Théorique, numérique
The research work is in the context of a fundamental theoretical challenge: mastering the projection of the high-dimensional phase-space dynamics of a system of >1000 atoms on 1 order parameter, a coarse-graining procedure that leads to a mathematical description in terms of Langevin equations (stochastic differential equations). Such projection is a key tool to study phase transitions, but also other activated processes in biophysics (e.g., protein folding or protein-drug interaction), chemistry (reactions in solution), nanoscience, and so on. This tool has many advantages, since it yields a physically intuitive (no black box) and parsimonious model of a complex phenomenon, while keeping rigorous mathematical foundations and preserving the real free-energy landscape and kinetic rates. The internship builds upon an intense research activity in the host group over the last years (see https://scholar.google.fr/citations?user=0fu5X-cAAAAJ&hl=en). The goal will be to advance the theoretical understanding and the exploitation of Langevin models as applied to a fundamental phase-transition mechanism, the nucleation of a critical solid cluster (and eventually of a solid phase) from a liquid. Notwithstanding the severe failure of simplified descriptions in terms of classical nucleation nucleation theory, the candidate will contribute to solve the open problem of faithfully modelling the nucleation mechanism of a realistic, complex system.