Convergence of field


#1

I am simulating for transmission and electric field in nanohole array. Simulation are running smoothly for small period say 400 nm or 600 nm. but for longer period like 800 nm field is not converged and ripples appear in field spectrum. What could be the possible cause for the problem?


#2

Hi,

Can you share the data so that the cause of the problem can be narrowed?

Regards,

Vivek


#3

Electric field data.pdf (174.5 KB)


#4

@eez158418

The problem could be that for the larger periods, the structure is supporting diffractive grating orders that are travelling at a steep angle. In this case, you should ensure that there is at least half a wavelength between the structure and the PML boundaries above and below the structure. You should also set the PML profile to “steep angle”:

And you can also try increasing the simulation time to make sure that the fields have enough time to decay before the maximum simulation time is reached.

Please try it out and let me know if you are able to resolve the problem.


#5

Thanks. I am trying.

I have one more query. I am running nested sweep using script. When simulation ends the field monitor shows field profile only for last sweep parameter. I want to save the field profile data for all wavelength and at each parameter swept. Could you please tell me that how can i save that data?


Collecting the field profile from a sweep
#6

I have split this off into a new topic and the reply is here:


#7

I tried it out but it did not resolve the problem. I am using metal in my structure and using conformal Mesh 1. If i use conformal mesh 0, the field was stabilized. can conformal mesh 0 also be used while using metals? if yes, how to make sure that it does not affect the results.


#8

If you are seeing ripples in the spectrum when using conformal variant 0, but you don’t see the ripples when using conformal variant 0, then that could indicate that the mesh is too coarse for the conformal variant 1 mesh refinement method to work well. You could try reducing the mesh step size.

For both conformal variant 0 (where staircasing is applied at metal surfaces) and conformal variant 1 (where conformal meshing is applied at metal surfaces), the results will eventually converge to the same result as you decrease the mesh step size around the surface of the metal, so either mesh refinement method could be used. An example is shown in Figure 4 here:
https://www.lumerical.com/support/whitepaper/fdtd_conformal_mesh_whitepaper.html

Figure 4 shows that results approach the same value, but a coarser mesh can be used to achieve the same accuracy level of accuracy when conformal meshing is applied.

Sometimes applying conformal meshing leads to artificial modes at the surface of the metal if the mesh is too coarse, and this could lead to extra peaks in the spectrum, so I suspect that may be what you are seeing. By performing convergence testing of the mesh step size using either conformal variant 0 or conformal variant 1, you can confirm that the results are accurate.

Hopefully this helps.