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 The information provided may be out of date.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.

Normal flux problem of transport of diluted species

Please login with a confirmed email address before reporting spam

Hi there,
I am solving possoin-nernst-planck equations using comsol.
When I combine Electrostatic module and transport of diluted species module together, I apply posive voltage and negative voltage on two opposite sides in es module and apply normal zero flux in transport of diluted species module. I use time dependent solver. Though I can get solution, I found normal flux is not zero at the boundary. I am so confused about this result since I predifined that.

A lot of thanks in advance.
Best,
Leo

9 Replies Last Post 24.03.2015, 11:33 GMT-4

Please login with a confirmed email address before reporting spam

Posted: 10 years ago 13.03.2015, 04:13 GMT-4
To add information, I have tick the migration under electric field box in transport of diluted species module and the flux is difined as Ni=-D*grad(c)-zuF*c*grad(V)+c*u(velocity vector). Velocity vector is defined as zero here.
Many Thanks!!!
To add information, I have tick the migration under electric field box in transport of diluted species module and the flux is difined as Ni=-D*grad(c)-zuF*c*grad(V)+c*u(velocity vector). Velocity vector is defined as zero here. Many Thanks!!!

Please login with a confirmed email address before reporting spam

Posted: 10 years ago 13.03.2015, 07:30 GMT-4
Hi

This is merely a guess, but how dense is your mesh close to the boundaries? You may define zero flux BCs, but COMSOL calculates the flux based on your mesh, which is prone to smallest changes in the mesh. I have with the flux BC a lot and never been fully content.

BR
Lasse
Hi This is merely a guess, but how dense is your mesh close to the boundaries? You may define zero flux BCs, but COMSOL calculates the flux based on your mesh, which is prone to smallest changes in the mesh. I have with the flux BC a lot and never been fully content. BR Lasse

Please login with a confirmed email address before reporting spam

Posted: 10 years ago 13.03.2015, 08:15 GMT-4
Dear Lasse,

Thank you so much for your reply!
My simulation geometry is just 2D 1 micrometer by 1 micrometer rectangular box. I use comsol predifined :Extremely fine size shown in the attachment.
I just wonder whether the unsymmetric oppsite boundaries(one with +1V the other with -1V applied) is possible for zero flux since I see many models in gallery are zero flux for symmetric boundaries.
Another question is does comsol first assume zero flux at BC then iterate?
Thank you!

Best,
Leo

Hi

This is merely a guess, but how dense is your mesh close to the boundaries? You may define zero flux BCs, but COMSOL calculates the flux based on your mesh, which is prone to smallest changes in the mesh. I have with the flux BC a lot and never been fully content.

BR
Lasse


Dear Lasse, Thank you so much for your reply! My simulation geometry is just 2D 1 micrometer by 1 micrometer rectangular box. I use comsol predifined :Extremely fine size shown in the attachment. I just wonder whether the unsymmetric oppsite boundaries(one with +1V the other with -1V applied) is possible for zero flux since I see many models in gallery are zero flux for symmetric boundaries. Another question is does comsol first assume zero flux at BC then iterate? Thank you! Best, Leo [QUOTE] Hi This is merely a guess, but how dense is your mesh close to the boundaries? You may define zero flux BCs, but COMSOL calculates the flux based on your mesh, which is prone to smallest changes in the mesh. I have with the flux BC a lot and never been fully content. BR Lasse [/QUOTE]


Please login with a confirmed email address before reporting spam

Posted: 10 years ago 14.03.2015, 09:09 GMT-4
In principle zero flux condition is possible, the concentration profile is then set so that the concentration gradient opposes potential gradient, making flux zero:

grad(c) + zF/(RT)*c*grad(U) = 0 --> grad(U) = - RT/(zF)*grad(ln(c))

I am not able to reply to your latter question, I do not know. Your mesh probably is very dense, but if your domain is that small, the minimum element size should perhaps further decreased in order to have better resolution at the boundaries.

br
Lasse
In principle zero flux condition is possible, the concentration profile is then set so that the concentration gradient opposes potential gradient, making flux zero: grad(c) + zF/(RT)*c*grad(U) = 0 --> grad(U) = - RT/(zF)*grad(ln(c)) I am not able to reply to your latter question, I do not know. Your mesh probably is very dense, but if your domain is that small, the minimum element size should perhaps further decreased in order to have better resolution at the boundaries. br Lasse

Please login with a confirmed email address before reporting spam

Posted: 10 years ago 14.03.2015, 10:12 GMT-4
Dear Lasse,

Thank you so much!
I have adoptted your suggestion of decreasing the minimal and maximal mesh sizes. I will report when I get the new resolution.
Many thank!

Best,
Leo
Dear Lasse, Thank you so much! I have adoptted your suggestion of decreasing the minimal and maximal mesh sizes. I will report when I get the new resolution. Many thank! Best, Leo

Please login with a confirmed email address before reporting spam

Posted: 10 years ago 17.03.2015, 08:26 GMT-4
Dear Lasse,

I have refined the mesh on the boundary, the flux do become smaller, but not within the error of comsol, I guess for my unsymmetric bc case, the flux might not be zero simultaneously...
Thank you so much!

Best,
Leo
Dear Lasse, I have refined the mesh on the boundary, the flux do become smaller, but not within the error of comsol, I guess for my unsymmetric bc case, the flux might not be zero simultaneously... Thank you so much! Best, Leo

Please login with a confirmed email address before reporting spam

Posted: 10 years ago 17.03.2015, 08:51 GMT-4
OK.

I still claim that your boundary conditions are quite OK :)

Lasse
OK. I still claim that your boundary conditions are quite OK :) Lasse

Please login with a confirmed email address before reporting spam

Posted: 10 years ago 17.03.2015, 10:41 GMT-4
Dear Lasse,

My next trial will be putting dielectric layer outside the two opposing boudaries to make grad(V) less steep, hope this will help get normal zero flux bc. I will update if it can work~~
Many thanks!

Best,
Leo
Dear Lasse, My next trial will be putting dielectric layer outside the two opposing boudaries to make grad(V) less steep, hope this will help get normal zero flux bc. I will update if it can work~~ Many thanks! Best, Leo

Please login with a confirmed email address before reporting spam

Posted: 10 years ago 24.03.2015, 11:33 GMT-4
Dear Lasse,

I have tried several ways but all failed. I tried to use pointwise constrain -user defined-constraint expression/constrain force expression to make boundary flux to be zero but also failed. I am so curious why pointwise constrain fail to give what it constains to?
Thanks a lot!

Best Regards,
Leo
Dear Lasse, I have tried several ways but all failed. I tried to use pointwise constrain -user defined-constraint expression/constrain force expression to make boundary flux to be zero but also failed. I am so curious why pointwise constrain fail to give what it constains to? Thanks a lot! Best Regards, Leo

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.