Silicon-based near-infrared light sources

New optical communication standards in datacenters and supercomputers as well as on-chip optical interconnects require optical components and light emitters that are compatible with standard Si technologies. In two of the contributing projects to the preceding Spezialforschungsbereich SFB 025 IRoN we showed that silicon-germanium (SiGe) quantum dots are a promising approach toward silicon-based light sources when embedded with high spatial precision in special optical components, so-called photonic crystals or photonic crystal resonators. In this stand-alone proposal we focus on the two most urgent problems of this approach on the route toward device implementations. First, we aim for improved emission line widths and temperature stability of the SiGe quantum dot emitters. For that purpose, strain engineering will be employed via modified fabrication parameters, and, in particular, by vertical stacking of two such quantum dots. In this way, we want to reduce the emission line width and enhance the yield of emitted light at higher temperatures. In particular, we will demonstrate photoluminescence emission at room temperature from a single quantum dot stack in a photonic crystal resonator. Simultaneously, we will improve the technological implementation of our photonic crystal resonators with the aim of enhanced quality factors which are a measure for enhance photoemission from such resonators. This task will require modifications of existing processing steps and additional steps to improve the perfection of the photonic crystals down to the nanometer regime. To reach the two aims, comprehensive structural and optical experiments are necessary in combination with simulations regarding the structural, electronic and optical properties of photonic crystal resonators with embedded quantum dot light emitters. In particular, transmission electron microscopy will be employed to determine on an atomic scale the local Ge concentration in a quantum dot stack as an input parameter for the simulations. With this project we attempt to expand the limits of Ge quantum dot light emitters into a regime were they will become useful for device applications in the telecommunication spectral range. In the long run, the results will contribute to monolithically integrated, Si-based devices for quantum cryptography or optical on-chip interconnects.

The ever increasing demands for data processing and -transmission become more and more impeded by limited band widths and Ohmic losses of electron transport in conventional semiconductor devices. To overcome these limitations, integrated optoelectronic solutions have recently been developed in the silicon-germanium (SiGe) material system with CMOS technology. But, with an efficient group-IV light emitter still lacking, hybrid approaches combining optoelectronic Si/Ge circuits with group-III-V light emitters are presently pursued. However, full exploitation of monolithic CMOS integration will ultimately require an efficient group-IV light source in the optical telecom range. Over the last years, the Institute of Semiconductor Physics at the Johannes Kepler University in Linz has significantly contributed to the field of group-IV light emitters based on site-controlled SiGe quantum dots (QDs). This preparatory work led to the demonstration of light emission from a single SiGe QD that was monolithically integrated with nm precision in the cavity of a photonic crystal resonator (PhCR). To further enhance the light emission yield, we studied in this project the structural and optical properties of deterministically positioned, vertically stacked SiGe QD pairs. In this approach, the lower QD induces a controlled amount of strain in the near surroundings of the upper QD. It was envisaged that this kind of strain engineering will improve the light emission efficiency, especially at higher temperatures. The layer sequence for the implementation of the QD stacks was derived from simulations of the expected electronic and optical properties as a function of layer geometry and composition. The simulations revealed that the Si spacer layer between the QDs must not exceed a few nm in thickness to suppress spurious light emission from the lower QD. With these design rules systematic series of QD-stack arrays were fabricated by molecular beam epitaxy (MBE), as well as reference samples containing only a single QD layer. Comprehensive experimental assessment showed that the stacked QD arrangement leads to a significant enhancement of the light emission yield at higher temperatures as compared to otherwise identical single-layer QD arrays. Moreover, room-temperature light emission could clearly be observed in SiGe QD-stacks. These promising results certainly require further improvements on the way toward a competitive group-IV light source for monolithic CMOS integration. A next step will combine the stacked-QD approach with controlled defect implantation into the upper QD. In the last years we have demonstrated the benefits of implanted, optically active defects in SiGe QDs. However, deterministic positioning of the implanted QDs, which is a prerequisite of CMOS integration, has not yet been possible. Now, a first feasibility study conducted in the present project showed that the combination of deterministically positioned QD stacks with defect implantation should indeed be possible.