Final Activity Report Summary - SIMGLASS (Molecular Simulation Study of Ageing and Plasticity in Glassy Materials) This project aimed at the development of a hierarchical computational approach for the prediction of the thermodynamic and structural properties of glasses, based on their composition and the conditions of their formation. In this work, computational methodologies founded on multidimensional transition state theory were extended and applied in order to understand and predict the phenomena of physical ageing and plastic deformation of glassy materials at temperatures below Tg. Our perspective, which allowed for the direct study of relaxation and deformation phenomena in glassy materials for temperature smaller than Tg, was based on the idea that the molecular configuration of a glassy material was locally trapped in the neighbourhood of a local minimum of the potential energy, i.e. it was inherent structure. The system was considered as fluctuating around such a minimum, or around a small number of neighbouring minima, in the complex multidimensional hypersurface spanned by all atom positions. Structural relaxation occurred as a result of infrequent transitions to other minima, which required overcoming high energy barriers. Plastic deformation also entailed transitions to neighbouring minima induced by the presence of external stresses. The project had two principal research objectives: 1. the computational study of physical ageing; and 2. the computational study of the mechanical properties of glassy materials. Our approach was founded on the multidimensional Transition state theory (TST) in order to trace elementary structural transitions between neighbouring energy minima in the configuration space of a glassy system and the calculation of the corresponding rate constants. Thermal fluctuations were taken into account by incorporating a Quasi harmonic approximation (QHA) for the vibrational motions of atoms around the configuration of mechanical equilibrium; thus, entropic as well as energetic contributions to the thermodynamic and dynamical properties were included. The computational methodology for the identification of saddle points and the tracking of transition paths was extended, so as to deal with the rugged multidimensional energy hypersurfaces of glassy materials in a manner that combined efficiency and predictive power. The duration of the project was three years. During the first year we developed the following tools that would give us the ability to reach our objectives: 1. generation of atomistic minimum energy configurations of our model amorphous polymer system, atactic Polystyrene (a-PS), under the condition of constant density; 2. implementation of the QHA for calculating the normal mode frequencies and the vibrational free energy; 3. development of Molecular mechanics' (MM) code for the density relaxation of the PS model system under imposed strain at finite temperature; 4. development and tests of MM-based software, capable of performing computational deformation experiments using the QHA in the PS model system; 5. serial and parallel code for linking pairs of neighbouring glassy minima via the dimer saddle point search method; 6.numerical experiments for the study of small deformations of glassy PS (elastic region); 7. formulation of a novel approach to Kinetic Monte Carlo (KMC), on the basis of Markovian web integration, which we anticipated to enable us to sample more efficiently the very broad time scales of relaxation phenomena of the glassy polymers. By the time of the project completion we were able to perform atomistic simulations based on QHA and to evaluate the elastic constants of our model PS system, including entropic contributions, from chemical constitution. Our subsequent goal was to combine this procedure with the algorithm that linked neighbouring inherent structures via the dimer saddle point search method and our novel KMC scheme. We anticipated that we would be able to describe the dynamical evolution of the systems under the imposition of external stress, or strain, at a prescribed rate. Moreover, our multilevel parallelisation, in combination with our novel KMC scheme, would allow us to simulate for the first time deformation over long time periods under conditions comparable to those utilised in most experiments as well as to physical ageing in our polymeric glass model.
