Difference between revisions of "Georg Rempfer"

From ICPWiki
Jump to navigation Jump to search
Line 15: Line 15:
 
Electrokinetic phemonena comprise three distinct physical effects: hydrodynamics, electrostatics, and diffusion. Their complicated interplay gives rise to a multitude of interesting consequences, important for example in DNA electrophoresis and electroosmotic flow. What makes such systems hard to simulate is the fact that effects on very different time and length scales are involved. Typically, the dynamics of ions involved happen on the scale of picoseconds and nanometers, while the dynamics of macromolecules such as DNA can easily exceed time scales of milliseconds or even minutes for the folding of some proteins, and length scales of micrometers.
 
Electrokinetic phemonena comprise three distinct physical effects: hydrodynamics, electrostatics, and diffusion. Their complicated interplay gives rise to a multitude of interesting consequences, important for example in DNA electrophoresis and electroosmotic flow. What makes such systems hard to simulate is the fact that effects on very different time and length scales are involved. Typically, the dynamics of ions involved happen on the scale of picoseconds and nanometers, while the dynamics of macromolecules such as DNA can easily exceed time scales of milliseconds or even minutes for the folding of some proteins, and length scales of micrometers.
  
Previously, the molecular dynamics simulation software ESPResSo ([http://www.espressomd.org www.espressomd.org]), developed mainly by the Institute for Computational Physics in Stuttgart, was able to treat the diffusive effects only by using explicit particles in simulations like the ones carried out as part of my {{Download|BSc_thesis_rempfer.pdf|bachelor's thesis}}. To overcome the limitations imposed by the excessive requirements in computational power of this approach, I implemented an algorithm to treat the neutral and ionic species in such a simulation using a continuum mechanical model as proposed by Capuani et. al. in 2004 (http://arxiv.org/pdf/cond-mat/0404289.pdf). I did so, relying on NVIDIA's CUDA framework for GPU computing to implement it as a module in [http://www.espressomd.org ESPResSo].
+
Previously, the molecular dynamics simulation software {{es}} (http://espressomd.org), developed mainly by the Institute for Computational Physics in Stuttgart, was able to treat the diffusive effects only by using explicit particles in simulations like the ones carried out as part of my {{Download|BSc_thesis_rempfer.pdf|bachelor's thesis}}. To overcome the limitations imposed by the excessive requirements in computational power of this approach, I implemented an algorithm to treat the neutral and ionic species in such a simulation using a continuum mechanical model as proposed by Capuani et. al. in 2004 (http://arxiv.org/pdf/cond-mat/0404289.pdf). I did so, relying on NVIDIA's CUDA framework for GPU computing to implement it as a module in {{es}}.
  
 
=== Bachelor's Thesis ===
 
=== Bachelor's Thesis ===

Revision as of 09:46, 31 May 2013

Georg rempfer.jpg
Georg Rempfer
PhD student
Office:1.035
Phone:+49 711 685-67702
Fax:+49 711 685-63658
Email:georg _at_ icp.uni-stuttgart.de
Address:Georg Rempfer
Institute for Computational Physics
Universität Stuttgart
Allmandring 3
70569 Stuttgart
Germany

Master Thesis

application_pdf.png"A Lattice Model for Electrokinetics" (2.23 MB)Info circle.png, 2013, Institute for Computational Physics, Stuttgart

Electrokinetic phemonena comprise three distinct physical effects: hydrodynamics, electrostatics, and diffusion. Their complicated interplay gives rise to a multitude of interesting consequences, important for example in DNA electrophoresis and electroosmotic flow. What makes such systems hard to simulate is the fact that effects on very different time and length scales are involved. Typically, the dynamics of ions involved happen on the scale of picoseconds and nanometers, while the dynamics of macromolecules such as DNA can easily exceed time scales of milliseconds or even minutes for the folding of some proteins, and length scales of micrometers.

Previously, the molecular dynamics simulation software ESPResSo (http://espressomd.org), developed mainly by the Institute for Computational Physics in Stuttgart, was able to treat the diffusive effects only by using explicit particles in simulations like the ones carried out as part of my application_pdf.pngbachelor's thesis (1.36 MB)Info circle.png. To overcome the limitations imposed by the excessive requirements in computational power of this approach, I implemented an algorithm to treat the neutral and ionic species in such a simulation using a continuum mechanical model as proposed by Capuani et. al. in 2004 (http://arxiv.org/pdf/cond-mat/0404289.pdf). I did so, relying on NVIDIA's CUDA framework for GPU computing to implement it as a module in ESPResSo.

Bachelor's Thesis

application_pdf.png"Lattice-Boltzmann Simulations in Complex Geometries" (1.36 MB)Info circle.png, 2010, Institute for Computational Physics, Stuttgart

The focus of this thesis lies on the simulation of processes in molecular dynamics that are governed by electrostatic and hydrodynamic interactions in volumes with boundaries. The first objective was to change the existing implementation of the Lattice-Boltzmann-Method in ESPResSo so that it can handle systems with arbitrarily complex boundary geometries.

Poiseuille flow in tilted channels

Additionally, a scenario involving the fluid-boundary interaction and the fluid-particle interaction was supposed to be developed for which analytical results can be obtained, so that the correctness of the implementation could be verified. This system consists of an electro-osmotic flow in an infinite slit pore. The analytical treatment was conducted by means of the electrokinetic equations.

Counterion distribution in slit pore

Teaching