Problems modelling THz structures using HFSS (version 11)

[lucyszyn]

hfss optical frequency metal

If you are interested in modelling terahertz structures then you may find significant errors in the THz frequency range when using some commercial software packages. For more information, you may found the following paper of interest:

Y. Zhou and S. Lucyszyn, “HFSSÔ modelling anomalies with THz metal-pipe rectangular waveguide structures at room temperature”, EM Academy’s PIERS Online Journal, vol. 5, no. 3, pp. 201-211, Mar. 2009.

This can be downloaded for free from:

**broken link removed**

 

[bitele]

room simulation in hfss

Dear lucyszyn,
thanks for your post.
but i don't think this paper is right because the theory used in the paper is not proven by experiment. as far as we know, thz is between microwave and optical frequency region, so it is not simply anlysised by foregoing theory.

 

[lucyszyn]

hfss terahertz

Dear Bitele,

Thank you for your comment. Unfortunately, I think you may have missed the point of the paper.

No matter what the mm-wave frequency, complete modelling of real structures requires the characterization of both intrinsic AND extrinsic effects. However, unless the intrinsic frequency dispersion model is accurate then there is no hope of developing accurate extrinsic models (e.g. to account for impurities, method of metal deposition, surface roughness, etc.). Fortunately, there are a number of well-understood intrinsic frequency dispersion models. These range from the simple (e.g. skin-effect) to the more complicated but still useful (e.g. Drude) to the very complicated all-encompassing (semi-classical).

This paper was written to specifically address just the intrinsic dispersion anomalies found at low THz frequency at room temperatures only. In this frequency-temperature regime the Drude model has stood the test of time (over a century). Measurement validation for this model has also been around for many decades and so it is considered textbook by physicists. For this reason, the Drude relaxation-effect model was adopted as the reference and the errors found when applying the over-simplified skin-effect model (employed by HFSS, by default, and used by microwave engineers) were reported in the paper.

The maximum frequency in the paper was set to 12 THz. This is because the Drude model, which accurately characterises intra-band transitions, may (depending on the metal) begin to break down with the onset of inter-band transitions above about 30 THz. Such transitions result in frequency resonance effects above around 100 THz. Therefore, to avoid the skirts from such resonances, 12 THz was considered a safe upper frequency limit.

It is hoped that this paper will educate the reader by illustrating the theoretical errors that can be expected by using the over-simplified skin-effect model to design structures up to THz frequencies at room temperature. Moreover, if the more complicated Drude model is adopted, more accurate extrinsic modelling can be developed. Only then can real structures be simulated accurately to predict measured performances.

As a word of caution, the experimental works of Tischer and subsequent modelling work of Wang, which promote the notion of ‘some kind of’ anomalous behaviour with ‘normal’ metals at room temperature, have already been discredited.