How to simulate finite size EBG surfaces in HFSS

Hi folks,

I wonder if anyone can shed some lights on this problem: how to simulate a FINITE size (in 2D X-Y plane) EBG surface in HFSS? Specifically, EBG formed by patches without via.

My previous simulation setup was using wave port on the top surface of the air box and radiation boundary around the air box. But the resulting reflection phase didn't make sense. I've also tried using incident plane wave, but it didn't provide reflection phase in the result.

Thanks for your help in advance.

Welcome joeacmilan,

Firstly, do you mean finite in two dimensions, or just one? I'm guessing from your setup that you're studying normal incidence reflection. Have you de-embedded the wave port to the surface of the EBG? You CAN get phase information from an incident plane wave, but it would be a lot of manual processing - better to stick with the wave port.

Hi, Thanks for the reply. Yes, it's finite in 2D and Im looking at the normal incident direction. I did de-embed it but the the reflected phase was bizarre. I am suspecting that the wave port or the boundary condition is not proper for this simulation. But I have no clue so far...

Does the waveport cover the entire top face? Placing wave ports beside radiation boundaries in these types of setups can cause strange field patterns to emerge from the port. Have you verified that the fields at the port are essentially plane waves? And out of curiosity, what is the size of your surface in unit-cells?

There's no overlap btw port and rad boundary. What kind of setup did you use in terms of boundary condition and port?

For normal incidence, I've always just simulated an infinite array. However, I do know from other experience that if you have a waveport touching orthogonal radiation boundaries (overlapping on the same surface is fine), strange effects will occur in the "corners".

I cannot use infinite array concept for my simulation as it is not fully periodic. that's why I m building a finite simulation model. I don't see any way to avoid touching of the top port and side rad boundary. Any suggestion?

1) Place a large empty "buffer region" around your EBG array, such that the array is electrically far from the side walls of the computational domain. Then, change the walls from radiation to a pair of perfect E's and perfect H's. This will allow for a perfect plane wave to be supported, while hopefully not disturbing your results too much.

OR

2) Shrink the size of the waveport (not the surface), and put a radiation on the top as well (such that they overlap). A medium sized waveport may emit better plane waves than a full-sized one for the reasons I stated in my last post.

OR

3) Create an empty surface right above the EBGs. Simulate using an incident plane wave, and do the math manually on the surface to find the phase (using Incident/Total fields). This might involve exporting to MATLAB, writing some scripts, etc.