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 The information provided may be out of date.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.

Metal-Insulator boundary

Please login with a confirmed email address before reporting spam

Hello everyone

I have a question about boundary conditions.
I have 3 layers (Insulator-metal-insulator) and this structure is placed in a solution that i want to sense its voltage with that metal layer.
Insulators are SIO2.
The problem is that when i choose dielectric shielding for insulator boundaries, the voltage on the metal, insulator and the solution will be all the same (and i was expected that it should not be the same on insulator layer!)
So i thought may be i should choose ground for insulator (!) and then the problem is solved. I have zero on insulator, different voltages on metal, and in solution.

But i beleive there is something wrong and i shouldn't work like this. Anyone has any idea?

Best
Bahar

13 Replies Last Post 26.03.2013, 16:22 GMT-4

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 10.05.2012, 12:14 GMT-4
Bahar,

do you need the specific boundary condition in your model? The SiO2 layer is already dielectric. Maybe you can keep it all at the default continuity.
Putting a ground BC to one of the insulators forces the adjacent metal close to ground potential, which is probably not what you want.

Cheers
Edgar
Bahar, do you need the specific boundary condition in your model? The SiO2 layer is already dielectric. Maybe you can keep it all at the default continuity. Putting a ground BC to one of the insulators forces the adjacent metal close to ground potential, which is probably not what you want. Cheers Edgar

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 10.05.2012, 15:50 GMT-4
Dear Edgar

I was thimking about the same issue when using ground layer. But the is the only case that i can reach to different voltage on isolator layer and metal layer. However, I, myself , know that there should be something wrong here.
I tried following what you have told me (not using dielectric shielding and leaving boundaries of SIO2 as they are. To me it should be more correct to do this. However, when i checked the results the still the voltage in all the system( isolator, metal, and solution) is almost the same and there is a small difference between them. However, the interesting part is that now the voltage is higher in isolator compare to the metal layer!!!
I will try to do this simulation with different boundaries to see what will happen but i will appreciate if you gave me any other comments.


Best
Bahar

Dear Edgar I was thimking about the same issue when using ground layer. But the is the only case that i can reach to different voltage on isolator layer and metal layer. However, I, myself , know that there should be something wrong here. I tried following what you have told me (not using dielectric shielding and leaving boundaries of SIO2 as they are. To me it should be more correct to do this. However, when i checked the results the still the voltage in all the system( isolator, metal, and solution) is almost the same and there is a small difference between them. However, the interesting part is that now the voltage is higher in isolator compare to the metal layer!!! I will try to do this simulation with different boundaries to see what will happen but i will appreciate if you gave me any other comments. Best Bahar

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 11.05.2012, 03:51 GMT-4
Bahar,

can you post the model? It is difficult to do more diagnosis without.

Cheers
Edgar
Bahar, can you post the model? It is difficult to do more diagnosis without. Cheers Edgar

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 11.05.2012, 11:03 GMT-4
Dear Edgar
I attached the model. You can see when i don't have any ground, the voltage is the same every where (and i am expecting that at least it should be different in insulators ). However, just today i tried to ground one of the boundaries and use that as a reference to see what will happen. So if you test the simulation by activating the ground boundary you can see wired results!!
In this case, the voltage in insulator layer is higher than voltage in metal layer (i checked that by line plots) and we will have nonsense variation of voltage inside the metal!
I would be grateful if you can give me your idea on this.
Best
Bahar


P.S. Just today i find out that university change our Comsol to version 4.2a, i set this topic for 4.2. Hope you can still run the model.
Dear Edgar I attached the model. You can see when i don't have any ground, the voltage is the same every where (and i am expecting that at least it should be different in insulators ). However, just today i tried to ground one of the boundaries and use that as a reference to see what will happen. So if you test the simulation by activating the ground boundary you can see wired results!! In this case, the voltage in insulator layer is higher than voltage in metal layer (i checked that by line plots) and we will have nonsense variation of voltage inside the metal! I would be grateful if you can give me your idea on this. Best Bahar P.S. Just today i find out that university change our Comsol to version 4.2a, i set this topic for 4.2. Hope you can still run the model.


Ivar KJELBERG COMSOL Multiphysics(r) fan, retired, former "Senior Expert" at CSEM SA (CH)

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 12.05.2012, 12:57 GMT-4
Hi

I can only advice yu to upgrae, and to patch it to the lates, next month I believe the 4.3 is coming, you never end chasing version with software ...

I do believe you are getting fooled, as I'm often in ES, do not forget it's Electro-STATICS and you do not get any current flow i your metals, therefore I always use EC (electric CURRENTS), then when I ground a boundary all my metal gets the same potential, to the resisitive value / ohm law results

In ES you do not need the internal material only the boundaries, and they must be fully selected to set them at a given charge (these are not moving on your conductors, at least that is how I see it ;)

Check carefully all hypothesis behind each physics, one can easily assume something against one of these, without really noticing it immediatly ;)

--
Good luck
Ivar
Hi I can only advice yu to upgrae, and to patch it to the lates, next month I believe the 4.3 is coming, you never end chasing version with software ... I do believe you are getting fooled, as I'm often in ES, do not forget it's Electro-STATICS and you do not get any current flow i your metals, therefore I always use EC (electric CURRENTS), then when I ground a boundary all my metal gets the same potential, to the resisitive value / ohm law results In ES you do not need the internal material only the boundaries, and they must be fully selected to set them at a given charge (these are not moving on your conductors, at least that is how I see it ;) Check carefully all hypothesis behind each physics, one can easily assume something against one of these, without really noticing it immediatly ;) -- Good luck Ivar

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 12.05.2012, 13:27 GMT-4
Dear Ivar

I appreciate your helpful reply. This ES sometimes confused me a lot. I also rather use EC since everything is much more clear for me in that. However, for this simulation i need to define a charge and in ES i can do that with using surface charge density or set a terminal as charge terminal. But I can't do this in EC and can't find out how it is possible to define a simple point charge in that. That was the reason i use ES here. Do you have any idea about this?

