Simulating defects in nuclear graphite

Graphite is vital to the current operation and development of new nuclear reactors due to its use as a neutron moderator and is a key structural component of the core. Despite this, graphite is not well understood on an atomic level as a result of the complex microstructure which influences graphite properties. Computer simulations are limited to small system sizes and short timescales using current modelling methods. Much of the simulated data available does not align with experimental results. Improving the knowledge of graphitic behaviour under irradiation is crucial to extending gas-cooled nuclear reactor life span in the UK. Molecular Dynamics (MD) is used to model graphite in this work. An important component of MD simulations is the choice of potential energy model used. Here, a recent ?fitting of the ReaxFF potential is used, while other potentials such as the AIREBO potential are also considered.

The ReaxFF Potential has been employed to model a variety of radiation induced process in graphite. Collision cascades were simulated over a range of energies for statistical analysis in both Bernal and rhombohedral graphites. Generally, the defects produced are point defects, formed through a fractal-like branching structure. These results are in agreement with the Kinchin-Pease model. Analysis of the 22.7 degree and 20.4 degree channels in rhombohedral and Bernal graphite types have shown that there is a significant distance difference in the channelling length due to the size difference in the maximum channelling cross section. X-ray diffraction patterns have been simulated with heating shown to shift signature 00l peaks to lower angles as a direct result of increasing interlayer lattice spacing. The introduction of vacancies and interstitials to otherwise pristine lattice have been observed to produce measurable differences to simulated XRD data. Interstitials cause all peaks to shift to lower angles, due to induced strain in both inter- and intra-layer directions, while vacancies experience less peak broadening as reflection planes are more readily maintained. 

Over extended timescales, the rate of diffusion for point defects has been considered, with migration values for monovacancies being predicted as significantly lower than the experimentally reported ones in literature, having been simulated through alternative methods. The same is true for a single interstitial, though experimental results align better in this case. We ascribe this difference to similar works to neglecting two factors; that defects migrate through a variety of structures, which are not necessarily the most stable, reducing the energies associated with the transition pathway, and through thermally induced curvature of the lattice which is able to lower energy barriers. The ReaxFF, hNN and AIREBO potentials have been considered to model the defects of strain on vacancy-type structures, with the hNN and AIREBO potentials being discarded for this purpose as they produce substantially lower relative energies for each defect than those reported in literature. 

Using the ReaxFF potential, the application of strain has been seen to alter the preferential state. We observe that the introduction of a vacancy-type defect into an otherwise pristine lattice deforms the lattice, inducing strain ?fields that can extend up to 25 Ang from the monovacancy defect and 40 Ang from the divacancy defect. These strain ?fields have sufficient intensity to induce a change in the structure, should another vacancy type defect be inserted nearby. 

Dienes Stone Wales defects have been introduced to the surface of carbon nanotubes of varying radii as part of a preliminary study in order to determine the defects of curvature on defect energy barriers. Through Nudged Elastic Band (NEB) calculations, we have observed an increase in the stability of the defect structure with increasing curvature, with the energy barrier to the defective state reducing by 16.4%. It becomes easier for the defect to form at high curvature and more difficult for the defect to transition back to pristine graphite.