Discussion Forum

Posted: 1 decade ago 31 mag 2012, 07:42 GMT-4

Hi I beleive you have the isue that in ES Electro STATIC you do not have any flow (currents), at least from my understanding. While in EC you need conductivity everywhere, so you cannot build up static charges, appart via I=dQ/dt. -- Good luck Ivar

Posted: 1 decade ago 31 mag 2012, 11:26 GMT-4

I see, thank you. But the problem seems to be more related to the material itself or at most to the conditions at the interfaces, because I find a charge distribution inside the insulator which has zero conductivity (or really low). I will post the solution as soon as I will figure out how to build a reliable capacitor made of resistive electrodes. Or has anybody already found solution to this problem? Thank you, good evening Cosimo

Posted: 1 decade ago 6 giu 2012, 05:05 GMT-4

Good morning you all, at the moment i am able to simulate the field and the voltage potential inside the capacitor but it seems that COMSOL cannot compute perfectly thin space charge distribution. Reminding you the problem, we would see the shape of the space charge distribution inside the metal electrode and the resistive layer of a capacitor with an applied voltage. That is I would see what is happening at nanometric scale. I tried reducing the dimensions of the elements in the mesh, but it doesn't help, resulting only in a increased computational complexity. I am using ES and EC module; solving Laplace should be enough.. Should I maybe use other modules? Or is more a problem of the space resolution of the solver? Is there anybody that could give any suggestion? Thank you for your interest, Cosimo

Posted: 1 decade ago 6 giu 2012, 16:23 GMT-4

Hi it would be much easer for us to understand if we got a model, or you explain all the steps to build a simple model showing your case. If you clear the solution and the mesh, it should fit and be uploadable to the Forum -- Good luck Ivar

Posted: 1 decade ago 8 giu 2012, 08:05 GMT-4

Hi all, we upload the file. The problem is that the charge distribution is a delta fuction across the interfaces between the elctrodes and the insulator. We were expecting to not find any charges inside the insulator and a larger space charge distribution in the resistive layer. We guess that somehow the results are limited by the resolution of the mesh. What do you think? Thank you for your attention Kind regards Cosimo

Attachments:

• 08_06_2012 TEST FILE UPLOAD 2.mph

Posted: 1 decade ago 8 giu 2012, 15:26 GMT-4

Hi I believe you have a problem by defining both ES and EC with the same dependent variable name V, as there is an unique soluton only for one or the other not for both physics driving the same depdendet variable. And normally you need then to drive one physics by the other or bi-directionally -- Good luck Ivar

Posted: 1 decade ago 11 giu 2012, 05:17 GMT-4

Hi, should we define then 2 variables? But how is it possible to link the two physics by having two different variables? Kind regards

Posted: 1 decade ago 11 giu 2012, 15:25 GMT-4

Hi for me you should have V1 and V2 as your physics do not consider the same material properties to solve the mdel and will therefore not give exactly the same results (in ES = static for me at least there is no current flow, only chanrge buildup) -- Good luck Ivar

Posted: 1 decade ago 12 giu 2012, 03:49 GMT-4

[quote]I am using ES and EC module; solving Laplace should be enough..[/quote] [quote]The problem is that the charge distribution is a delta fuction across the interfaces between the elctrodes and the insulator. We were expecting to not find any charges inside the insulator and a larger space charge distribution in the resistive layer. We guess that somehow the results are limited by the resolution of the mesh. [/quote] The right-hand side of the Laplace equation is zero. In other words, the space charge term is zero. So this is not a problem of mesh resolution: the problem you're solving doesn't allow for space charges to begin with. You can have an interface charge, which you call a "delta function across the interfaces", though mathematically(!) these are defined at the boundary and at the boundary only. In the EC case, boundaries will be charged if they are discontinuous in conductivity. In the ES case, only exterior boundaries (electrodes) can be charged, since there is no charge transport through the bulk.

Posted: 1 decade ago 19 giu 2012, 09:50 GMT-4

Hi John, thank you of your reply. The point is that the peak in the distribution of charges is also getting a bit inside the insulator, we were expecting that delta fuction at the interface, only on the metal side instead.That's why I thought that the solver couldn't solve nanometric layer. We have also being trying to keep the problem only wtih one physics at each time as you suggested, but we didn't solve. For the moment I should say that is difficult to simulate space distribution of charges inside materials! Any suggestion is welcome. thank you Cosimo