Discussion Forum

If I were doing it, I would probably model the cell phone by a simple small dipole antenna. Comsol has an example, called dipole_antenna.mph, in the Application Library (under RF Module, Antennas). You don't need to bother with the perfectly matched layers in that example. Just look at how they set up the source (which uses a lumped port). Note that you can make the antenna smaller if you want; it will simply radiate less efficiently, which is probably fine for your purposes. And it is pretty easy to figure out how much power is being radiated. I hope that helps. Good luck.

Many thanks dear Koslover, Your advice is really clear and helpful; I've found dipole_antenna.mph and I dragged it into my big COMSOL project. Please, because I'm a beginner with COMSOL, Would you please help in these too:

1- When minimizing the size of antenna to be suitable for my project, should I minimize each element in the Geometry (cylinders, sphere, ...) or just part of them?
2- From my foreknowldge, I knew that I must make the area in which the antenna will radiate Perfectly Matched Layer (PML) to abosorb all the radiated power from the antenna; So, should I make all the room boundaries PML? 3- I want to study the effect of the radiated field from the antenna on very small objects (in micrometer scale); when simulating those micro-objects, I can't see them when they compared to the dimension of the antenna; So, How to let COMSOL show me those micro-objects when studying on them?

I appreciate so much your kindness and support. Many Blessings. Kyrillos

1. In regard to antenna size, choose a size that is not so large that it needs an excessively large number of mesh elements to represent, but not so small that it needs excessively tiny (relative to the rest of the problem space) mesh elements either.
2. You probably don't need to use PMLs at all. Just use scattering boundary conditions on the room walls, asuming you want them to have a ~minimal impact on the model.
3. Your comment about the micrometer scale is unexpected. Are these isolated micrometer-scale objects? If so, unless you are in the reactive field of the subject antenna (< lambda/2pi away), your "very small objects" will effectively see plane waves, locally. In that case, why bother with modeling a dipole source (or other antenna) at all? Instead, use Comsol's "scattered field" formulation, and choose a "background wave type" as linearly polarized plane wave. You can find more about doing that in the help system as well as elsewhere in this forum. But frankly, if you don't really care about the "room" itself, and your actual volume of interest is sufficiently small compared to a wavelength (i.e., if all you really care about is modeling how the isolated micrometer-scale objects behave when exposed to oscillating EM fields, and you have no real interest in the EM waves per se) then you might want to consider using the AC/DC module instead (after all, it will let you specify whatever frequency you want). In that case, you could set up your volume to yield either a (more or less) uniform oscillating E field or H field, depending on the physics most relevant to your phenomena of interest. And with a small volume, you wouldn't have trouble "seeing" your small objects in it. Let the physics of your problem guide you toward making the correct modeling choices. There would be big differences in choosing an approach to modeling, for example: (1) a desk covered with a stack of electronics boards with tiny microcircuits on each of them, (2) a cloud of metallic dust in a machine shop, or (3) a single speck of pollen, floating in the air. But it would seem that all of those fit your problem description.

Dear Koslover, My deep appreciation for your most detailed explanation and patience toward me; I'm so grateful to you.

Kyrillos