PDRA: Oxy-carbide fuel performance, comparison to oxide and carbide

Imperial College London

First advertised here: 14 June 2024


Carbide nuclear fuels have been developed since the 1960s and were originally proposed as an alternative fuel to oxide for fast neutron spectrum reactors. More recently, oxy-carbides have been considered for several non-water cooled and advanced Generation IV reactors. In comparison to conventional oxide fuels their higher thermal conductivity, higher fissile element density, and radiation tolerance make carbides and possibly oxy-carbides strong alternate fuel candidates. The beneficial properties of pristine UC compared to conventional oxide fuels could allow the fuel to operate at higher burn-ups, although, it is unclear how defect formation alters properties. Both stoichiometry changes and formation of Pu occur because of fuel burn-up.


Atomic scale simulations are used to predict how specific defect configurations influence properties of (U/Pu)C and selected oxy-carbides (specified by others carrying out experiments). Simulations will identify the temperature dependence of thermophysical properties of stoichiometric and non-stoichiometric compositions. Key temperature dependent properties are: lattice volume (and hence the linear thermal expansion coefficient), specific heat and thermal conductivity.  Modern fuel performance codes employ knowledge of the physical processes that underpin property behaviour to inform constituent parts of the performance codes.  This approach can be particularly useful where limited data is available on which to base empirically fitted fuel behaviour property models (which is the case for uranium and plutonium carbide and oxy-carbides).


Empirical potential based calculations are used to carry out molecular dynamics simulations of temperature dependence (employing the Large-scale Atomic/Molecular Massively Parallel Simulator (LAMMPS) simulation package). We compare these predictions to available experimental data for UC and address ambiguities between the high temperature experimental measurements. New interatomic potentials will be derived to deal with oxy-carbides as well as plutonium containing compounds (developed by fitting to the experimental PuC lattice parameter data).  This provides the capability to predict thermophysical properties for PuC and mixed (U,Pu)C (10% and 20% additions of PuC to UC).

For the electronic component of the specific heat capacity and thermal conductivity, electronic band structures will be calculated using density functional theory (DFT) to solve the Boltzmann transport equation of both UC and PuC (employing the Vienna Ab-Initio Simulation Package, VASP). DFT calculations are also used to examine the defect energies of UC, comparing to other DFT and MD studies. Non-stoichiometric simulation cells will be generated by randomly assigning defect species to interstitial sites or by removing lattice atoms, until the desired defect concentration is reached. Similarly, mixed nitride systems are created by randomly assigning Pu atoms to U sites.


Interim report/update due Sep 24

Final project report due Mar 25.


All computing will be carried out using the Imperial College High Performance cluster.  All codes used are free to use for UK based academics for research purposes.