Numerical techniques for electromagnetic applications in microelectronic and radar imaging systems

In this thesis, the application of numerical techniques to electromagnetic problems in microelectronic and radar imaging systems are investigated. In particular the following problems are studied: (1) Dielectric rib waveguide discontinuities are analyzed with the Finite Difference Time Domain (FDTD) method. The application of Berenger's Perfectly Matched Layer to multi-layered dielectrics is analyzed and the specific conditions needed to successfully match the multiple dielectric layers are determined and justified. An FDTD method to find the fundamental mode's spatial distribution is used to excite the discontinuity problem. It is shown that the computational domain can be reduced by twenty percent over Gaussian excitations. The effects of rib waveguide bend discontinuities and the effects of the rib geometry to the bend loss are presented. (2) An Impedance Boundary Condition (IBC) for two dimensional FDTD simulations containing thin, good conductor sheets is developed. The IBC uses a recursive convolution scheme based on approximating the conductor's impedance as a sum of exponentials. The effects of FDTD parameters such as grid size and time step on simulation accuracy are presented. The IBC is shown to accurately model the conductor loss over a wide frequency range. The verification is performed by comparing the quality factors of rectangular resonant structures determined by the FDTD simulation and analytical methods. (3) Phase unwrapping techniques for the inversion of terrain height using Synthetic Aperture Radar Interferometry (InSAR) data are analyzed. The weighted least squares and branch cut phase unwrapping techniques are specifically studied. An optimal branch cut method and a hybrid least squares/branch cut method are presented and used to unwrap the phase of both simulated and real SAR interferograms. When used to invert terrain height, these new SAR phase unwrapping methods offer over fifty percent reduction in root mean square (rms) height error compared to the straight least squares method and over thirty-five percent reduction in rms height error compared to the weighted least squares method based on coherence data weighting schemes. Thesis Supervisor: Professor Jin Au Kong Thesis Supervisor: Dr. Y. Eric Yang 19971208 001 Numerical Techniques for Electromagnetic Applications in Microelectronic and Radar Imaging Systems by Jerome J. Akerson B.S. Electrical and Computer Engineering, University of California, Davis, 1984 M.S. Electrical Engineering and Computer Science, Massachusetts Institute of Technology, Cambridge, 1990 Submitted to the Department of Electrical Engineering and Computer Science in partial fulfillment of the requirements for the degree of Doctor of Philosophy at the Massachusetts Institute of TechnologyFebruary 1998 © Jerome J. Akerson, 1998. All rights reserved. The author hereby grants to MIT permission to reproduce and distribute publicly paper and electronic copies of this thesis document in whole or in part, and to grant others the right to do so.