Scientific Software

Currently (2014-) at University College London I am developing software for structure prediction of nanoclusters.

Nanoclusters, or small nanoparticles, have typical dimensions below 2–5 nanometers, a size regime where current experimental techniques are insufficient for accurate and comprehensive structure characterisation. Fortunately, that is where computational approaches can usefully complement and aid experimental studies.

At the research group at UCL I am responsible for the designing and developing software of such computational approaches, which includes global optimisation techniques (KLMC); software for the back-end and front-end of the WASP@N web interfaced database to search, analyse and disseminate nanoclusters; a novel scientific software package based on statistical mechanics (GCAT) to evaluate the stability of the nanoclusters' compositions, determine their average properties and estimate free energies.

Previously (2011-2014), before I have joined UCL, during my PhD at Loughborough University I was developing and maintaining software to model physical processes occurring during radiation damage events.

Radiation events are usually driven by energetic particles and their effect on the nearby atoms in physical objects. Within the first few ps of a radiation event, also known as the ballistic phase, the initial damage region becomes highly distorted and leads to a thermal spike, a quick local temperature increase, and thermally driven processes occur which can be modelled using the Molecular Dynamics technique. After the ballistic phase, the system is left with metastable defects as it reaches local minimum and the processes of defect recombination, migration and clustering take place on much longer time scales - the so called recovery phase, which cannot be simulated using MD and other techniques, such as Kinetic Monte Carlo, have to be employed.

In order to model radiation damage events, two in-house software packages were exploited. LboMD was used to model the ballistic phase and LAKMC was used to model the recovery phase. To read more about the software and the project, please check my PhD thesis.