Characterisation of the Free Energy Landscapes of Polymeric Polymorphs using data driven methods
The higher dimensional the free energy landscapes (FELs) are often expressed in terms of essential degrees of freedom known as collective variables (CVs). Data-driven techniques provide a systematic route to construct the free energy landscape without the need for extensive a priori intuition about the system. We are looking at the polymorphic transitions of polymer melts using data driven methods to overcome the two main problems: identifying goods CVs and the inherent longer timescales of the transitions.
Connecting thermodymanics to the dynamical properties of supercooled liquids
In liquid state theory, there is a well-known connection between dynamical and structural properties. However, glass-forming systems such as the Kob-Andersen model with interactions described by Lennard-Jones (LJ) and Weeks-Chandler-Andersen (WCA) potentials exhibit seemingly similar structural properties and entirely different dynamical behaviour. Earlier we have shown that this difference in dynamics is orgininated from a thermodynamical quantity. Now, we use the Kirkwood-Buff solution theory to compute thermodynamic quantities which are related to the microscopic structure of the systems.
Crystal structure prediction using basin hopping method
Organic molecules can be stable in different crystalline form known as polymorphs. Finding these polymorphs could be extemely computationally expensive as they are separated by high energy barriers. We use powerful basin-hopping- global-optimization to predict different low energy structures for benzene crystal.
- Papers
- The Journal of Physical Chemistry A (2021)