Learning Forum

[sagar]

I am trying to generate effective mode area data by sweeping the radius of the pillar but with PML boundary condition is unable to generate data set while with metal BC it is working fine. while by varying radius values manually I am able to generate mode data with PML.

[Guilin Sun]

This is mainly because the number of trial modes are small when you use PML BCs. Please increase it and try again.
In Metal BCs it can easily find modes enclosed with metal box, whereas in PML NCs there are many PML modes which need to be removed. When the calculated modes reaches a certain number, it stops even though the reported number of modes is fewer than expected.
To test, please choose a specific sweep parameter, and then manually calculate the modes with PML BCs. you will find that it may not find the desired mode if the number of trial modes is small. Please increase it till you can find the modes.

[sagar]

I put the number of trial mode is 20, but all the time it is giving only two modes for lower diameter value and for higher diameter it is giving around four to five trial modes. Where with metal BC all 20 modes are forming. But as I am selecting only first two modes, in diameter sweep it should come.

[Guilin Sun]

I guess you know that the number of reported modes is fewer using PML than Metal BCs as I explained previously.

[sagar]

Yes the number of reported modes will be fewer in case of PML, that I have understood. So I have checked the diameter value from where it is able to generate the modes before the parameter sweep. From that cutoff value I did the diameter sweep and chose first two modes those are forming for all the values of diameter of the waveguide. So corresponding parameters like effective area, ng , neff all there value should come in the result. But I am getting this error.

[Guilin Sun]

Did you tried the file with that sweep parameters separately ? if so, please increase the number of trial modes again, as the random seeds may create different PML modes.

[sagar]

I did the simulation by putting mesh around the waveguide and with PML BC and did the radius sweep, and it is working fine. I calculated effective mode area, neef, and ng there are not much changes with PML BC vs Metal BC. The only slight fluctuation is between 80 to 150 nm is observed in the case of PML BC even there is a slight deviation for mode 1 and mode2, whereas in the case of the metal curve is very smooth and selected mode 1 and 2 are giving same data.
What exactly happening with metal? As suggested on the Lumerical website metal BC conditions are preferable to PML.

[Guilin Sun]

I would suggest that you separate/isolate the radius which gives you suspicious/undesired result. In particular check the mode to see if it is consistent with the previous one, and neff is consitent.