We model the interaction of light and matter as well as chemical reactions in real time. Our aim is not only to unravel the underlying processes but also to control them. The systems under consideration reach from diatomics to DNA. Accordingly, different techniques like quantum dynamics or molecular dynamics calculations are employed.
SHARC - Surface Hopping including ARbitrary Couplings
We have developed a nonadiabatic ab initio molecular dynamics (MD) method including spin-orbit coupling (SOC) and laser fields. This mixed quantum classical dynamics method is used as a general tool for studies of excited-state processes. Intersystem crossing - e.g. a transition from a singlet to a triplet state - can be investigated within the given framework. The laser interaction is treated on a non-perturbative level that allows to consider nonlinear effects like strong Stark shifts. As MD allows for the handling of many atoms, the interplay between triplet and singlet states of large molecular systems are accessible.
Ab initio molecular dynamics software SHARC: sharc-md.org
Neural networks and machine learning
Artificial neural networks are computer algorithms mimicking the operating mode of our brain. These neural networks belong to the vast field of machine learning and have been employed for a wide range of applications. We use them to predict potential energy surfaces and other molecular properties. These neural networks offer the exciting advantage of delivering highly accurate results at little computational cost.
SERS and related spectrosc
opies
Surface enhanced Raman scattering (SERS) is a promising technique making use of nanotechnology for various applications like specific DNA detection or development of new materials. The basis for SERS is the Raman effect, where photons are inelastically scattered from atoms or molecules. The sensitivity of Raman scattering can be drastically increased by the SERS effect, allowing even for single molecule detection. The reason for the enhancement is thought to be due to two phenomena: 1.) Electromagnetically induced surface plasmons and 2.) chemisorbtion, i.e. building of chemical bonds with the surface and charge-transfer complex formation. We are working on an implementation of both effects in quantum mechanical simulations to disentangle the different contributions.
Laser control of chemical reactions
One of the central problems in chemistry is to control the outcome of reactions. This goal can be achieved using laser light. In general, every molecular system can be modified by a molecule-field interaction to yield a desired product. This universality arises from the large variety of laser parameters, which can be adapted to yield electro-magnetic fields of the most different shapes and colors (see e.g. the Wigner representation of a third-order chirped pulse on the left). In this way, chemical bonds can be broken and formed selectively. We use quantum dynamics (QD) simulations to unravel the mechanism underlying such photodissociation and association processes.
De-novo 
enzyme design
Catalysis is one of the most important processes in chemical reactions. In nature, the task as biocatalysts is fulfilled by enzymes (usually proteins), providing the highest efficiency and selectivity. Hence, enzymes are employed also in the laboratory for chemical reactions under mild conditions. However, enzymes are not available for all important chemical reactions. Therefore, the field of de-novo enzyme design envisages to computationally identify an amino acid sequence, which yields an enzyme for a specifically targeted, artificial reaction. The necessary knowledge of the reaction's transition state as well as the protein folding is obtained by different tools ranging from ab initio quantum chemistry to bioinformatics.
