Discussion Forum

You can have your patch antenna and computational domain (vacuum, I assume) surrounded by perfectly matched layers. Then it is possible to solve for the eigenfrequencies and eigenmodes (or characteristic modes). Without further information (a model file the best), it is hard to see where your problem is.

Hi Pu,

Thanks for the reply. I am attaching my model file. Characteristic modes are supposed to be real values, and I am getting imaginary eigenfrequencies. Also, do you know how to export the eigenvectors for the corresponding eigenvalues?

Jose Santos.

I have a brief look at your model. First of all, I don't understand why you are using a 2D model. Shouldn't a patch antenna be a real-world 3D thing? And the eigenfrequencies actually should be complex, as the modes are loosing energy through radiation. About exporting eigenvectors, I'm not sure what you meant.

Best

Hi Pu,

I was trying to model the patch antenna as a crossection modeling only the conducting element. I undestand now that the eigenfrequencies should be complex, but according to documentation, the eigenfrequency and the eigenvalue study should be similar. However, I am getting different results with them. Do you have any suggestion of how I should model this? Also, where can I find more information about the eigenfrequency solver and its physical interpretation?

Respectfully, Jose Santos.

I don't get the cross section modeling thing. Can you explain it a bit?

Eigenfrequency and eigenvalue modules are indeed the same in essence. The results should be same. The only differences are some definitions.

I'm not aware of COMSOL calculation of characteristic modes, but the software is commonly used to calculate quasi-normal modes of metallic nanostructures at optical frequencies (similar to your study). You can find relevent information on the internet.

Hi Jose,

Have you found any way to export the eigenvectors for the corresponding eigenvalues?

Best,