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 The information provided may be out of date.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.

Surface Charge--Visualization and Computation

Please login with a confirmed email address before reporting spam

I have constructed a model that is a modified transistor and only want to see the effect of electrostatics on it (not an actual functioning transistor). The element of the model that is most significant to me is a conductor with a floating potential. One surface is connected to the air, and the other surfaces touch oxide. I am trying to compute the surface charge (by integrating surface charge density) on two of the surfaces touching oxide. I am also trying to create a picture of the charge distribution on the floating conductor.

I tried to leave the conductor domains out of the solution (by deselecting them under my Electrostatics tab). My problems are the following:

1) When I try to integrate the surface of my conductor, I get the error message "Failed to evaluate variable. Variable: mod1.es.nD" etc.

2) When I try to create the pictures, I am unable to get the view that I created in the definitions tab (that hides other entities/objects) to also be used in my 3D Plot Group (Surface), so there are objects/domains in the way of my plot. I am also hesitant that the parts I can see are working, based on problem 1). In other words, if it cannot evaluate the variable for surface charge density, then I don't know that is plotting it, and I'm only seeing a uniform color across that part of the device.

Is it possible to read this information (surface charge density) off of a floating conductor using COMSOL? If so, are there any tips on how to treat the conductor and solve it correctly to get this information?

I had experimented with using the polarization option instead of surface charge density. It has the same units as es.nD (surface charge density), but isn't quite what I am looking for.

Thanks!

6 Replies Last Post 2012年8月1日 GMT-4 12:05

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 2012年7月27日 GMT-4 07:39
Is it possible to read this information (surface charge density) off of a floating conductor using COMSOL?

Yes, it is possible, and the way to do it is pretty much what you described. So there's nothing wrong with your approach per se.
[quote]Is it possible to read this information (surface charge density) off of a floating conductor using COMSOL?[/quote] Yes, it is possible, and the way to do it is pretty much what you described. So there's nothing wrong with your approach per se.

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 2012年7月27日 GMT-4 16:00
Thanks. I guess the trick is trying to find out what's going wrong. I have been able to plot the surface charge density on every surface except the conductor. There's always an error. I wonder if it has something to do with the program not solving on the conducting domains? Maybe there's some obscure selection I must make in the solution tab. It's not immediately clear to me. Also, the integration on the surface of the conductors does not work either.
Thanks. I guess the trick is trying to find out what's going wrong. I have been able to plot the surface charge density on every surface except the conductor. There's always an error. I wonder if it has something to do with the program not solving on the conducting domains? Maybe there's some obscure selection I must make in the solution tab. It's not immediately clear to me. Also, the integration on the surface of the conductors does not work either.

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 2012年7月30日 GMT-4 03:04
I have been able to plot the surface charge density on every surface except the conductor.

In electrostatics, you should see no surface charges (or rather, interface charges) except for the surfaces of conductors.

I wonder if it has something to do with the program not solving on the conducting domains?

No, that's exactly what the program is supposed to do: solve the Laplace equation for the electric potential inside a volume (the dielectric) bounded by surfaces of conductors (where the value of the electric potential is a given constant).

You seem reluctant to upload your model, which you will have your reasons for. Regardless, the best thing to do in a case like this is make a simplified model. I'd suggest a (metallic) sphere inside a box. Apply voltage and ground at opposite faces of the box, set the spherical surface to "floating potential" (zero total charge), solve the electrostatic problem outside the sphere, and plot the surface charge density. You should see a sphere that is positively charged on one half of its surface, negatively charged on the other. This should help you figure out what's wrong with your model.

[quote]I have been able to plot the surface charge density on every surface except the conductor.[/quote] In electrostatics, you should see no surface charges (or rather, interface charges) except for the surfaces of conductors. [quote]I wonder if it has something to do with the program not solving on the conducting domains?[/quote] No, that's exactly what the program is supposed to do: solve the Laplace equation for the electric potential inside a volume (the dielectric) bounded by surfaces of conductors (where the value of the electric potential is a given constant). You seem reluctant to upload your model, which you will have your reasons for. Regardless, the best thing to do in a case like this is make a simplified model. I'd suggest a (metallic) sphere inside a box. Apply voltage and ground at opposite faces of the box, set the spherical surface to "floating potential" (zero total charge), solve the electrostatic problem outside the sphere, and plot the surface charge density. You should see a sphere that is positively charged on one half of its surface, negatively charged on the other. This should help you figure out what's wrong with your model.

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 2012年7月30日 GMT-4 12:20
Thanks so much for your responses!

