NuFuel & MMSNF 2015

First Workshop on Research into Nuclear Fuel in Europe
and Materials Modeling and Simulation for Nuclear Fuels Workshop
Karlsruhe, Germany, November 16th to 18th, 2015

Updated: Tue 08 Dec 2015, 14:27

Talk 5.8: Calculation of diffusion coefficients in ceramic oxide fuels using adaptive kinetic Monte Carlo modelling

Daniel Griffin, Mark Bankhead
  • National Nuclear Laboratory (NNL), Warrington, UK


Fuel performance modelling codes are reliant upon knowledge of material properties of the fuels, such as thermal conductivity or fission product diffusion coefficients. Diffusion coefficients can be difficult to measure experimentally and therefore it is often desirable to use modelling techniques, such as molecular dynamics, to support, or in place of, experimental measurements. However, the timescales and computational effort required to reliably calculate diffusion coefficients, particularly for cationic and fission product species, can be prohibitive. This work investigates the use of adaptive kinetic Monte Carlo (AKMC) modelling [1] as an alternative to classical molecular dynamics (MD) for the calculation of diffusion coefficients in ceramic oxide fuels. Whereas classical MD simulates molecular motion in ‘real time’, thus limiting practical simulations to timescales in the region of 10-9–10-6 s, AKMC is able to target the ‘rare-event’ atom translations associated with species diffusion and hence enables access simulations upwards of 10-3 s in length. In theory, the longer timescales accessible with AKMC enable statistically reliable calculation of diffusion coefficients for species present in irradiated nuclear fuels that have previously been challenging to obtain via atomistic modelling approaches.

Since oxygen diffusion in oxide fuels is understood to be several orders of magnitude faster than most other cationic or fission product species present, and is therefore readily accessible through classical MD, a comparison of oxygen diffusion coefficients in UO2 as calculated using AKMC [2] and MD is presented to demonstrate this novel application of the AKMC modelling approach.

  1. G. Henkelman and H. Jónsson, Long time scale kinetic Monte Carlo simulations without lattice approximation and predefined event table, J. Chem. Phys., 2001, 115, 9657
  2. D. Gunn, N. Allan and J. Purton, Adaptive kinetic Monte Carlo simulation of solid oxide fuel cell components, J. Mater. Chem. A, 2014, 2, 13407