I am trying to use EME to simulate a long MMI on ARROW layers. However when I include the ARROW layers in the EME geometry, the solution becomes empty in the multimode section. If I do not include the ARROW layers in the EME geometry does it include the underlying layers in the calculation?

Hi @mattstott.bcm
Could you please explain more what is ARROW layers. What do you mean? Did you mean “Arrow waveguide” ?

ARROW layers refers to the periodic layers under the core that make up the
ARROW waveguide. These layers however are not on top as the air cladding
confines the mode in the oxide core.

Hi @mattstott.bcm
I made a research in Lumerical’s Knowledge base https://kb.lumerical.comenindex.html?nanophotonic_applications_arrow_waveguide.html and i found out that this oxide layer can also be above the core waveguide and on top there is air.
@bkhanaliloo I would like to ask, ARROW layers are there only under the core of the ARROW waveguide?
Also, is it possible to use these ARROW layers in the EME geometry?

Dear @konslekk

My knowledge is limited for the ARROW waveguide designs, but based on this Wikipedia page, they are under the core material:

However, it looks like people use coating on top of the core material, as is shown in the link you provided in the previous post.

I am not seeing any reason why not. If we can calculate the modes in FDE, we should be able to use EME for light propagation.

@mattstott.bcm: Can you please upload your simulation file for a review? I am not quite sure about the problem but hopefully we can figure it out.


Yes here is the simulation file for a review. Currently the EME geometry encompasses the ARROW waveguides which results in a weird solution. To obtain a solution I change the confines to not include the ARROW waveguides, change the z min in EME setup to -1.399 um. My question is toward the EME solver. When the geometry does not include the ARROWs does it still consider the underlying structure when solving the for the propagating modes?
LongARROWbasedMMI.lms (342.8 KB)


Hi @mattstott.bcm

Thank you for sharing the simulation file. Just by looking at it, I found a few critical error in your file. I will point out a few problems in this post, but I strongly recommend you to study our getting started examples:

  • EME boundaries are touching the geometry on top (+z). You need to leave at least half wavelength gap between geometry and EME boundaries
  • The injected mode from port object is not set properly. Expand MMI_Input and introduce a new cell to make sure that light is injected properly. You can visualize injected mod by right clicking the port object.
  • Since the Arrow section has high index material, you need to force software to look for the modes around the index of interest (rather than just increasing the number of modes). This idea is explained in this post: How to set searching index in EME by scripting?

After all, bare in mind that you will need to perform convergence testing as is explained here:

Please go ahead and keep me updated with your results. I will be glad to be of a help if you had further questions.