Learning Forum

[JakeConnors]

I am interested in simulating a laterally periodic (i.e. periodic in a plane usually perpendicular to a wave propagation direction) at a range of frequencies, from when the typical length scale of the periodicity is both electrically small and electrically large. The notion is that for these laterally periodic structures, scattering will occur beginning at some wavelength comparable to the pitch (or pitch ~ lambda/2).

I have attempted previously to do this using a model (picture attached) with master and slave boundary conditions on the lateral edges, and Floquets ports on the top and bottom surfaces. When looking at the S-parameters for such a model, the scattering that one would expect is not present even at very high frequencies where you would expect the structure to be multi-moded. Including extra modes in the Floquet port does not help as these modes do not have the same periodicity as the structure and the coupling to them from the TE00 mode is basically zero up to numerical precision.

What I have now tried is to replace the Floquet ports with a plane wave excitation on one side and a PML layer on the opposite side, but have not been able to figure out how to plot the scattered wave or extract any useful information from such a model. Any suggestions as to how to approach such a calculation or suggestions of a different modelling approach entirely would be greatly appreciated.