Note: This discussion is about an older version of the COMSOL Multiphysics® software. The information provided may be out of date.

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.

Identity pair as port source in Comsol 4

Please login with a confirmed email address before reporting spam

Hi all

Does anybody knows how to set "identiy pair" as "port" in Comsol 4.0a or 4.1? I am calculating the reflectance and transmittance in Comsol. Previously, I used Comsol 3.5a for this work (refer to: www.comsol.com/showroom/gallery/6141/ ) and it worked quite well. However, in the new Comsol, it seems that this function is lost and if I do this no field can be seen. I have tried other ways to calculation R and T, e.g., using scattering boundary or port without setting PML, however, the results are not very good due to the reflection of the scattering interface under tilted incidence.

Actually, what I am interested in is putting a PML above a port source and the incident light go only into the structure without going into the PML (only reflected light will enter PML). In old version of Comsol, we cannot set interior boundary as port, so "assembly" must be used. In the new Comsol, however, although interioe boundary can be set as port, the light goes to both directions.

It will be very appreciated if any of you can provide me some information about this problem.

Cheers
Xiaofeng LI


10 Replies Last Post 23 ott 2013, 03:33 GMT-4

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 10 nov 2010, 02:03 GMT-5
A good question.
Maybe the bug of V4. Contact the support please.
A good question. Maybe the bug of V4. Contact the support please.

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 29 set 2011, 17:46 GMT-4
Dear Xiaofeng LI,
Have you solved the problem?
I am finding out the way solving the same problem.
Dear Xiaofeng LI, Have you solved the problem? I am finding out the way solving the same problem.

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 18 ott 2011, 15:03 GMT-4
Hi,
I have exactly the same problem with 4.2. Please let me know if you get any answer from support or any other source.

Thans
Hi, I have exactly the same problem with 4.2. Please let me know if you get any answer from support or any other source. Thans

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 19 ott 2011, 04:53 GMT-4
Hi, I am still using my Comsol 3.5a for this problem, although there may be an alternative way to do it in Comsol 4.1 or 4.2. But as I know, even in the latest 4.2 version, there has the same problem. Hope Comsol support can provide more information on that.
I will change to 4.2 also and do the similar calculation. If anyone of you found the solution, please kindly attach it here. Thanks!
Hi, I am still using my Comsol 3.5a for this problem, although there may be an alternative way to do it in Comsol 4.1 or 4.2. But as I know, even in the latest 4.2 version, there has the same problem. Hope Comsol support can provide more information on that. I will change to 4.2 also and do the similar calculation. If anyone of you found the solution, please kindly attach it here. Thanks!

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 24 ott 2011, 09:12 GMT-4
Hi I got a very detailed answer from support(attached). Unfortunately it hasn't worked for me. I keep on writing to support and trying to understand...

The files that they sent me a pretty big... so I upload only one of them. The method in both of them was pretty the same

COMSOL's support:

Ports on identity boundaries may make a return in one form or another at
some point, but they are currently not supported. Instead, we recommend a
different approach described in the two models and the write-up here below.

Scattering on single scatterers in free space is preferably modeled by
sending in a plane wave and solving for the scattered field. If the object
is placed on a substrate, the correct expression for the background field
becomes more complicated. It needs to be the correct superposition of an
incident and a reflected wave in the free space domain, and a transmitted
wave in the substrate.

A simple and general way to avoid deriving and entering the analytical
background field is to use a full field solution in the background
materials. This is done in the Electromagnetic Waves (emw) interfaces in
both the provided models. Because only the background materials are present
in this interface, the wave remains plane and no PMLs are needed.

The Electromagnetic Waves (emw2) interface solves the scattered field
problem with the emw solution as the input. The scattered field solution of
this simulation is due only to the presence of the scatterer. PMLs capture
this in all directions.

If you remove the side PML domains and use the same side boundary
conditions in both interfaces, you are instead modeling an array of
scatterers. The 2d_array_on_substrate model is set up for perpendicular
wave incidence, but with Floquet conditions it would work for oblique
incidence too. This is an alternative to the multi-port method shown in the
Plasmonic Wire Grating model, and becomes an attractive option if there are
many diffractive orders present.
Hi I got a very detailed answer from support(attached). Unfortunately it hasn't worked for me. I keep on writing to support and trying to understand... The files that they sent me a pretty big... so I upload only one of them. The method in both of them was pretty the same COMSOL's support: Ports on identity boundaries may make a return in one form or another at some point, but they are currently not supported. Instead, we recommend a different approach described in the two models and the write-up here below. Scattering on single scatterers in free space is preferably modeled by sending in a plane wave and solving for the scattered field. If the object is placed on a substrate, the correct expression for the background field becomes more complicated. It needs to be the correct superposition of an incident and a reflected wave in the free space domain, and a transmitted wave in the substrate. A simple and general way to avoid deriving and entering the analytical background field is to use a full field solution in the background materials. This is done in the Electromagnetic Waves (emw) interfaces in both the provided models. Because only the background materials are present in this interface, the wave remains plane and no PMLs are needed. The Electromagnetic Waves (emw2) interface solves the scattered field problem with the emw solution as the input. The scattered field solution of this simulation is due only to the presence of the scatterer. PMLs capture this in all directions. If you remove the side PML domains and use the same side boundary conditions in both interfaces, you are instead modeling an array of scatterers. The 2d_array_on_substrate model is set up for perpendicular wave incidence, but with Floquet conditions it would work for oblique incidence too. This is an alternative to the multi-port method shown in the Plasmonic Wire Grating model, and becomes an attractive option if there are many diffractive orders present.


Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 19 ott 2012, 09:44 GMT-4
I can report that this approach does not work for me in COMSOL 4.3 because the PML now cannot be assigned to a single emw (Electromagnetic waves) interface but instead is applied to all. Please, help me to find how to solve current problem in 4.3
I can report that this approach does not work for me in COMSOL 4.3 because the PML now cannot be assigned to a single emw (Electromagnetic waves) interface but instead is applied to all. Please, help me to find how to solve current problem in 4.3

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 29 nov 2012, 06:50 GMT-5
Dear Gary Zaiats ,

I have got your answer from Comsol Discussion Forum.I run your attach and got some sesults.Now I want to get other files that you said ''The files that they sent me a pretty big... so I upload only one of them. The method in both of them was pretty the same". I want to solve the problem of transmission index and refrective index. If you have some spare time ,sent me the files ,OK? Thanks a lot. my email adress is " 0516zhoujian@tongji.edu.cn ".
Dear Gary Zaiats , I have got your answer from Comsol Discussion Forum.I run your attach and got some sesults.Now I want to get other files that you said ''The files that they sent me a pretty big... so I upload only one of them. The method in both of them was pretty the same". I want to solve the problem of transmission index and refrective index. If you have some spare time ,sent me the files ,OK? Thanks a lot. my email adress is " 0516zhoujian@tongji.edu.cn ".

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 29 nov 2012, 06:57 GMT-5
I can advise to build the model using the scattered field approach. I used it to solve a periodic system of graphene ribbons and it worked perfectly in 4.3a.
I can advise to build the model using the scattered field approach. I used it to solve a periodic system of graphene ribbons and it worked perfectly in 4.3a.

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 7 giu 2013, 04:13 GMT-4
To Mychailo:

the problem illustrated in the 2D_array_on_substrate.mph file is exactly what I need to implement. However, as you observed it is not possible to set-it up in the version 4.3.

Did you find a different way to set-up the problem or simply by upgrading to version 4.3a you could use the again the same settings of the example?

Please let me know if it worth installing the new version :)
To Mychailo: the problem illustrated in the 2D_array_on_substrate.mph file is exactly what I need to implement. However, as you observed it is not possible to set-it up in the version 4.3. Did you find a different way to set-up the problem or simply by upgrading to version 4.3a you could use the again the same settings of the example? Please let me know if it worth installing the new version :)

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 23 ott 2013, 03:33 GMT-4
In the model [2d_array_on_substrate.mph], using PEC, the scattered light from scatter will be made to perpendicular to the plane of PEC, is it correct?


Hi I got a very detailed answer from support(attached). Unfortunately it hasn't worked for me. I keep on writing to support and trying to understand...

The files that they sent me a pretty big... so I upload only one of them. The method in both of them was pretty the same

COMSOL's support:

Ports on identity boundaries may make a return in one form or another at
some point, but they are currently not supported. Instead, we recommend a
different approach described in the two models and the write-up here below.

Scattering on single scatterers in free space is preferably modeled by
sending in a plane wave and solving for the scattered field. If the object
is placed on a substrate, the correct expression for the background field
becomes more complicated. It needs to be the correct superposition of an
incident and a reflected wave in the free space domain, and a transmitted
wave in the substrate.

A simple and general way to avoid deriving and entering the analytical
background field is to use a full field solution in the background
materials. This is done in the Electromagnetic Waves (emw) interfaces in
both the provided models. Because only the background materials are present
in this interface, the wave remains plane and no PMLs are needed.

The Electromagnetic Waves (emw2) interface solves the scattered field
problem with the emw solution as the input. The scattered field solution of
this simulation is due only to the presence of the scatterer. PMLs capture
this in all directions.

If you remove the side PML domains and use the same side boundary
conditions in both interfaces, you are instead modeling an array of
scatterers. The 2d_array_on_substrate model is set up for perpendicular
wave incidence, but with Floquet conditions it would work for oblique
incidence too. This is an alternative to the multi-port method shown in the
Plasmonic Wire Grating model, and becomes an attractive option if there are
many diffractive orders present.


In the model [2d_array_on_substrate.mph], using PEC, the scattered light from scatter will be made to perpendicular to the plane of PEC, is it correct? [QUOTE] Hi I got a very detailed answer from support(attached). Unfortunately it hasn't worked for me. I keep on writing to support and trying to understand... The files that they sent me a pretty big... so I upload only one of them. The method in both of them was pretty the same COMSOL's support: Ports on identity boundaries may make a return in one form or another at some point, but they are currently not supported. Instead, we recommend a different approach described in the two models and the write-up here below. Scattering on single scatterers in free space is preferably modeled by sending in a plane wave and solving for the scattered field. If the object is placed on a substrate, the correct expression for the background field becomes more complicated. It needs to be the correct superposition of an incident and a reflected wave in the free space domain, and a transmitted wave in the substrate. A simple and general way to avoid deriving and entering the analytical background field is to use a full field solution in the background materials. This is done in the Electromagnetic Waves (emw) interfaces in both the provided models. Because only the background materials are present in this interface, the wave remains plane and no PMLs are needed. The Electromagnetic Waves (emw2) interface solves the scattered field problem with the emw solution as the input. The scattered field solution of this simulation is due only to the presence of the scatterer. PMLs capture this in all directions. If you remove the side PML domains and use the same side boundary conditions in both interfaces, you are instead modeling an array of scatterers. The 2d_array_on_substrate model is set up for perpendicular wave incidence, but with Floquet conditions it would work for oblique incidence too. This is an alternative to the multi-port method shown in the Plasmonic Wire Grating model, and becomes an attractive option if there are many diffractive orders present. [/QUOTE]

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.