Hands-On Tutorial: Density Functional Theory and Beyond, Concepts and Applications
Fritz-Haber-Intitut der Max-Planck-Gesellschaft, Berlin, Germany
First-principles computational methods are instrumental for a fully predictive modelling of (bio)molecular and materials properties from the atomic scale on upwards: on their own (to predict specific properties for given materials or even materials for a specific purpose), as a fully independent complement to experimental science and traditional analytic theory, or as a means to evaluate and improve the accuracy of larger-scale model representation in a multiscale approach. The length scales, time scales, and accuracy requirements of first-principles predictions continue to evolve at a rapid pace, as do the methods which are used. For many years, the Fritz Haber Institute has organized a successful series of "Hands-On" tutorials, in order to allow new researchers (predominantly PhD students and young postdocs) to learn the basics up to the state of the art of the field as effectively as possible, lowering the otherwise considerable entry barrier to a field that acts as a moving target.
The tutorial concept consists of two main tracks:
· morning sessions with overview lectures on the basics and most important developments in the field, given by experts from different areas
· afternoon hands-on tutorials to apply these concepts directly, with immediate access to tutors and experts
In addition, participants are given the opportunity to present their own research or research goals as posters. A formal poster session will be held at the outset of the tutorial, but importantly, all posters remain available directly at the conference venue throughout the tutorial, offering ample time for discussions both among participants, as well as with the experienced tutors and lecturers at hand.
While the hands-on tutorials will necessarily be based on one particular implementation of DFT and beyond, we emphasize that the overall goal of the tutorial is a general introduction to first principles, not merely to teach a specific code. After completion, participants are expected to have obtained a good understanding of all the most important tools available: e.g., LAPW, plane waves, atom-centered basis sets, grid based approaches, etc. Based on this information, each participant should be equipped to choose the method(s) best suited for their own research direction.
Regarding the basic concepts of electronic structure theory, participants will gain a sound understanding of the basic yet powerful techniques behind ground state DFT, i.e., total energies, forces, relaxation, ab initio molecular dynamics, for molecules and solids, including the most important types of implementation. While ground state DFT remains the workhorse of the field, algorithmic developments to overcome specific accuracy limits, coupled with the continuing increase of the available computational power, continue to thrive, often as add-ons to DFT. Particularly important developments include: the proper description of quasiparticle excitations in the GW approximation (both non-selfconsistent or "self-consistent" in some way); Green's function based approaches to the ground state energy itself (RPA and beyond), or to transport properties; explicitly time-dependent methods for electrons; the limits of the Born-Oppenheimer approximation for nuclear motion in current simulations (e.g., in molecular spectroscopy), and how to overcome them; or, hybrid length- and time-scale bridging techniques that use first-principles input to parameterize computationally simpler Hamiltonians with first principles accuracy. Important examples of the latter approach include semi-empirical corrections for long-range dispersion interactions, or simplified parameterized interaction Hamiltonians for structure prediction in materials science.
The Harnackhaus conference center of the Max Planck Society has proven to be an excellent venue for a tutorial of the present scope. Not only does the center provide excellent conference facilities and sufficient accommodation for all participants in one location as a group, allowing to maximize the scientific interaction. In addition, the close proximity of the Harnackhaus conference center to the Fritz-Haber Institute is a key asset: It ensures a thorough, effective preparation of all computational equipment and exercises needed for the course well in advance, as well as the possibility to have many experienced lecturers and tutors on hand already, an essential prerequisite for a successful event.
The duration of the tutorial will be ten days.
Heiko Appel (Fritz Haber Institute of the Max Planck Society (FHI), Berlin) - Organiser
Matthias Scheffler (Fritz-Haber-Institut der Max-Planck-Gesellschaft) - Organiser
Volker Blum (Duke University) - Organiser