Quantum-mechanics based theories such as density-functional theory now allow one to address nanostructures of up to a few thousand atoms, as well as periodic solids, surfaces, and nanowires, their defects, electronic structure, optical properties etc. The field is moving fast, fueled by the advances of computers, and by the algorithms, approximations, and computer codes developed by groups such as ours.
Work in the group falls into three categories:
- Materials and interfaces. We use techniques based on density functional theory and many-body theory to study the structure, electronic and/or chemical properties of inorganic materials. Current work in the group focuses on organic-inorganic hybrid semiconductors, on materials for photovoltaic applications, graphene growth on SiC for electronic applications, and the structure and mechanisms behind photocatalytic activity of graphitic carbon-nitride materials.
- Molecular structure and spectroscopy. Density-functional theory and many-body techniques applied to molecular systems. Work in the group has focused on structure prediction of the "building blocks of life", amino acid and peptide molecules. Together with Thomas Theis and the group of Warren Warren here at Duke, current work focuses on predicting nuclear spins states and time evolution for applications in nuclear magnetic resonance.
Pushing the computational methods that form the basis of our field.
This work includes density-functional methods, the main workhorse of computational materials science today, but also methods that go beyond density-functional theory to enable us to capture many-body electron correlation, electronic or optical properties of materials.
The most important development in our group is the "FHI-aims" all-electron electronic structure package, of which Volker is the coordinator (together with Matthias Scheffler in Berlin) and lead developer. FHI-aims enables accurate simulations of systems comprising thousands of atoms and scales up to computer platforms with thousands of CPUs or more. The code originated at the Fritz Haber Institute of the Max Planck Society in Berlin, Volker's scientific home for many years, and is now heavily developed in Volker's group at Duke, in Matthias Scheffler's group at FHI, and in several other groups around the world.
Within a NSF SI2-SSI open-source infrastructure project "ELSI" (Electronic Structure Infrastructure), we are now taking this work to the next level for large-scale, accurate, efficient, scalable electronic structure theory.
A key solver supported in ELSI is the ELPA library, which facilitates the solution of real symmetric and complex Hermitian eigenvalue problems on today's largest supercomputer architectures. A current overview of ELPA and its capabilities was published as a psi-k.org Scientific Highlight in December, 2014.