De-Embedding Method for Strongly Coupled Cavities

The well-known cavity perturbation method, in which a cavity resonator is partially filled with a material sample is used to determine material parameters (permittivity, electrical conductivity, and permeability) at microwave frequencies. These parameters are inferred from the differences in resonant behavior between the empty cavity and the material-loaded cavity. When applying the method, one has to couple the cavity to external measuring equipment (e.g., a vector network analyzer) to obtain the S-parameters of the cavity. The insertion of a coupling device (e.g., a thin stub) detunes the cavity. If the cavity is loaded with a large sample or if one has no access to the empty cavity-both cases may occur in in-process monitoring applications-one must remove the detuning effect from measured S-parameters to produce accurate estimates of the sample material parameters. Several methods for this de-embedding have been proposed in the literature. They work well for loose or not-too strong coupling, but less so for strong coupling, which may occur in in-process monitoring situations. We have examined the known de-embedding procedures with respect to their ability for coping with strong coupling by thin stubs and propose a method that works in this case.