Functionalization of Si baseline technology
Silicon (Si), dominating the platform for integrated circuitry (IC) technologies over the last 50 years, is running into fundamental physical limits with further miniaturization (“More Moore”) and / or functionalization (“More than Moore”). In this respect, right material candidates must be identified to enable new innovative functions without compromising the high performance of the Si chip baseline technology. One of such potential candidates is Germanium (Ge) experiencing a renaissance in modern microelectronics due to its superior optoelectronic properties with respect to Si (band gap, charge carrier mobility etc.). In consequence, Ge is a promising material for building up future photonic modules within the “More than Moore” approach [1-3]. Despite clear advantages of Ge integration in Si technology (identical diamond crystal structure with Si, CMOS compatibility due to no contamination risks etc), it faces also some true challenges limiting the optoelectronic performance.
The major “stumble block” is given by a large lattice mismatch of 4.2 % between Ge and Si, resulting in too high defect densities  as well as complex, strain driven SiGe interdiffusion phenomenon . The Ge thin film quality in terms of defect density (scattering and recombination centres) and purity (SiGe alloys adopt Si optoelectronic properties at rather small Si interdiffusion concentration) is of utmost importance. In this respect, Si CMOS compatible Ge heteroepitaxy approaches for high quality Ge micro- and nanostructure growth with reasonable thermal budget must be developed. Such techniques like direct deposition with cyclic annealing , SiGe  and oxide buffers , selective growth in mesa windows  or aspect ratio trapping  were investigated in the past to realize high quality Ge on Si substrates but limits in terms of defects and purity still exist. In the present project, alternative advanced Si CMOS compatible Ge heteroepitaxy approaches are investigated to push the quality of Ge micro- and nanostructures on Si(001) beyond the state-of-the-art.
Nanoheteroepitaxy for fully relaxed Ge
A very promising method to reduce the density of threading dislocations is to grow Ge selectively in a limited area. The compliant behavior of Ge/Si nanostructures offers a promising vision: In contrast to the classical Ge deposition on bulk Si substrates, the strain energy accumulated in the Ge epilayer is partially transferred to the free-standing Si nanostructure. This strain partitioning phenomenon is at the very heart of the compliant substrate effect and, if strain energy levels are correctly balanced, offers the vision to grow defect-free nanostructures of lattice mismatched semiconductors on Si .
Fig. 1 Scanning and transmission electron microscope images
of Ge nanoclusters on free-standing Si(001) nanopillars.
Thus, the Ge nanoheteroepitaxy at IHP aims on the experimental verification of the vision to realize Ge/Si nanostructures of high structural quality for future photonic applications within this innovative growth approach. Certainly, the successful establishment of such a complex growth scenario will also be of high importance for the integration of other lattice mismatched semiconductors (e.g. GaAs) on the Si technology platform.