Jayeeta Ghosh

State point dependence of systematically coarse–grained potentials

We apply systematic structural coarse–graining based on optimizing a potential against the structure obtained in atomistic simulations to the small organic glass former ortho-terphenyl (OTP). Atomistic radial distributions at various temperatures ranging from below the glass transition temperature to the equilibrium liquid show hardly any change with temperature. These pair distribution functions are used as targets to be reproduced by a mesoscale model of OTP which is formulated by replacing each benzene ring with a single interaction center. The potentials are obtained by Iterative Boltzmann Inversion of the distribution functions. The resulting potential depends not only on the structure but also implicitly on the temperature at which it was optimized. Potentials optimized in the liquid range lead to crystalline structures if used below the glass transition requiring independent optimizations in the glass. We compare potentials optimized in both ranges to study the system over the whole temperature range. The dynamic mapping turns out to be different for different mapping potentials.

A comparative molecular simulation study of the glass former ortho-terphenyl in bulk and freestanding films

We performed molecular dynamics simulations of the low-molecular weight organic glass former ortho-terphenyl in bulk and freestanding films. The main motivation is to provide molecular insight into the confinement effect without explicit interfaces. Based on earlier models of ortho-terphenyl we developed an atomistic model for bulk simulations. The model reproduces literature data both from simulations and experiments starting from specific volume and diffusivity to mean square displacement and radial distribution functions. After characterizing the bulk model we form freestanding films by the elongation and expansion method. These films give us the opportunity to study the dynamical heterogeneity near the glass transition through in-plane mobility and reorientation dynamics. We finally compare the model in bulk and under confinement. We found qualitatively a lower glass transition temperature for the freestanding film compared to the bulk.

Simulations of glasses: multiscale modeling and density of states Monte-Carlo simulations

We show the application of two recent developments in molecular simulations, density of states (DOS) (Wang–Landau) Monte-Carlo and multiscale modeling, to the understanding of the glass transition. First, we review DOS Monte-Carlo using the two-dimensional Ising model without external field on lattices of varying size. We point out that we can analyze the resulting densities of state in a canonical and a microcanonical way starting from the same simulations. The heat capacity is discussed in both ensembles. Results from both ensembles of off-lattice simulations of a model binary glass former are compared as well. Subsequently, a self-consistent systematic mapping procedure for molecular models from the atomistic to the mesoscale is presented. It allows to efficiently derive mesoscale models with fewer interaction centers from atomistic models preserving the molecular identity. We use the optimization of a corresponding mesoscale model for atactic polystyrene (a good glass former) in the melt as an example. Simulations of different temperatures using this model allow some insight into the glass formation of this system. We point out the strengths and weaknesses of these approaches and give an outlook towards their combination.

Comparing the density of states of binary Lennard-Jones glasses in bulk and film.

We used Wang-Landau density of states Monte Carlo to study a binary Lennard-Jones glass-forming mixture in bulk and films between noninteracting walls. Thermodynamic properties are calculated using two different ensembles and film data are compared with the bulk. Bulk properties are in good agreement with previous simulations. We confirm the formation of a glass using various properties, e.g., energy, heat capacity, and pressure with temperature. We find a change in slope in the energy per particle and pressure as a function of temperature. We do not find any defined crystal structure. A higher glass transition temperature is found for the film.

Molecular Dynamics Simulation of the Glass Transition of Ortho-Terphenyl in Bulk and Thin Films

The glass transition temperature in thin film depends strongly on film thickness and interaction with the substrate and it is normally a priori not clear which way it deviates from the bulk value. This causes new challenge in the technological advancement of smaller and smaller electronic devices. In this study molecular dynamics simulations of a low-molecular weight organic glass former, ortho-terphenyl, are carried out in bulk and freestanding films. The main motivation is to provide insight into the confinement effect without interface interactions. Based on earlier models of ortho-terphenyl we developed an atomistic model for bulk simulations. The model reproduces the literature data from simulations as well as experiments. After characterizing the bulk model we form a freestanding film. This film gives us the opportunity to study the dynamical heterogeneity near the glass transition by in-plane mobility and reorientation dynamics. We also develop a structurally coarse-grained model for this glass former based on our atomistic model to study bigger system for a longer period of time.

Density of States Simulations of Various Glass Formers

The behavior of glassy systems in bulk and especially in confined geometries has received considerable attention over the last decades because of the technological importance and inherent complexity of the systems near or below the transition temperature. Confined glasses have been studied using different theoretical and experimental techniques which helped shape our understanding; but still huge gaps remain. In this work we are using the Wang–Landau Monte Carlo approach to study different model glasses. General Monte Carlo fails to sample all relevant regions of phase space; the application of this method gives us the opportunity to directly estimate the density of states and consequently any other thermodynamic properties. We can calculate properties in different ensembles using the same simulation runs. This random walk algorithm is designed to visit all energy states with equal probability to produce a flat histogram. We can estimate the density of states on the fly whenever any energy state is visite...