Regensburg 2019 – scientific programme
Parts | Days | Selection | Search | Updates | Downloads | Help
MM: Fachverband Metall- und Materialphysik
MM 4: Methods in Computational Materials Modelling (methodological aspects, numerics)
MM 4.6: Talk
Monday, April 1, 2019, 11:45–12:00, H45
On finding the dipole in the haystack of MD trajectories — •Frederic Felsen, David Egger, and Christoph Scheurer — Theoretical Chemistry, Technical University Munich
Aiming at implicit solvation models for large scale simulations of chemical reactions, we took a first step in the development of an automated solvent characterization based on classical MD trajectories. By modifying a method for the detection of semi-rigid domains in macromolecular structures [1], we propose a coarse graining procedure for decomposing a charge neutral atomistic system into a multi-center dipole representation. The procedure is formulated in terms of a series of linearly constrained quadratic programming problems, which describe a clustering of the particles with respect to three major properties: spatial compactness, as the multi-center expansion is in terms of ideal point dipoles, structural rigidity, to ensure a valid mapping over the whole trajectory, and local charge neutrality. Solving these optimization problems by means of an active set algorithm yields the dipole domain decomposition of the system of interest. Based on such a dipole representation, an expression for the system’s polarization can be obtained which is directly related to the dielectric function of the medium [2]. Investigating molecular (liquids and macromolecules) as well as crystalline (MoS2 sheet structures) test systems, the key features of the coarse graining will be discussed.
[1] S. Bernhard, F. Noé, PLOS ONE, 5, (2010).
[2] D. Bonthuis, S. Gekle, R. Netz, Langmuir, 28(20), (2012).