Which considerations are required for Periodic Structure simulations?



Hi all,

I read the Lumerical posts about the periodic structures and about the boundary conditions and I understand them. In order to be able to simulate complex periodic structures I decided to start with something simple: a single gold nanoparticle in an homogeneous medium (let’s say air) with anti-symmetric and symmetric boundaries and record the reflectivity and transmission. I expect to see the plasmon resonance of the gold particle in this monitors.

However, I see in the time monitor that has a ripple structure as can be seen in the next figures.

And in the transmission:

My hypothesis is that it requires more time for simulation or to level down the Autoshoff Level but they are at 1000 fs and 1e-7 respectively. It seems energy is being reflected all the time by the gold sphere. So maybe I’m doing something wrong with the simulation region…

How can this be solved? Thank you very much.

Best regards,



Dear @G.W.CastellanosGonz1

Clearly your field remains in the simulation regions, most probably due to poor absorbing PML but it can be for different reasons. Can you please upload your simulation files for a review?


Hi! Sorry for the delay, was really busy last days. I made some changes in that simulation I referred too and a workmate helped me to improve the simulation region. However, I still have some doubts about my simulation. Hereby I upload and if you take a look I would really appreciate that. Thanks!

P.S.: The goal of that simulation is to see if the plasmon resonance can be observed in the reflectivity and transmission monitors

test(3).fsp (2.3 MB)


Dear @G.W.CastellanosGonz1

Your simulation file is set properly. As you said, since you are exciting the plasmonic resonant modes of the gold disk, your simulations takes some time to reach the auto shutoff level. However, once you are waiting long enough, the auto shutoff level becomes low to consider that energy has escaped from the simulation region. Please see the transmission plots below (I am using a slightly coarse mesh):

I guess the ripples are very minimal and you can ignore them. Otherwise increasing the simulation time might improve the results.

There is one more concern about these type of simulations where you are using plane wave source: the periodicity of the simulations. With the current settings you are assuming that your device is repeated many times and its periodicity is defined by the size of the FDTD region (1000nm). This means that changing the FDTD region will change the results. I thought mie scattering example which uses TFSF source will be a good link to learn more about a different approach for these simulations.

Please let me know if you have any other questions.