Inteference due to glass thickness



I am trying to get rid of interference from my simulation. I followed this post

and extended my pml from +z direction inside glass to a distance half the maximum wavelength and I still get the oscillation. Also, Besides the oscillation, I also think that the results inside the Solar cell is not correct.

Can someone explain what am I doing wrong?



Hi @rumman18,

Sorry for the wait. I am checking your file at the moment and I will get back with some comments by tomorrow.


Hi @rumman18,

I checked your results and I noticed that the transmission measured by the DFT monitors was larger than 1 for certain frequencies. This is probably because you used the stabilized profile for the PML, which is not the most appropriate choice in this case. You should only use this profile for simulations that are unstable due to the PML; in your case I would suggest using the “steep angle” profile as explained here:

I have some additional suggestions for improving the simulation (see Base_Model_modFG.fsp (358.9 KB) ):

  • The mesh you are using is not fine enough to resolve the thin layers (ZnO, for example). Therefore, I added a mesh override region around silicon and ZnO to refine the mesh along the z direction.
  • The material fit for silver can be improved by using the imaginary weight option, as you can see in the Material Explorer.
  • I noticed that the grooves in the structure seem to be at random positions as you can see in the index profile from “monitor_2”. Is this intended?

After these changes I got the following results for transmission and reflection:

Hope this helps!


@fgomez thank you so much!!!


@fgomez yes it’s intended. Quick follow up question. If my structure is random along xy plane, should I still use periodic boundary condition or should I just use PML in all direction?


Hi @rumman18,

If you have a random structure you can use a super-cell approach where you use periodic boundary conditions. The structure inside the periodic boundaries will be a super-cell that is repeated periodically. The size of the super-cell should be large enough to avoid coupling between the super-cells. This would be the right approach if you want to use a plane-wave source to illuminate the structure (since, conceptually speaking, the wavefronts from a plane wave source extend to infinity).

You could also use PML boundaries to simulate a finite region, but in this case you would need to use a source that is also finite in extension, for example a Gaussian beam.

Hope this helps!