A Novel Approach To Constructing The Green's Function for Layered Media and Its Application to MMIC, RFIC, and EMC Problems

摘要

Green's functions for layered media play an important role in RF and microwave circuit applications and they are time-consuming to construct because of the tedious nature of the Sommerfield Integrals. Several techniques for expediting the construction of the layered-medium Green's functions have been proposed in the literature [1,2], notable among them is the closed-form Green's function approach [2]. However, these techniques are still not as efficient as one would desire, especially when the source and observation points are not strictly located in the same plane. Furthermore, if the computational domain size is large and the observation points are many wavelengths away, the accuracy of the closed-form Green's function is known to suffer. In this work we present a novel approach to constructing the layered medium Green's function using the BOR (body of revolution) version of the FDTD, which has several desirable attributes: (a) It poses no difficulty when handling n-layer problems, even when n is large; (b) it is easy to analyze lossy layers, including RAM materials; (c) the location of the observation point can be arbitrary in terms of horizontal and vertical distances from the source; (d) the electric and magnetic fields (all six components) are computed directly at all observation points in the computational domain and, hence, they can be conveniently used to compute MoM matrix; (e) Surface wave contributions, which play an important role at large distances from the source, can be obtained without any difficulty; (f) the dielectric layers can be truncated at an arbitrary radial distance from the source to estimate truncation effects and, (g) the mutual interaction between two sources (e.g., antennas) can be obtained by applying the Reaction concept once the field due to the first source at the location of the second one has been obtained. The Green's function results are then combined with superposition to construct the fields due to a distributed source.