In this thesis, the successful implementation of optical excitation for time-resolved Scanning Tunneling Microscopy (STM) is presented. The fruitful combination of these two experimental methods allows investigating photo-induced dynamic processes on the nanosecond time scale with atomic resolution. The optical setup provides a great versatility regarding the adjustment of excitation parameters such as optical pulse height, pulse width or pulse repetition rate to the experimental needs. Moreover, for the first time, it is possible to disentangle and quantify thermally induced effects, e.g. thermal expansion of the STM tip, from the originally inquired signals, resulting from photo-triggered charge dynamics at the sample surface.Using continuous wave optical illumination at the Gallium-Arsenide(110) (GaAs) surface, it is proven in this thesis, that the presence of excited holes, accumulating in the tip-induced Space Charge Region (SCR) beneath the STM tip not only creates a Surface Photovoltage (SPV) but also opens an additional tunneling channel. Current dependent studies show that this extra tunneling process exhibits a high impact on the charge concentration at the surface. Assuming a steady state between charge generation via optical excitation and charge annihilation via the tunneling process, parameters such as the diffusional or field driven transport rate towards the surface are determined.The build-up and relaxation of an SPV is connected to various microscopic processes, e.g., charge transport or carrier recombination. With pulsed optical excitation, these mechanisms can be probed and disentangled. Whereas for low tunnel rates the decay of the system after optical excitation is mainly determined by the charge annihilation via the tunneling process, for high rates the recharging of dopants becomes visible. Studies of this ionization process for donors, positioned at different depths beneath the surface, reveal a significant local inhomogeneity. By applying a field driven ionization mechanism, it turns out, that the dopants cannot be treated independently from each other. Instead, the ionization dynamics have to be treated as an interacting network of coupled, locally fixed and randomly distributed charge centers.