I will recheck everything and i will follow your advice on boundaries issue.
Once a gain thanks a lot for your reply

Best
Bahar
Dear Ivar I appreciate your helpful reply. This ES sometimes confused me a lot. I also rather use EC since everything is much more clear for me in that. However, for this simulation i need to define a charge and in ES i can do that with using surface charge density or set a terminal as charge terminal. But I can't do this in EC and can't find out how it is possible to define a simple point charge in that. That was the reason i use ES here. Do you have any idea about this? I will recheck everything and i will follow your advice on boundaries issue. Once a gain thanks a lot for your reply Best Bahar

Ivar KJELBERG COMSOL Multiphysics(r) fan, retired, former "Senior Expert" at CSEM SA (CH)

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 12.05.2012, 14:58 GMT-4
Hi

well ES allows you to define a charge because it cannot migrate away ;)

in EC you can define your charge via a voltage and a capacitance, or just that current is charges per seconds.
Again EC you have the charges flowing away, but regularly replaced, so If you define a charge alone in EC, normally it will indeed "discharge" via the current flow.

--
Good luck
Ivar
Hi well ES allows you to define a charge because it cannot migrate away ;) in EC you can define your charge via a voltage and a capacitance, or just that current is charges per seconds. Again EC you have the charges flowing away, but regularly replaced, so If you define a charge alone in EC, normally it will indeed "discharge" via the current flow. -- Good luck Ivar

Ivar KJELBERG COMSOL Multiphysics(r) fan, retired, former "Senior Expert" at CSEM SA (CH)

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 12.05.2012, 16:25 GMT-4
And I would add

to get ES to look alike EC you must select all boundaries around any metal, as charges tend to concentrate on the external boundaries, then you can even drop the interiour of conductors

--
Good luck
Ivar
And I would add to get ES to look alike EC you must select all boundaries around any metal, as charges tend to concentrate on the external boundaries, then you can even drop the interiour of conductors -- Good luck Ivar

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 13.05.2012, 04:29 GMT-4
Hi Bahar,
I have a problem about input reactions in comsol, my problem is that the rate constant of my reaction is based on m^3/s
but in comsol its unit m^3mol-1.s-1.Do you know how do i should input these rate constant?

So thanks,
Azadeh
Hi Bahar, I have a problem about input reactions in comsol, my problem is that the rate constant of my reaction is based on m^3/s but in comsol its unit m^3mol-1.s-1.Do you know how do i should input these rate constant? So thanks, Azadeh

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 16.05.2012, 10:09 GMT-4
Hi Ivar

I appreciate your helpful comment. Right now I'm trying to move my simulation to EC to find out what will be the result there. I will update here about the results :)
Hi Ivar I appreciate your helpful comment. Right now I'm trying to move my simulation to EC to find out what will be the result there. I will update here about the results :)

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 16.05.2012, 10:12 GMT-4
Hi Azadeh

I can't understand the problem. Can you give us more information. And I think if you open a new post with relevant title more people can help you :)

Best
Bahar
Hi Azadeh I can't understand the problem. Can you give us more information. And I think if you open a new post with relevant title more people can help you :) Best Bahar

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 26.03.2013, 15:30 GMT-4
I have read this conversation as it relates to a problem I am facing.

In regards to,


"And I would add

to get ES to look alike EC you must select all boundaries around any metal, as charges tend to concentrate on the external boundaries, then you can even drop the interiour of conductors"


what does selecting all boundaries around metal mean? Do you mean grounding the metal surfaces?

Khyati

I have read this conversation as it relates to a problem I am facing. In regards to, "And I would add to get ES to look alike EC you must select all boundaries around any metal, as charges tend to concentrate on the external boundaries, then you can even drop the interiour of conductors" what does selecting all boundaries around metal mean? Do you mean grounding the metal surfaces? Khyati

Ivar KJELBERG COMSOL Multiphysics(r) fan, retired, former "Senior Expert" at CSEM SA (CH)

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 26.03.2013, 16:22 GMT-4
Hi

ES electrostatic is set up without current flow (thereof the "static" name) that is your conductor does not lead electrons, but you can charge up a surface. Therefore any metal conducting surface can be ignored (as bulk item) you need to define only the surrounding boundary and apply a voltage, assuming its constant over the boundary (or defined by a field if you know that)

EC is solving for current flow, hence your bulk conductor will transmit electrons (except in high frequency regimes, when the skin effect will limit the current flow to the external wire boundaries, then again its enough to model the surfaces, and you might ignore the conductors (use surface physics in that case)

Note that you get also singularities at any sharp edges, and even if sparking is known to start at sharp edges, the field what you see are not necesaily numerically correct just on the sharp vertexes, you should only trust them "qualitatively"

--
Good luck
Ivar
Hi ES electrostatic is set up without current flow (thereof the "static" name) that is your conductor does not lead electrons, but you can charge up a surface. Therefore any metal conducting surface can be ignored (as bulk item) you need to define only the surrounding boundary and apply a voltage, assuming its constant over the boundary (or defined by a field if you know that) EC is solving for current flow, hence your bulk conductor will transmit electrons (except in high frequency regimes, when the skin effect will limit the current flow to the external wire boundaries, then again its enough to model the surfaces, and you might ignore the conductors (use surface physics in that case) Note that you get also singularities at any sharp edges, and even if sparking is known to start at sharp edges, the field what you see are not necesaily numerically correct just on the sharp vertexes, you should only trust them "qualitatively" -- Good luck Ivar

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.