|Phone:||+49 711 685-67715|
|Fax:||+49 711 685-63658|
|Email:||mkuron _at_ icp.uni-stuttgart.de|
Institute for Computational Physics
I am a PhD student in Christian Holm's group, working on lattice-Boltzmann simulations of cooperative behavior of active particles.
"Efficient Lattice Boltzmann Algorithms for Colloids Undergoing Electrophoresis" (file does not exist!), 2015, Institute for Computational Physics, Stuttgart (Download will be available within the next few months)
For this thesis, waLBerla, a highly-scalable grid framework for applications such as lattice-Boltzmann and solving partial differential equations, was extended so that it can be used for simulating the electrokinetics of active colloids.
"Like-Charge Attraction in DNA" (2.3 MB), 2013, Institute for Computational Physics, Stuttgart
For this thesis, the MMM1D algorithm was ported to GPGPU. This resulted in a 40-fold performance increase over the previous implementation in ESPResSo and now allows for Molecular Dynamics simulations with electrostatic interactions in 1D-periodic geometries with several thousand particles.
Using this, simulations with various simple DNA models were performed. These simulations show that charge discretization and phase shifts between DNA molecules, modeled as rods, have a significant influence on their attractive properties, an effect that previous works disregarded as it was computationally too expensive, even though it turns out to be too large to neglect for realistic results. Curling up the discretely charged rods into helices, thus making the most accurate model of DNA that could be simulated with the limits of time and resources for this thesis, reveals further geometry dependencies and again a strong influence of a phase shift between the two helices. For phase shifts of 180°, the results for the continuous rods are mostly recovered for large Bjerrum lengths, but for any other phase shift, the forces are weaker, albeit still attractive.