zur Hauptseite

Institute for Advanced Simulation (IAS)

Navigation und Service


29.04.2015 13:30 Uhr
29.04.2015 14:30 Uhr
Jülich Supercomputing Centre, Hörsaal, Geb. 16.3, R. 222
1. Vortrag:
Calculation of Selected Thermodynamic Properties of Intermetallic Phases and Oxides for Advanced Energy Materials
Prof. Lorenz Singheiser, IEK-2

Determination and calculation of phase diagrams is essential to develop new materials for energy application. Complex phase diagrams are currently based on experimental determination of thermodynamic properties of key phases and intermetallic precipitates and described by the Calphad Method, which is based on appropriate physical models to inter- and extrapolate thermodynamic data and to calculate phase diagrams. This approach is quite time consuming and expensive. Therefore DFT calculations are interesting to calculate thermodynamic properties of intermetallic phases and compounds and to use these data as part of thermodynamic assessment of phase diagrams. In the frame of two research projects this approach has been applied to two different systems:

The first system considered are complex Fe-Cr-Si-Nb-W alloys, which are interesting as interconnect materials for future Solid Oxide Fuel Cells and future high strength steels for power plant applications, where high strength is strongly dependent on fine precipitates of Laves phases with defied composition. The alloy system consisting of a ferritic matrix and precipitates of Laves phases and mu phases shows a couple of discrepancies with respect to composition and existence regimes of different precipitates. Using DFT calculations thermodynamic data of selected Laves and mu phases have been calculated and the corresponding phase diagrams have been re-assessed.

The second example is represented by candidate anode materials for future battery systems. Using LixCoO2 as a well-established anode material thermodynamic properties, specifically heat capacities, have been calculated for discrete stoichiometries using DFT calculations and compared with experimental data. Calculated data are in good agreement with experimental data taking into account the simplifications as implemented in the quasi-harmonic model. To pursue this matter further doped LixCoO2 compounds will also be studied. In a separate work package, the enthalpies of formation of different compositions of Li-Co oxides in the range of x in [0.5;1] have been calculated within the GGA+U framework in an effort to model the metal-insulator transition within this compositional range. The knowledge of the correct evolution of enthalpies of formation are important for cell design and safe operation of future battery systems. The results offer a solid basis for future development of new anode materials for electrochemical devices.

Ankündigung als pdf-Datei:
 VSR Seminar, talk Lorenz Singheiser, on 29 April 2015 (PDF, 54 kB)
Wednesday, 29 April 2015, 13:30 - 14:00
Jülich Supercomputing Centre, Hörsaal, building 16.3, room 222
2. Vortrag:
Protein Structure Refinement at CASP11
Jun.-Prof. Gunnar Schröder, ICS-6

The prediction of protein structures from the amino acid alone is still a big challenge. In case a protein structure with a similar sequence is known, it can be used as a template to build a homology model. But homology models often have errors such as wrong loop conformations, incorrect side-chain packing and shifts of secondary structure elements. Our approach to structure refinement is based on molecular dynamics (MD) simulations. However, standard room-temperature MD simulations are usually not successful at refining protein structures, as the resulting structural ensemble is broad and we are interested only in the most probable structure, i.e. the minimum of the free energy landscape. We therefore simulate an ensemble of multiple copies which reduces structural fluctuations while sampling the conformational space. In addition the individual copies of the protein are not identical but chosen to have homologous sequences. With this the ensemble of protein structures effectively moves on a energy landscape averaged in sequence space.

The CASP (Critical Assessment of Techniques for Protein Structure Prediction) experiment is a biannual blind test for protein structure prediction. In the refinement category, the goal is to further improve a previously predicted model. In the talk our structure refinement method will be presented and results from the recent CASP11 experiment will be discussed.

Wednesday, 29 April 2015, 14:00 - 14:30
Jülich Supercomputing Centre, Hörsaal, building 16.3, room 222
Ankündigung als pdf-Datei:
 VSR Seminar, Talk Gunnar Schröder on 29 April 2015 (PDF, 28 kB)

Alle Interessierten sind herzlich eingeladen.