The simulation of microstructural evolution of materials at temperatures appropriate for fusion operating conditions requires a critical re-examination of fundamental concepts underpinning linear elasticity theory. We find that linear elasticity theory predicts straight dislocations to be unstable with respect to thermal perturbations, which is in striking disagreement with molecular dynamics simulations. Consequently, a dislocation dynamics model based solely on linear elasticity does not provide a sound foundation for the treatment of stochastic thermal fluctuations of dislocation lines.
We propose an atomistically motivated extension of linear elasticity theory that takes into account the physical properties of the edge dislocation core. The resulting model predicts perturbation formation energies qualitatively and quantitatively consistent with predictions derived from molecular dynamics, without need of introducing any empirical parameters. Furthermore, the model offers a first-principles derivation for the commonly used line-tension approximation for the dislocation core energy, and thus represents a fundamental step towards unifying the atomistic and continuum scales in modelling defects and dislocations.