QCM-2014-2-0036

In order to get a deep understanding on how the sub-units are connected with each other over the time, we need to fragment the whole density in its local electronic densities. The possibility of computing the local dipole moments (local multipoles in a general case) over the time will give a valuable information in order to discern between different mechanisms of the EET in plants.
A preliminary implementation of a part of the code that enable to compute these local properties has been recently included on the source code. However, nowadays it is still not adapted to deal with very big systems with millions points of the mesh.
New strategies for the fragmentation of the total density will be implemented, such an efficient and parallel version to compute Bader surfaces following a near-grid method without lattice bias.
Besides, with the well defined local domains, we will be able to implement algorithms to compute local properties such as potential interaction or current density between sub-units.

Both implementations will be tested first over small clusters of couple of chlorophylls (~300-600 atoms) up to the first steps of one monomer of the global system which involves ~6000 atoms.
This kind of calculations requires a big amount of computational time and resources due to the huge number of involved states as well as the huge number of space points needed to cover the whole system volume (~1548 angs³). This fact makes necessary the use of Supercomputing facilities.