Browsing by Author "Schopf, Daniel"

Now showing 1 - 3 of 3

Open Access
Classical interaction potentials for diverse materials from ab initio data : a review of potfit
(2015) Brommer, Peter; Kiselev, Alexander; Schopf, Daniel; Beck, Philipp; Roth, Johannes; Trebin, Hans-Rainer
Force matching is an established technique to generate effective potentials for molecular dynamics simulations from first-principles data. This method has been implemented in the open source code potfit. Here, we present a review of the method and describe the main features of the code. Particular emphasis is placed on the features added since the initial release: interactions represented by analytical functions, differential evolution as optimization method, and a greatly extended set of interaction models. Beyond the initially present pair and embedded-atom method potentials, potfit can now also optimize angular dependent potentials, charge and dipolar interactions, and electron-temperature-dependent potentials. We demonstrate the functionality of these interaction models using three example systems: phonons in type I clathrates, fracture of α-alumina, and laser-irradiated silicon.
Open Access
Effective potentials for numerical investigations of complex intermetallic phases
(2013) Schopf, Daniel; Trebin, Hans-Rainer (Prof. Dr.)
The class of Complex Metallic Alloys (CMAs) is interesting for its wide range of physical properties. There are materials that exhibit high hardness at low density or good corrosion resistance, which is important for technological applications. Other compounds are superconductors, have strong anisotropic transport coefficients or exhibit a novel magnetic memory effect. The theoretical investigation of CMAs is often very challenging because of their inherent complexity and large unit cells with up to several thousand atoms. Molecular dynamics simulations with classical interaction potentials are well suited for this task – they can handle hundreds of thousands of atoms in reasonable time. Such simulations can provide insight into static and dynamic processes at finite temperatures on an atomistic level. The accuracy of these simulations depends strongly on the quality of the employed interaction potentials. To generate physically relevant potentials the force-matching method can be applied. A computer code called potfit has been developed at the Institute for Theoretical and Applied Physics (ITAP) especially for this task. It uses a large database of quantum-mechanically calculated reference data, forces on individual atoms and cohesive energies, to generate effective potentials. The parameters of the potential are optimized in such a way that the reference data are reproduced as accurately as possible. The potfit program has been greatly enhanced as part of this thesis. The optimization of analytic potentials, new interaction models as well as a new optimization algorithm were implemented. Potentials for two different complex metallic alloy systems have been generated and used to study their properties with molecular dynamics simulations. The first system is an approximant to the decagonal Al-Pd-Mn quasicrystal. A potential which can reproduce the cohesive energy with high accuracy was generated. With the help of this potential a refinement of the experimentally poorly determined structure model could be performed. The second class of potentials was fitted for intermetallic clathrate systems. They have interesting thermoelectric properties which originate from their special structure. Silicon- and germanium-based clathrate potentials were derived and the influence of the complex structure on the thermal conductivity has been studied.
Open Access
Effektive Potenziale für komplexe metallische Phasen
(2009) Schopf, Daniel
Die Xi-Phasen des Aluminium-Palladium-Mangan (AlPdMn) sind Approximanten für einen dekagonalen Quasikristall mit einer Gitterkonstanten von 1.6 nm in der periodischen Richtung. Diese Systeme können jedoch aufgrund ihrer Größe nicht mit ab-initio-Methoden berechnet werden, die Einheitszellen dieser Phasen umfassen jeweils mehrere hundert Atome. Aus diesem Grund ist man für Molekulardynamiksimulationen auf die Verwendung effektiver Potenziale angewiesen. In dieser Arbeit werden die dazu benötigten Potenziale mit Hilfe der Force-Matching-Methode erzeugt. Dabei werden die Parameter eines Potenzials so angepasst, dass sie die quantenmechanisch berechneten Referenzdaten wie Kräfte, Energien und Spannungen möglichst gut wiedergeben können. Das Programm potfit wurde um die Möglichkeit erweitert, die Parameter analytischer Potenziale zu optimieren. Es wurden neue analytische EAM-Potenzialmodelle entwickelt und an verschiedenen metallischen Systemen ausführlich getestet. Für die Systeme Magnesium-Zink und Aluminium-Palladium wurden verschiedene Potenziale erzeugt und miteinander verglichen. Für die Xi-Phasen des Aluminium-Palladium-Mangan, insbesondere xi und xi', wurden Potenziale erzeugt, die für die Strukturoptimierung verwendet werden. Die ab-initio berechneten Referenzdaten können dabei sehr gut reproduziert werden.