This project aims at the ab-initio calculation of defect levels within the band-gap of novel semiconducting materials. This knowledge is important for the understanding of the mechanisms of de-excitation of electrons (holes) in the conduction (valence) band. Defect-mediated de-excitation determines the luminescence properties of semiconductors and can also influence the performance of photovoltaic materials. A precise determination of the defect levels and their interaction with radiation is therefore of high importance but remains a challenge for many materials of current interest. Already the precise calculation of band-gaps in pure (defect-free) materials is a challenge because it requires the use of the methods ofmany-body perturbation theory (MBPT) in order to take into account the effects of electron correlation. The calculation of defect levels with these methods adds additional complexity due to the requirement of large super-cells.Defect related optical properties will be studied using two complementary approaches: (1) The approach of solid-state physics: MBPT (“GW+Bethe-Salpeter”) starting from a plane-wave pseudopotential calculation on the level of density functional theory (DFT). (2) In parallel we will use the methods of quantum-chemistry in the embedded-cluster approximation (where the defect and its immediate neighborhood are represented by wave-functions and the surroundings are represented by point-charges and pseudo-potentials). The quantum-chemistry methods (multi-configuration selfconsistent field and multi-reference configuration interaction) allow for relaxation in the excited state which can be quite important to understand the shift between peaks in optical absorption and emission (luminescence). Both techniques will be further developed and adapted for the use on defects. The goal is to reach a precision that allows to associate measured luminescence spectra with specific defect geometries. In parallel, we will test and develop approximate methods: the use of hybrid-functionals in DFT has been quite succesful in the past for the calculation of band-gaps, but its performance for defects remains to be tested.A material of prime interest at the university of Luxembourg are the chalcopyrites Cu(In,Ga)Se2 and related materials which are used for the production of thin-film solar cells. Defects in these materials are still purely understood, but are known to strongly influence the performance of the solar cells. An important part of this project will thus be devoted to these materials, complementing the intense ongoing research in the Laboratory for Photovoltaics of the University of Luxembourg. Other systems to be studied are defects in hexagonal boron-nitride and boron-nitride nanotubes where we will investigate in particular the relaxation of the geometry in the excited state. Finally, we would like to compare the performance of the periodic supercell approach and the embedded quantum-chemistry approach for the study of excitons and colour centers in alkali halides.