in keV, with and without electronic stopping. The number of defects from the arcDPA formula for Fe is also plotted for comparison [18]. The total ES losses from the LAMMPS simulations are compared with the total ES losses obtained from the NRT model to validate the fix_esfriction code added to LAMMPS.…”

“…LAMMPS [1] is an open source software capable of massively parallel atomistic simulations. We use it widely in our radiation damage studies, not only to simulate colision cascades [17,18], but also to simulate defect diffusion [19,20]. It has several package extensions amongst which the two temperature model (TTM) by Duffy et.al., is also implemented.…”

Kinetic Monte Carlo Simulations of Irradiation Effects

“…IV-3 Typical damage accumulation / microstructure evolution: impact of dose, temperature, impurities … Many KMC studies has also been dedicated to characterize damage accumulation in various materials to model specific experiments such as ion implantation in Si for instance (Hobler and Otto 2003) or to investigate the impact of various parameters: the irradiation technique (ions versus neutrons , ions versus protons (Fluss et al 2004), continuous versus pulsed (Perlado et al 2003)), the neutron spectrum (Soneda et al 2003) (Choi and Joo 2013) (Figure 8), the dose, and dose-rate (Soneda et al 2003) (Chiapetto et al 2016a), the temperature (Soneda et al 2003) (Arévalo et al 2007) (Chiapetto et al 2016a) (Castin et al 2018), impurities or solute atoms (Alonso et al 2000) (Gámez et al 2007) (Castin et al 2018), the effect of annealing (Caturla et al 2000b) (Fluss et al 2004) or to compare materials (Caturla et al 2000a). The irradiation is typically modeled using cascade databases obtained either through MD (Perlado et al 1998) (Caturla et al 2000a) (Caturla et al 2000b) (Wirth et al 2001) (Marian et al 2001) (Arévalo et al 2007) (Chiapetto et al 2016a) (Warrier et al 2016) (Castin et al 2018) or the BCA method (Hobler and Otto 2003) (Otto et al 2005) (Hobler et al 2005) (Becquart et al 2006) (Valles et al 2015) (Payet et al 2016). The effect of pulsed irradiation has been studied in…”