Most of the physical mechanisms which control material evolution can be described with kinetic based approaches at the atomic scale: phase transformations, diffusion of vacancies, deformation, dislocation nucleation and propagation, microcavities, microcracks. Classical methods like molecular dynamics are however limited by high frequencies phonons. Currently, the time scale reached by these methods is of the order of several nanoseconds. Different approaches have been investigated to go beyond this limit (Diffusive Molecular Dynamics, ART Monte Carlo, Phase Field Crystal) but neither emerged as being clearly the most appropriated. In that context, we would like to develop a new approach, still at the atomic scale, but based on a Langevin dynamics, which is of first order in time. This is equivalent to remove kinetic energy. The main interest is to keep a discrete description of matter and a continuous description of atomic positions without having to follow fast oscillations caused by phonons. Accessible time scales should be therefore several orders of magnitude larger than those reached by classical methods. A noise term must be however introduced in the kinetics, and has to be controlled to guaranty the convergence towards to the thermodynamical equilibrium. First results tend to show that this method is promising to study materials subjected to different thermal and mechanical loadings [C. Baruffi et al. , Mater. Theory 3, 4 (2019)].
This thesis is based on theoretical, numerical and practical phases. In the theoretical phase, a technique of scale change will be applied to derive Langevin equations from Newton equations. In this way, the expression of the Langevin equations, in particular the noise term will be deduced as well as a physically justified interatomic effective potential. To do this, a procedure inspired by R. Zwanzig (J. Stat. Phys. 9, 215 (1973)) will be used. In the numerical phase, interatomic effective potentials often used for investigations of alloys (Lennard-Jones, Embedded Atom Model (EAM) or Second moment approximation of the tight binding model (TB-SMA)) will by determined with the help of molecular dynamics simulations. In the practical phase, microstructural evolutions during thermal or mechanical loadings (phases, variants, grains, dislocations) will be reproduced in alloys for aerospace applications
job : PhD (3 years)
Location : LEM, ONERA, Châtillon, France
Knowledge in solid state physics and/or material science, Capability in programing under unix/linux environment with scientific languages
Academic degree : Master Degree
Contacts : Mathieu Fèvre, Alphonse Finel (Email Us)