I tried the simplified model as you suggested, with a copper sphere inside of a box of air. I set the boundaries of the box to ground, except for one side which I gave a potential. I set the boundaries of the sphere to floating potential, as in my other model. I used electrostatics to solve only the domain of air, not the domain inside of the conductor.

When I tried to do a surface plot on the sphere of surface charge density, I recreated the same error from my other model. I attached this simplified model to this post. It's a bit strange, but I'm guessing there's a fundamental error in my geometry or boundaries and materials set-up that is causing the error in both models.
Thanks so much for your responses! I tried the simplified model as you suggested, with a copper sphere inside of a box of air. I set the boundaries of the box to ground, except for one side which I gave a potential. I set the boundaries of the sphere to floating potential, as in my other model. I used electrostatics to solve only the domain of air, not the domain inside of the conductor. When I tried to do a surface plot on the sphere of surface charge density, I recreated the same error from my other model. I attached this simplified model to this post. It's a bit strange, but I'm guessing there's a fundamental error in my geometry or boundaries and materials set-up that is causing the error in both models.


Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 2012年7月31日 GMT-4 03:49
I ran your model, as is, and got the result you can see in the attached figure. I guess this what you (and I both) were expecting.

However, I'm using the latest Comsol version, 4.3, whereas you seem to be on 4.2a. Luckily, I still have the old version installed, and when I ran the model there, I could reproduce the error you described ("Failed to evaluate variable. [...] mod1.es.nD"). I would think this is (or was) a bug.

As to if and why there's a bug, you can only get a definitive answer from Comsol's developers, via Support. But if I was to venture a guess, it would be that, in the old version, the surface charge density was not defined properly on (some or all) interior boundaries.

I suspect that because, when you switch on "Equation View" (from the "Show" drop-drown menu in the Model Builder tab), you can see that the surface charge density is defined whenever you add a boundary condition such as "Ground" or "Electric Potential", but not for the "Floating Potential". This is different in 4.3.

The Equation View is also your way in to working around the bug in 4.2a. Just define a new variable (say, nD) on all boundaries, and copy the expression from the Equation View: es.unx*(down(es.Dx)-up(es.Dx))+es.uny*(down(es.Dy)-up(es.Dy))+es.unz*(down(es.Dz)-up(es.Dz)). By the way, this is just the defintion of the surface charge density, n·(D₁-D₂), at an interface between media 1 and 2 with normal vector n, expressed in terms of Comsol's internal variables. You can also plot the above expression directly.
I ran your model, as is, and got the result you can see in the attached figure. I guess this what you (and I both) were expecting. However, I'm using the latest Comsol version, 4.3, whereas you seem to be on 4.2a. Luckily, I still have the old version installed, and when I ran the model there, I could reproduce the error you described ("Failed to evaluate variable. [...] mod1.es.nD"). I would think this is (or was) a bug. As to if and why there's a bug, you can only get a definitive answer from Comsol's developers, via Support. But if I was to venture a guess, it would be that, in the old version, the surface charge density was not defined properly on (some or all) interior boundaries. I suspect that because, when you switch on "Equation View" (from the "Show" drop-drown menu in the Model Builder tab), you can see that the surface charge density is defined whenever you add a boundary condition such as "Ground" or "Electric Potential", but not for the "Floating Potential". This is different in 4.3. The Equation View is also your way in to working around the bug in 4.2a. Just define a new variable (say, nD) on all boundaries, and copy the expression from the Equation View: es.unx*(down(es.Dx)-up(es.Dx))+es.uny*(down(es.Dy)-up(es.Dy))+es.unz*(down(es.Dz)-up(es.Dz)). By the way, this is just the defintion of the surface charge density, n·(D₁-D₂), at an interface between media 1 and 2 with normal vector n, expressed in terms of Comsol's internal variables. You can also plot the above expression directly.


Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 2012年8月1日 GMT-4 12:05
Thank you so much for your help! This makes me feel a little better about myself since I couldn't figure out where I was going wrong.

I was having a problem defining my own variable (and with the variables unx, uny, unz), but plotting the expression manually gives me exactly the result I needed.

Thanks for helping me solve this problem! I hadn't done much with the equation view before, but now I see it's value.
Thank you so much for your help! This makes me feel a little better about myself since I couldn't figure out where I was going wrong. I was having a problem defining my own variable (and with the variables unx, uny, unz), but plotting the expression manually gives me exactly the result I needed. Thanks for helping me solve this problem! I hadn't done much with the equation view before, but now I see it's value.

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.