Open Collections

Khan, Shaheer

University of British Columbia

Inverse design is a computational method that systematically explores a design space to find optimal device geometries based on specific performance criteria. In silicon photonics, inverse design often results in devices with sensitive design features that are altered significantly during the fabrication process, leading to performance loss. This limits the applicability of these devices for practical use. This thesis demonstrates a novel approach, fabrication-aware inverse design (FAID), to mitigate this issue by integrating lithography models into the inverse design process. FAID optimizes the figure of merit of a lithography-predicted geometry to yield devices that adjust for fabrication-induced performance degradation. The approach is validated by designing and fabricating several devices, demonstrating that FAID significantly reduces performance discrepancies between simulated and fabricated results for both deep-ultraviolet (DUV) lithography and electron-beam lithography (EBL). The integration of global and hybrid optimization techniques further enhances the ability to explore the design space effectively, enabling designers to discover performance bounds for real-world applications. The results highlight FAID’s potential to enable inverse design to generate high-performance photonic devices that are viable for practical and commercial use in scalable silicon photonic fabrication.

Text

2024-12-10

Attribution-NonCommercial-NoDerivatives 4.0 International

http://hdl.handle.net/2429/89820

Electrical and Computer Engineering

University of British Columbia

UBCV

http://creativecommons.org/licenses/by-nc-nd/4.0/