Si-Ge Quantum Dot Laser: What Can We Learn From III-V Experience?

Abstract

Si-Ge system offers a significant extension to traditional Si-based microelectronics. However, applications of the system would be further greatly expanded if it can be used for high-speed optical transmitters and interconnects. A straightforward idea to achieve this goal is to use the device designs, which are already successfully applied for direct- gap semiconductor materials, particularly, to III–V materials. These are diode lasers based on double heterostructures [1] and on heterostructures with reduced dimensionality [3], so called, quantum wells (QWs), quantum wires (QWWs) and quantum dots (QDs). An idea to achieve lasing in indirect gap materials by using double hetrostructure concept was first mentioned by Kroemer in 1963 [2]. In his paper H. Kroemer proposed to use the double heterostructures (DHS) for carrier confinement in the active region of the diode laser and wrote that “laser action should be obtainable in many of the indirect gap semiconductors and improved in the direct gap ones, if it is possible to supply them with a pair of heterojunction injectors”. Attempts to achieve lasing SiGe-Si DHSs and QWs did not result is significant success, however, as also in the case of other types of indirect-gap materials, for example, AlGaAs DHSs with high Al content (x>0.5), or in type-II GaAs-AlAs quantum QWs. A different approach to achieve lasing in semiconductors was first proposed by Basov, Vul and Popov in 1959 [4], who considered unipolar carrier injection. Population inversion between ionised impurities and free carriers was thought as a gain mechanism through impurity ionisation upon application of pulsed electric field. Boundaries of the sample providing the reflection of light were proposed for a laser feedback mechanism. For Si-based optoelectronics such an opportunity is particularly important, because optical transitions in the latter case are linked only to one band and the problem of indirect crystal band structure in silicon is lifted. In 1971 an extension of the unipolar laser approach was proposed by Kazarinov and Suris [5]. The authors proposed to use population inversion between different electron subbands in a specially designed QW superlattice. The laser based on such approach (cascade laser) was realised in 1985 by Faist et al. [6]. The success of the cascade laser is linked, however, to direct-gap III–V materials and not to Si-based systems, in spite of the fact that the hystory of intraband lasing in indirect gap materials (e.g. in p-doped Ge) is quite long [7].