Aspects of Integrated Amplification for Silicon Photonics

Abstract

The exponential rise in global data traffic and the growing reliance on digital applications is pushing the bandwidth demands within data centers. The traditional hierarchical network architecture of data centers, primarily relying on electrical interconnects, faces scalability challenges including power dissipation, bandwidth limitations, and cooling requirements. Optical interconnects, using fibers and pluggable transceivers, emerge as a promising solution to these challenges, offering advantages such as electromagnetic interference resistance, high bandwidth, and efficient energy usage. This thesis explores the design, fabrication, and characterization of opto-electronic devices to be used as components for optical transceivers on a silicon photonics platform, which leverages the mature complementary metal-oxide semiconductor fabrication technology. Chapter 2 introduces the basics of waveguide theory alongside the principles behind defect-based avalanche photodiodes, phototransistors, and two-level system optical amplifiers. Chapter 3 details the design, simulation, and characterization of a high-responsivity silicon/germanium phototransistor, achieving over 1000 A/W in performance. Chapter 4 discusses the design and measurement of an all-silicon avalanche photodetector for near-infrared wavelengths mediated by deep-level defects. In Chapter 5, the focus shifts to enhancing the previously mentioned photodetector’s responsivity through slow light gain with subwavelength grating waveguide structures, with details on its design, simulation, fabrication, and characterization. Chapter 6 explores the development and analysis of an erbium-doped waveguide amplifier on a hybrid silicon nitride - tellurite platform, incorporating erbium ions via ion implantation. This thesis makes contributions toward realizing efficient silicon photonics-based data communication infrastructure, supporting the escalating demand for bandwidth while mitigating power consumption and improving system scalability.