New strategies for dislocation density reduction in monolithic III/V epitaxy on Si – FILTER

The combination of the Silicon technology with III-V semiconductors is expected to have a major impact for the realization of high-volume / low-cost integrated circuits. Photonic chips based on the already mature Silicon photonics capabilities together with integrated III-V lasers or photodetectors will soon be the foreground for optical-interconnects or lab-on-chip sensors. While heterogeneous integration - where the III-V device or material is bonded to the Si circuit - already showed promising results, a more direct integration scheme is highly desirable in order to improve the yield, the integration density and to decrease the cost of devices.
In this context, the monolithic integration by epitaxial growth of high-quality III-V heterostructures directly on Si has been intensively pursued in the past decade. The main challenge remains a drastic reduction of the threading dislocation density (TDD). These line defects are created at the III-V/Si interface due to the large lattice-mismatch and propagate through the epilayers, severely degrading the device performance and lifetime. Despite a considerable amount of work over the past years, TDDs in the range of 1E9 to 1E10 /cm2 are still generally present after growth of about 1 µm of III-V material. We propose to explore a radically new design and growth strategy with the main objective to reduce TDDs down to below 1E5 /cm2.
The groundbreaking concept in this project is to initiate in a controlled manner the interaction between on the one hand threading dislocations and, on the other hand, other types of extended defects such as anti-phase boundaries or misfit dislocation arrays at interfaces between specially designed interlayers. These interlayers can be inserted directly at the III-V/Si interface or inside the III-V buffer layer and should act as sink for the threading defects.
To this end, MBE growth strategies and new structure designs will be proposed by the French group thanks to an extensive study by advanced transmission electron microscopy techniques of dedicated samples carried out by the German group. The progress made on the TDD reduction will be assessed by the fabrication and testing of lasers demonstrating the impact of the proposed project. The III-V material used here will be GaSb, which can readily serve as a starting point for many optoelectronic devices in the mid- to far-infrared. Even so, the filter design rules derived on the basis of detailed microstructure analysis as well as the growth techniques developed throughout the project are expected to be applicable to other compound semiconductor families and have consequently a major impact on a wide range of applications from data/tele-communication to sensing among others. Furthermore, the developed methodology of dynamic microscopy in three dimensions is in itself an important step towards a complete and efficient determination of the structure-function relationship which can be applied to many materials combinations.