Optical Activity of Group III-V Quantum Dots Directly Embedded in Silicon

Abstract: Optically active III-V group semiconductor quantum dots (QDs) are the leading element of the upcoming safe quantum communication. However, the entire electronic and IT infrastructure relies on silicon-based devices, with silicon also providing a natural platform for photonic integration. Combining semiconductor optics with silicon electronics is thus a major technological challenge. This obstacle cannot be directly solved because silicon is optically inactive. Interfacing III-V quantum dots with silicon is thus a sought-after solution. A radical approach is to embed III-V material grains directly into silicon. The first realization of such technology was developed, and it gave InAs and core-shell InAs/GaAs QDs embedded in Si with bright and narrow single-QD emission lines. No theory has been given, though, and, as we show here, it is not even obvious if and how such QDs can be optically active. We first use general arguments, also supported by atomistic calculations, that InAs/Si QDs cannot confine both carrier types unless the structural strain is mostly relaxed, meaning many defects at the interface. This explains the lack of light emission from those dots. Then we show that the InAs/GaAs/Si QDs can confine both carrier types. Their electron states are, however, highly influenced by $k$-space valley mixing, which impacts emission spectra and deteriorates optical properties. We propose to overcome this by adding an additional wider-bandgap material layer.