Menu

Discussion Forum

New Discussion
查找

Discussion Closed This discussion was created more than 6 months ago and has been closed. To start a new discussion with a link back to this one, click here.

Unexpected S-Parameters in Waveguide Simulation

Posted 2023年5月4日 GMT-4 07:57 Wave Optics, Results & Visualization Version 5.6 4 Replies

Christopher Waas
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Hi all,

I'm trying to simulate a 3D nanophotonic waveguide. The structure is very simple, a slab of material in perfect vacuum. I applied a PML all around, and now I want to investigated the propagation of supported modes. For that, I follow the post in Copying and Resuing Boundary Mode Analysis Results, copying faces and using withsol() to "listen" to the correct mode.

I created an excitation port (1) and an observer port (2) at the same face, as well as some additional ports. Mainly, I would expect S11 and S21 to be the same value, and I would expect that the excited mode propagates very well as the effective refractive index has a small imaginary part.

The Multislice seems to show a nice propagation, yet the S-parameters are way worse than I would expect. Especially the difference between S11 and S21 is surprising to me.

Am I missing something in the way the S-parameters are calculated or how the ports act? Or are my expections wrong? I'd be glad about any input!

Thanks!

Attachments:

4 Replies Last Post 2023年5月11日 GMT-4 06:45
Dave Greve
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 2 years ago 2023年5月5日 GMT-4 16:00

I can't run your model because I don't have wave optics. But- I don't understand why you have so many ports.

Seems to me you have one input port and one output port. You only need more ports if there are multiple propagating modes.

When set up properly with linear materials and a single mode type you should find S11=S22 and S12=S21 (or is there a minus sign, I forget).

I can't run your model because I don't have wave optics. But- I don't understand why you have so many ports. Seems to me you have one input port and one output port. You only need more ports if there are multiple propagating modes. When set up properly with linear materials and a single mode type you should find S11=S22 and S12=S21 (or is there a minus sign, I forget).
Christopher Waas
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 2 years ago 2023年5月8日 GMT-4 08:22

Hi Dave,

thanks for your answer. Indeed, I added ports to check for the cross-talk between the two allowed propagating modes. One port (port 6) was only added after I realized that the other S-parameters do not behave in the way I would expect. With this one, I wanted to monitor if the wave propagates as I would imagine.

Do ports influence the solution? Or are they only evaluated in post-processing?

Hi Dave, thanks for your answer. Indeed, I added ports to check for the cross-talk between the two allowed propagating modes. One port (port 6) was only added after I realized that the other S-parameters do not behave in the way I would expect. With this one, I wanted to monitor if the wave propagates as I would imagine. Do ports influence the solution? Or are they only evaluated in post-processing?
Dave Greve
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 2 years ago 2023年5月8日 GMT-4 22:38
Updated: 2 years ago 2023年5月8日 GMT-4 22:41

Do ports influence the solution? Or are they only evaluated in post-processing?

You can see that the ports are part of the calculation because for every port there is a corresponding S parameter that appears under "dependent variables."

I think it is fair to say it this way- Comsol calculates a solution to Maxwell's equations including outgoing waves at the ports with magnitude and phase that are variables.

What this means- if physically there should be outgoing waves and the port does not allow them the result will be incorrect. This could happen, for example, if you specified an output port with an electric field perpendicular to the electric field for the input port.

I think what you should be doing in your calculation is doing a boundary mode analysis for an input port consisting of the diamond and the surrounding vacuum. Same thing on the output port. If there was something that would result in mode conversion to another mode (say a disconinuity or obstacle of some sort) then there could be additional modes that propagate. In that case you would need to explicitly add additional ports.

Note that when you do the boundary mode analysis, you specify the effective mode index or out of plane wavenumber. These specifications allow you to select particular modes.

>Do ports influence the solution? Or are they only evaluated in post-processing? You can see that the ports are part of the calculation because for every port there is a corresponding S parameter that appears under "dependent variables." I think it is fair to say it this way- Comsol calculates a solution to Maxwell's equations including outgoing waves at the ports with magnitude and phase that are variables. What this means- if physically there should be outgoing waves and the port does not allow them the result will be incorrect. This could happen, for example, if you specified an output port with an electric field perpendicular to the electric field for the input port. I think what you should be doing in your calculation is doing a boundary mode analysis for an input port consisting of the diamond and the surrounding vacuum. Same thing on the output port. If there was something that would result in mode conversion to another mode (say a disconinuity or obstacle of some sort) then there could be additional modes that propagate. In that case you would need to explicitly add additional ports. Note that when you do the boundary mode analysis, you specify the effective mode index or out of plane wavenumber. These specifications allow you to select particular modes.
Christopher Waas
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 2 years ago 2023年5月11日 GMT-4 06:45

Hi Dave,

thanks for the reply. I am indeed doing a boundary mode analysis and propagating the solution to the various numeric ports. I now realized that having multiple "observer" ports listening to the same mode is probably not a good idea. They seem to influence each other. If I disable most of the off-ports in my model, the results make sense. Mainly port 2 in my model seemed to have caused the issues.

Thanks for your input!

Hi Dave, thanks for the reply. I am indeed doing a boundary mode analysis and propagating the solution to the various numeric ports. I now realized that having multiple "observer" ports listening to the same mode is probably not a good idea. They seem to influence each other. If I disable most of the off-ports in my model, the results make sense. Mainly port 2 in my model seemed to have caused the issues. Thanks for your input!

Note that while COMSOL employees may participate in the discussion forum, COMSOL® software users who are on-subscription should submit their questions via the Support Center for a more comprehensive response from the Technical Support team.

Suggested Content