This thesis deals with exciton relaxation processes in thin polycrystalline films and matrix-isolated molecules of the perylene derivatives PTCDA (3,4,9,10-perylenetetracarboxylic dianhydride) and MePTCDI (N,N'-dimethylperylene-3,4,9,10-dicarboximide). Using femtosecond pump-probe spectroscopy, transient absorption spectra, excitonic relaxation in the lowest excited state subsequent to excitation, and exciton-exciton interaction and annihilation at high excitation densities have been addressed. Transient absorption spectroscopy in the range 1.2eV-2.6eV has been applied to thin polycrystalline films of PTCDA and MePTCDI and to solid solutions of PTCDA and MePTCDI molecules (monomers) in a SiO2 matrix. We are able to ascribe the respective signal contributions to ground state bleaching, stimulated emission, and excited state absorption. Both systems exhibit broad excited-state absorption features below 2.0eV, with dominant peaks between 1.8eV and 2.0eV. The monomer spectra can be consistently explained by the results of quantum-chemical calculations on single molecules, and the respective experimental polarization anisotropies for the two major transitions agree with the calculated polarizations. Dimer calculations allow to qualitatively understand the trends visible in the experimental results from monomers to thin films. The broad excited state absorption band between 1.8eV and 2.0eV allows to probe the population dynamics in the first excited state of thin films. We show that excitons created at the Gamma point relax towards the border of the Brillouin zone on a 100fs time scale in both systems. Excitonic relaxation is accelerated by increase of temperature and/or excitation density, which is attributed to stimulated phonon emission during relaxation in k-space. Lower and upper limits of the intraband relaxation time constants are 25fs (resolution limit) and 250fs (100fs) for PTCDA (MePTCDI). These values agree with the upper limit for the intraband relaxation time of 10ps, evaluated from time-resolved luminescence measurements. While the luminescence anisotropy is in full accordance with the predictions made by a luminescence anisotropy model being consistent with the exciton model of Davydov-split states, the pump-probe anisotropy calls for an explanation beyond the models presently available. At excitation densities 10^(19)cm^(-3), the major de-excitation mechanism for the relaxed excitons is exciton-exciton annihilation, resulting in a strongly reduced exciton life time. Three different models for the microscopic behavior have been tested: a diffusion-limited annihilation model in both three and one dimensions (with diffusion constant D as fit parameter) as well as a long-range single-step Förster-type annihilation model (with Förster radius RF as fit parameter). For PTCDA, the latter two, being structurally equivalent, allow to fit a set of multiexponential decay curves for multiple initial exciton densities with high precision. In contrast, the three-dimensional diffusion-limited model is clearly inferior. For all three models, we extract annihilation rates, diffusion constants and diffusion lengths (or Förster radii), for both room and liquid helium temperature. Temperature dependence and orders of magnitude of the obtained parameters D or RF correspond to the expectations. For MePTCDI, the 1D and the Förster model are in good agreement for a smaller interval of excitation densities. For a initial exciton densities higher than 5 x 10^(19)cm^(-3), the 3D model performs significantly better than the other two.

We have measured the optical properties of films of the organic semiconductors PTCDA (3,4:9,10-perylene-tetracarboxylic dianhydride) and HBC (peri-hexabenzocoronene), prepared by Organic Molecular Beam Expitaxy (OMBE), on different substrates by means of Differential Reflectance Spectroscopy (DRS). The optical setup enables us to directly follow the thickness dependent optical properties of the organic films, starting from submonolayer coverage up to thicker films on the order of 20 monolayers (ML) film thickness. Due to the different optical nature of the different substrates used, i.e., mica, glass, Au(111), and HOPG, the direct interpretation of the DRS signal is not feasible. Therefore, we have proposed a method by which the calculation of the optical constants n (index of refraction) and k (absorption index) of thin films on arbitrary substrates from just one spectral measurement (in our case the DRS) becomes possible. The results fulfill a priori a Kramers-Kronig consistency and no specific model is needed to express the spectral behavior of the optical constants. Based on our method, we have successfully calculated the optical constants, and therefore the absorption behavior, of films of different thickness of PTCDA on mica, glass, Au(111), and HOPG, as well as of HBC on mica, glass, and HOPG. Extrinsic effects due to island growth or the presence of a polarizable substrate (screening) have been accounted for. We have introduced a finite dipole model which considers the extended geometry and anisotropy of the organic molecules. The calculated absorption behavior is discussed in great detail in terms of spectral changes with varying film thickness, different growth modes, degree of ordering of the films, interactions with the substrates and oscillator strength. A direct observation of a monomer-dimer transition in solid films could be observed for the first time. Our results indicate an exciton delocalization over about 4 molecules for both molecules.