Engineering and technology
- Optical fibre communications
Recently, they have also enabled novel applications such as accelerators for artificial intelligence and LiDAR sensors for autonomous vehicles. These applications benefit from the miniaturization of complex components such as lasers, detectors and modulators, and their high-density integration on a single die offered by typical silicon photonics platforms.
Within imec’s silicon photonics platform, Germanium is intensively used to realize high speed photodetectors, avalanche photodetectors and Franz-Keldysh effect modulators that support data-rates of 50 Gbps and beyond. These devices rely on bulk Ge(Si) material. Even though they support very high data-rates, bulk Ge(Si) modulators have moderate static performance and require a relatively high voltage swing, when compared to silicon-microring-modulators.
It has been shown that exploiting the sharp exitonic absorption peaks in multi-quantum well Ge/SiGe devices can overcome this drawback. Therefore, photodetectors (PDs), electro-absorption modulators (EAMs) and electro-optic modulators (EOMs) exploiting this so-called quantum-confined Stark effect (QCSE) have the potential to overcome the drawbacks of bulk Ge devices and provide wider wavelength tunability, lower losses and higher extinction ratio. Initial results of such devices fabricated in imec’s silicon photonics platform have garnered attention in the optical communication and the optical accelerator community.
In this thesis, the PhD candidate will pioneer next-generation waveguide-coupled Ge multi-quantum well QCSE devices on silicon and explore device concepts for optical communications and artificial intelligence applications. Specifically, the student will investigate devices such as QCSE EAM-PD codesign and QCSE EOM. The work will require material and multi-physics modelling of the multi-quantum well stacks, optimization of these stacks, modelling, design, and characterization of devices and validating through appropriate transmitter-receiver demonstrators for each target application.