2. Electronic structure of solids and nanostructures
In the last 20 years the importance of computational materials science has grown tremendously. This was made possible mainly by three developments: i) the power of modern computer hardware has continued to grow exponentially and has reached a level where large scale simulations of “real” materials are possible, ii) the continuous development of density functional theory (DFT) allows a more accurate description of the electronic structure leading eventually to a predictive power of ab initio methods, iii) developments of new algorithms, their efficient implementation in modern computer programs and a reasonably user-friendly interface which allows also non-experts to use such codes.
Over the last years the group of Peter Blaha has developed WIEN2k, a computer program to calculate the electronic structure and several related properties of solids. It is based on the very accurate augmented-plane-wave method and has become one of the most widely used program packages in this research area. It has been licensed by more than 1250 groups worldwide and is well known for its applicability and predictive power for all classes of materials. Part of its success is also due to good support and continuous development of WIEN2k, in particular with technical enhancements (new algorithms, speed, parallelization, userfriendliness), new DFT functionals (better GGAs or hybrid DFT-functionals) and new features and properties which can be calculated.
Besides code development we are of course also applying our code to all kinds of different materials problems, ranging from magnetic and highly correlated materials to various other metallic, semiconducting or insulating solids up to surfaces, interfaces and nanostructures. Most recently we could unravel the structure of the BN/Rh(111) “nanomesh” and explain the functionality of the nanomesh when trapping isolated molecules. This is a nanostructure with a periodicity of 3.2 nm, which requires simulations of more than 1100 atoms / unit cell and can only be treated on modern computer clusters with (at least) 100 cores, still requiring a few months of computer time.
Cooperations with other DFT experts of the CompMat consortium (Mohn, Redinger) have tradition and a new collaboration (Held) has been started and is funded by the FWF GK computational materials science to integrate DFT and dynamical mean field theory (DMFT). Collaborations with numerical mathematicians can always be very fruitful as the joint development of an improved preconditioner for the iterative diagonalization has shown (Koch). After the first meeting of the CompMat consortium we definitely see also possible collaborations and synergy with the groups of Fidler and Kozeschnik.