Date of Award
Doctor of Philosophy (PhD)
Electrical, Computer & Energy Engineering
The high index contrast of the silicon - silicon dioxide material system allows for dense integration of optical waveguide devices. Possible applications include intra-chip, inter-chip and fiber optic interconnection systems. Optical intra-chip interconnections become more desirable as the complementary metal-oxide-semiconductor (CMOS) circuit density puts ever tighter constraint on on-chip interconnection performance. Board level, rack level and rack-to-rack data center interconnections are ever more constrained by space and bandwidth to which silicon photonic modules may offer an improvement. As fiber optic systems serve smaller and smaller area systems, integrated switching systems that are enabled by silicon photonic devices involving wavelength division multiplexing (WDM) become more desirable.
In this thesis, we firstly take a brief review of the development history of information technology, optical communication and silicon photonics. Secondly we examine the optical performance of an array of photonic devices which are the basic building blocks for silicon photonic circuits. Thirdly we turn the attention to the fabrication related issues. Silicon photonic circuits are prone to the thermal and fabrication induced process variations. We discover the process variation exhibits a “random walk” pattern with spatial extent at wafer scale. Fourthly we propose a simple method to extract fundamental parameters out of fabricated silicon photonic devices. Based on the systemic wafer-scale measurement results, our method combines the advantage of both numerical simulation and simple analytical modeling techniques. Lastly, we propose a variation-aware on-chip interconnect design for multi-core processors. This design adapts to on-chip thermal and process variation effects, pointing to the improvement of wafer-scale fabrication yield and interconnect network communication throughput.
Chen, Xi, "Process Variation in Silicon Photonic Devices" (2013). Electrical, Computer & Energy Engineering Graduate Theses & Dissertations. 72.