Seminar 20th May 2011 3:30 p.m. Building 58, Room 1007
What is Wrong with Density Functional Theory?
Professor Nicola Marzari, Director, Materials Modelling Laboratory
Department of Materials, University of Oxford
- Web page
- AMBER, C, CASTEP, Complex Systems, Computer Science, CVS, Density functional Theory, FFT, Finite elements, Fortran, GPU, GPU-libs, HECToR, HPC, Iridis, Materials, Metals, Molecular Dynamics, Multi-physics, Multi-scale, Nanoscale Assemblies, Onetep, OpenMP, ProtoMS, Scientific Computing, Software Engineering, Spintronics, SVN, Systems biology, Xmgrace
- Chris-Kriton Skylaris
Complex Systems Simulation Seminar Series (CS^4)
Electronic-structure modeling has become a very powerful tool to understand, predict, or design the properties of complex materials and devices. It is also an imperfect tool, with many open and urgent challenges in our quest towards qualitative and quantitative accuracy, and in our ability to perform quantum simulations under realistic conditions.
Several of these challenges stem from the remnants of self-interaction in our electronic-structure framework, leading to qualitative failures in describing e.g. mixed-valence complexes, electron-transfer excitations, and even single-particle energies. I’ll discuss these effects in realistic case studies, and suggest possible solutions based on constrained density functional theory (DFT), on extended Hubbard functionals, or on imposing a generalized Koopmans’ condition. I’ll also highlight how the calculation of magnetic properties (NMR/EPR) or the use of wavefunction techniques can provide stringent validation criteria for novel developments.
Available from 3:30pm, lecture starts at 4pm.
Multidisciplinary Research Co-ordinator
University Strategic Research Groups
Research and Innovation Services