Material fit



i’m trying to fit the sampled data with FDTD model. i’ve tested the customary ways related to adjusting the data but unfortunately resulted in failure. would you please check it out or give some useful recommendations about modifying the results?
here is my structure (in this structure the real part is not important so i set it to zero for all range of frequencies.)

impatiently waiting for your response.


Gomez.Far.fsp (411.6 KB)


Dear @s_eskand

Since the real part of the conductivity is set to zero and you have a broadband frequency range, fitting material data will be very challenging. You can read more about this scenario here:

If you know the real part of the conductivity, I strongly recommend you to import them in sampled data (rather than setting them to be zero). Please note that material fit tries to fit both real and imaginary part of the data:

If you still couldn’t get a good fit, one workaround is to run simulations for narrower bandwidth or even single frequencies. Also, if you still want to set the real part of the conductivity to zero, it will be a good idea to set the it to a small value (like 1e-6) rather than zero.

I hope this was helpful.


Thanks bkhanaliloo. i was very useful. i did as you say. i ran the simulation for a single frequency and set the real part to 1e-6. the result got better, but there’s still problem. before that, let me just give you a general description of my work.the proposed structure is an infinite sheet of graphene which acts as a switch, something like the following structure:
the operating frequency (the frequency used in the simulation) must result in ON state, expecting the profile of the E-field to be something like the following picture:

in the ON state, as you are aware, we expect to observe the resonances of the surface waves when propagated on the surface of the graphene sheet as the above picture indicates. but my result is this:

which shows no maxima and minima and thus it is not a surface wave.
would you please help me solve this problem?
for better help, i attached the simulation file:
graphene_waveguide_switch.fsp (731.3 KB)

thank you for your attention to this matter.


Dear @s_eskand

In the simulation file you sent me, you have a graphene layer with periodic FDTD BCs. Since this is an infinite graphene layer, what do you expect for the modes? Once I am using a mode source, the injected mode is not confined and it does not make much of sense. Below is the screenshot of the mode profile in the source:

Regarding the material fit, I noticed that it fails to give a proper fit if your source is single frequency. Instead, I tried to use a narrowband pulse (say 0.1 THz bandwidth) and it could give me a proper fit. Also, I modified the simulation file in KB page and ran the simulations with source frequency of 10 THz. The major change was to increase FDTD region so that I can capture a few oscillations. Here are the results with the simulation file:

graphene_waveguide_switch_10THz.fsp (3.8 MB)

Please let me know if you have further questions.


I think if you want to simulate an infinite graphene sheet, you can use a 2D simulations and set all the BCs to PML. Ultimately you expect that injected mode will be a slab mode confined inside graphene.



Dear bkhanaliloo.
i really don’t know how to appreciate your cooperation. i’m very proud to be one of your students.
as it seems, i’d better tell you the whole story of my work:
i want to observe the switching effect in an infinite sheet of graphene using magnetic biasing. with the presence of magnetic field, the surface conductivity of the graphene is changed drastically which is not defined in Lumerical yet. but as your colleague, fgomez, said, i can calculate the surface conductivity for a bunch of magnetic fields and frequencies and import them using 2D sampled data material. i followed the same procedure but still couldn’t get the desired results.
anyhow, the proposed structure is nothing but just a theoretical graphene-based switch. the infinite sheet of graphene is surrounded by air with no substrate, that’s why i disabled substrate, graphene middle and guide. the following picture is the imaginary part of the surface conductivity when subjected to a dc magnetic field (3 Tesla) as a function of frequency:

when the image sign is positive, tightly confinement of TM mode is expected to propagate, representing ON state. in this case, i’m expected to observe the oscillation of surface waves propagating along the surface of the graphene. on the contrary, when the image sign is negative, poor confinement of TE mode is expected to propagate, representing OFF state. in this case, i’m expected to observe the weak oscillation of surface waves propagating along the surface of the graphene. that is a theoretical background of my work.
following your suggestion, i set the bandwidth and frequency to 0.1 THz and 3 THz, respectively, resulting in a perfect fit. furthermore, i ran the 2D simulation and set the BCs to PML, but nothing is observed. i even increase the FDTD region in the hope of capturing a few oscillations but again resulted in failure.
i attached the text file of real and imaginary part of the surface conductivity (above picture) imported in the software, together with the updated simulation file.
3T.txt (12.2 KB)
magnetic biasing.fsp (743.6 KB)

i would be very glad if you continue helping me.

thank you again for all you’ve done for me.


Dear @s_eskand

Thanks for the detailed explanation of the problem. It was quite helpful.

I modified the simulation file as we planned already:

magnetic biasing_2D.fsp (314.9 KB)

I had one more inquiry. Do you expect that the “magnetically modified graphene” material to support a mode at this wavelength? When I tried to inject a mode source, I couldn’t see any confined mode, and I suspect maybe because the material is in its “off” state for this particular wavelength.

For next steps, I guess a reference showing a supported mode will be quite useful. If the goal of project is to find these surface modes, maybe you can change the source wavelength and look for the modes at other wavelengths.