left-handed materials simulations using CST MWS

Hi,
I was trying to reproduce the simulation results of T. Weiland et al. in the paper
T weiland et al. J. appl. Phys. Volume 90 Number 10 2001.

The authors try to simulate a split ring resonator (SRR) in MWS. I find that the results of the resonance frequency of the SRR are not very convincing. The authors place the SRR in a computational cell where they do not specify the dimensions of the latter. They only hint the boundary conditions!

I started with some rough cell size and found that the results are pretty much dependent on the computational cell size. If you change the later, you get some shifted results.

My question is:

If you have such a problem at hand, which cell size do you choose to obtain an accurate result? lamda, lamda/2,...?

What kind of surrounding space is good? Do we have to find it by trial and error?

Background properties:
->Surrounding space

lower x upper x (open boundary)
--- ---
lower y upper y (electric boundary conditions)
--- ---
lower z upper z (magnetic boundary conditions)
--- ---


Thanks for any insights.

Hello em_solver,

I'm simulating similar unit cells, can you give more details or post your model here?

Regards,
P.

As I mentioned, I'm first trying to reproduce Weiland's results and make sure I got it correct. As for my own stuff, it's not published yet so I can't just place it online.

I have some experimental results that I want to double check/confirm with MWS. But sorry I can't give you hints on this one. Let's stick to Weiland's model for the time being.

That's OK,

Can you post the article? I'll perform my own analysis and see what comes out.

P.

Yes please post the article or even an example project of the article in CST.

OK here are the MWS basic model files! Thanks.

Here goes the article:

I can reproduce fig 6 (left) [Split ring without the rod] but slightly shifted resonance frequency.

I believe that it's dependent on the computational cell size.

But I cannot reproduce fig 7 (left) [Split ring with rod]. And it looks like the graph has been truncated in the paper.

Thanks for the help.

Hello,

I suggest trying "Adaptive Meshing -> Energy based" - Because this will refine your mesh in points where the energy distribution is very high.

Additional it compares the deviation of the S-parameter between two passes and will stop when the changes are less than 2% (default value) depending on the discretisation.

Maybe it can optimise the structure to get a reasonable resonance frequency.

Greetings,

Frank

Good point. It might work for the figure 6 graph as the deviation I got to the original value is small. But as for figure 7 graph, I'm far from it so I doubt it works in this case.

MWS seems to be very good for detecting decrease in transmission peaks but for increase in transmission, I would say it's pretty tricky.

I even asked the authors for the parameters they used but I never got a reply from them. Pff...

Hey Pushhead, did you have any luck so far with the model provided? I could get this in HFSS but as for MWS, I'm still stuck. Did anybody got luckier and want to post back the improved model?

At last I finally cracked the puzzle! It's harder with MWS but trying hard enough and you can get the hang on it! :D

I'm curious - can you please explain us the "magic"?.

I tried the modell several times with slight modifications but I was unable to the get the resonance frequency given by the paper.

I am a little bit curious, can meta materials be modelled in standard software packages such as microwave studio or HFSS?

Hello,

Any hint about how to solve out the problem of simulation?