Terahertz Polarizer with Graphene



I am trying to achieve the results of the following paper:
07305209.pdf (556.7 KB)

I am first trying to see the simple case of one layer graphene. Now, to make the calculations faster I am using the 2D model of graphene which gives me a good fit for 0.1 - 5 THz range.

Now, when I go onto simulate the structure, I am getting a transmission plot like this:

The authors in the paper have plotted the Transmission in dB, but still I believe my simulation are far from those the author reported.

Can you please look into my simulation file and tell me where I am going wrong and what do I need to correct?

My simulation file: graphene-polarizer.fsp (913.4 KB)

Does Lumerical provide a way, in which I can plot the transmission in dB, as the authors did in the paper?



Any help anyone?


I used a plane wave source.
I changed the fdtd span to 3*3 micron.
The simulation file is here: graphene-polarizer-mod.fsp (256.3 KB)


Hi @sourangsu.banerji and @m.dezyani,

Thanks for the improved simulation file @m.dezyani! You are right, here we should use the regular plane waves source and set the x span and y span of the simulation region to match the period of the unit cell.

I just wanted to add a couple of suggestions (see graphene-polarizer-modFG.fsp (262.9 KB)):

  • To create the slot you can use a rectangle of etch material (mesh order 1) right on top of graphene. In this way you do not create a hole in the silicon substrate, just in graphene.

  • To get more accurate result I would refine the mesh in the x and y directions for the slot.

  • To get the plot in dB you can use the following script lines:

    Tres = getresult(“Transmission_2”, “T”);
    plot(Tres.f/1e12, 10*log10(abs(Tres.T)),“f (THz)”, “Transmission (dB)”);


Thanks you both @fgomez and @m.dezyani I appreciate both of your help but there was a reason as to why I was using a mode source. I don’t know fully if it was correct or not but I will explain as to why.

In the paper they said that for a y polarization, I would get the curve similar to the one, you have plotted. And for x polarization, I would get a nearly straight line.

I am attaching the graph from the paper:

Now from this post, I got the idea about how I can change my polarization for a plane wave:

But I am getting actually opposite results for both x and y polarization.


This is my modified file:
graphene-polarizer-modFG_ver1.fsp (265.0 KB)

Thanks once again for all your help.


Any help on this issue guys! I am really struggling on this issue for a long time and cannot seem to figure out why am I having the drop for x-polarization when the paper reports the drop in transmission for y polarization and no drop for x polarization.

Thanks in advance!


Hi @sourangsu.banerji,

Sorry for the long wait. I think the problem is a confusion with the orientation of the coordinate system. Your simulated structure is rotated 90 degrees with respect to the diagram in the plot from the paper: the long side of the slot is aligned with the y axis in your simulation, while in the paper’s diagram it is aligned with the x axis. Therefore, the polarization definitions are just inverted.

In your latest file, I noted that you used PEC for the “2d slot”. If the slots are simply holes in the graphene layer, I think you should the approach I described in my previous answer:


i am trying to learn FDTD software. I used your simulation and tryed to simulate it. but my time of simulation is too long. Can you help me?3.fsp (256.1 KB)


I change source frequency.
Attach file:3-modified.fsp (251.1 KB)


thank you for your reply. I am wondering why you have used mesh1 and mesh2? and how did you find out the location of the meshes?
I would really appreciate if you could help me.


The interface between to material is the best location for using override mesh.