Atomistic Contribution To The Understanding Of Metallic And Silica Glasses PDF Download

Amorphous materials are ubiquitous in everyday life. They comprise "hard" and"soft" glasses. Hard amorphous systems are usually seen as structure materials, with properties and use comparable to those of crystalline solids. Soft glasses are usually seen as complex fluids, described in terms of their rheological properties with the corresponding practical applications (concrete, paints, drilling mud, cosmetic gels, creams or foams, etc). Amorphous materials can either present a solid-like behaviour or flow depending on their mechanical load: all are yield-stress fluids. Their usage limits are often defined by the occurrence of shear-banding, an extreme form of localization seen in molecular glasses as well as in granular materials. There is now considerable evidence that they are consequences of the existence of a disordered structure at the level of the elementary constituents (atoms, particles,...). Studies of plasticity in amorphous solids, are still hampered by the lack of any identifiable defect responsible for the plastic response. It is now acknowledged that plasticity is the net result of local rearrangements, or "shear transformations", involving small clusters of (say a few tens of) particles. These rearrangements are thermally--activated and are ubiquitous processes in the structural relaxation and deformation of glasses at low temperatures. Unfortunately, they take place over timescales long compared to those accessible to direct Molecular Dynamics simulations. Some extremely promising new tools, however, are opening the route towards accelerated algorithms for the simulation of thermal systems. They are based on numerical methods developed over these last two decades to determine thermally activated transitions in atomic systems. Of particular interest here is the Activation-Relaxation Technique (ART), an eigenvector-following method that allows the identification of activated states and paths in the potential energy landscape of atomic systems. In this study, we will show that although an exhaustive search for saddle points in case of disordered solids is unfeasible (because of the exponential number of activated states), ART can identify enough saddles to build statistically relevant samples, from which stationary distributions can be computed. The purpose of this strictly numerical thesis was the prediction of thermally activated kinetics in glasses such as those encountered experimentally. The nature of such miscroscopic events occuring in disordered systems was studied both under mechanical stress and in ageing conditions. We investigate two quantities that describe thermally-activated events within the harmonic approximation of the transition state theory, i.e. activation energy and attempt frequency.Since in the definition of an attempt frequency the curvature of the initial minimum and the saddle point are present, we wanted to see if there was a relation between attempt frequencies and activation energies of a given event in two types of systems: metallic glasses and silica glasses. Such correlation had been observed before for a wide range of phenomena and is referred to as the Meyer-Neldel compensation rule. We also attempt to answer if the simple BKS potential without Ewald summation is able to reproduce polyamorphism observed in silica glasses subject to hydrostatic compression and characterized mainly in terms of coordination numbers. Apart from thermally activated processes, the structural analyses of metallic and silica glasses were performed. The short and medium range orders were characterized using two methods: Voronoi tesselations for metallic glasses, providing us information about near neighbor conformations, and in case of silica, statistics of ring distributions.