Coupled Molecular Dynamics and Finite Element Methods for the simulation of interacting particles and fields

The dynamical simulation of many particle systems is currently a widespread technique in many fields: e.g. nuclear and atomic physics, computational material science, computational chemistry, molecular biology and pharmacology. Under the locution Molecular Dynamics (MD) we can regroup a variety of approaches and numerical codes, whereas the commonalities are: 1) the atomistic (or nuclear) resolution (i.e. particles are atoms or nucleons), 2) the force derivation, starting from the systems configuration, through semi-classical (also called semi-empirical) or quantum mechanics based theoretical frameworks, 3) the (generally explicit) numerical integration of the Newton-like equations of the motions to simulate the system kinetics. Within this scheme methodology variations can be found in the literature, but it is undoubtedly valid to qualify the MD meaning in the field of the scientific computation. The general scope of this Thesis work is the extension of the MD methods to the study of kinetics of larger particle (i.e. from mesoscopic dimensions and above), where effective particle-particle interactions are mediated by a field evolving self-consistently with the many particles system. This objective is mainly motivated by the applications of the method to control and predict the manipulation of mesoscopic (electrically) neutral particles by means of electromagnetic (e.m.) interactions: i.e. exploiting the so called dielectrophoresis (DEP) phenomena in the systems of electromechanical particles (EMPs). This is the specific case of study here considered, but in principle the methodology can applied after suitable adaptation to also other systems.