Coding Solvation Workshop

Implicit Solvation Functionality in FHI-aims: Kirkwood Multipole Expansion Mode

Markus Sinstein(Technical University of Munich)
Karsten Reuter (Technical University of Munich)
Harald Oberhofer (Technical University of Munich)

Implicit solvation models - used within first-principles calculations for the
solute - provide a numerically most efficient way to effectively capture
solvation effects. To this end we have implemented the multipole expansion
(MPE) model, originally introduced by Kirkwood, into the full-potential
density-functional theory (DFT) code FHI-aims. We will briefly present the
theoretical background of the model and focus mainly on the implementation
which fully exploits the optimized multi-center multi-pole decomposition of the
electron density performed within FHI-aims. This leads generally to an
insignificant computational overhead compared to the underlying DFT calculation
for the solute. Aiming to minimize the number of free parameters inevitably
connected with such implicit models, we use an isodensity definition of the
solvent cavity. Finally, as to the other parameters, we present an efficient
parametrization scheme based on experimentally measured hydration energies of
small organic molecules.