Diverging issue in graphene

I am going to simulate the transmission of graphene on photonic crystals. In the higher value of graphene chemical potential (above 0.3ev) the simulation runs with no problem but in the chemical potentials below 0.3 ev it is stopped due to diverging problem.
I already have changed all boundary conditions to metal but still faced with the same problem. Reducing the dt stability time even to 0.1 also did not solve the issue. I also tried to change the pml profile and number of PML layers but that also was not helpful. I would be thankful if you suggest me a solution.
My file is uploaded here.

Dear @Leili.AbdollahiShira

Thank you for providing me with your troubleshooting steps. It was very useful.

I could not see the simulation file but based on your descriptions, the problem seems to be due to material fitting. There was a similar case to this which I discussed how we can solve this type of divergence problem:

Can you please take a look at it and let me know if you still had a problem.


Thank you for your reply. Yes, indeed this solved the diverging problem.
But still I do not get the correct absorption. eg if I run the structure with graphene with chemical potential of 0.7ev I expect to obtain much less absorption than graphene with 0.1 ev chemical potential. The differences in my simulation are very small. graphene on PhC.fsp (629.4 KB)
Again I upload my file. Hope this time you can see it. I would appreciate if you help me to figure out the problem.

Dear @Leili.AbdollahiShira

Yes, you are right. I didn’t see much difference between the two cases.

I had a few comments about your simulation file: I think it will be a good idea to increase the FDTD region. At the moment the distance between waveguide and PML in the x-direction is only 200nm, and PML might affect the waveguide modes. A good way to choose the gap distance it to plot the mode source in the log scale and make sure that waveguide mode has decayed enough (few order of magnitude) before reaching the PML.

Moreover, do you have any reference that shows the results will change with the current geometry? I am not quite sure if this layer is large enough to modify the wavguide and transmission results. Please keep me updated and I will be glad to be of a help.


Thank you for your note.

I tried different thing and large PML-structure distance as well, but did not get that much improvement.
I see lumerical has already simulated graphene modulator (https://kb.lumerical.com/en/other_application_electro-optic_modulator_using_.html).
I would like to ask the fsp file of this graphene modulator. In principle it is the same as mine. So I can get idea for my own device.

I don’t have a reference but on the ring, very high extinction ratio has been reported. So on the PhC also we should see the same effect.


Dear @Leili.AbdollahiShira

Thanks for the update and applying my recommendations.

I guess it is a very good idea to compare the FDTD results with the results of the KB example on electron optical modulator. To create the FDTD file, simply select the geometry and mesh override in Mode, and then copy (ctrl + c) - paste (ctrl - v) them into FDTD. You can then add FDTD simulation region, monitors, and sources to capture transmission.

Another good step to try will be to replicate the results of the ring resonator case. In your geometry, I am not quite sure how good the PC cavity is. Moreover, the graphene layer is small and is located outside the cavity. However, in the ring case, light will circumvent inside many times before escaping, and will interact with the graphene layer in each cycle. This means that even if there is a small difference between two types of graphene, the effect will be enhanced within thousand or millions of cycle (depending on the quality of the ring resonator) inside the ring. It looks to me that this is not the case in your current PC cavity geometry.

Please let me know of your thoughts, or if you had any questions.