Learning Forum

[Shangxuan Yu]

I am trying to do a mode simulation for a multimode waveguide, the radiation loss shows negative value, I don't understand why is that. I tried both loose and tight bend and varying the simulation region size, but couldn't get rid of this issue. I used PML for all boundaries, and I don't undertsand where does the gain come from.

[anna.wirth-singh]

Hi, thanks for your question. My first suggestion is that the PML boundaries may be too close to the waveguide. Would you please try increasing the size of the simulation region and see if that affects your result? This article may be helpful for explanation: Working with lossy modes and dB/m to \(\kappa\) conversion – Ansys Optics The simulation region should be large enough that the field decays to nearly zero at the boundaries. I hope this suggestion helps. 

If that is not the issue, unexpected gain can also be caused by material fitting issues. Are you using the built-in Si - Palik material for the waveguide, and what material is your substrate?

Best,

Anna

[Shangxuan Yu]

Hi Anna,

Thanks for the advice!

From the log plot, it seems the field has fully decayed with current simulation region size

I also tried to double the simulation region to 20 by 20 um, there seems no effect

 

 

The materials I used silicon for the waveguide and SiO2 for the substrate, and they both come from the material library, and I checked that the imaginary part is literally 0

[anna.wirth-singh]

Hi Shangxuan,

I apologize for the delay in responding to you, and thank you for sharing that additional information. I've looked into it a bit more, and since the imaginary part of the effective index and the loss/gain are very small (1e-8), this is most likely due to a numerical error. That is, this relatively small negative loss is probably below the level of precision that can be obtained, and most likely does not indicate any issues in the simulation. While there are still some negative values when the PML boundary is moved further away from the mode, the numbers are closer to zero, so this is another indication that the loss here is very close to zero. 

Best,

Anna