Frequency-domain solvers are used extensively
for modeling arbitrary metal-based terahertz structures. Four
well-known commercially available electromagnetic (EM) modeling
software packages include HFSS™, CST Microwave
Studio®, EMPro, and RSoft. However, there are a number of
operational issues that relate to how they can be used to obtain
more meaningful and accurate results. Even experienced users
of these and similar software packages may not fully appreciate
some of the subtle ambiguities in defining boundaries and material
parameters for use in THz applications. To this end, a detailed
comparative study has been undertaken, in consultation with all
four vendors. First, in order to avoid introducing ambiguities,
frequency dispersion in materials has to be clearly defined from
first principles; in both intrinsic and effective forms. Different
frequency dispersion models are then introduced for ‘metal-like’
materials. To act as benchmark structures, conventional air-filled
metal-pipe rectangular waveguides, associated cavity resonators
and a spoof surface plasmon waveguide have been simulated,
using a raft of different approaches; with a view to illustrating
quantifiable weaknesses in commercial software packages for
simulating arbitrary metal-based THz structures. This paper
highlights intuitive and logical approaches that give incorrect
results and, where possible, makes recommendations for the
most appropriate solutions that have hitherto not been given in
Technical Notes.