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 The information provided may be out of date.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.

Far field of an optical grating

Please login with a confirmed email address before reporting spam

Dear All,

I'm studying what happens when a gaussian beam impinges a lithium niobate sample in which it is written a refractive index pattern; the sample can rotate respect to the beam direction, and I want to compare the simulated far field with the experimental data. The comsol model (which is attached to this comment) follows more or less the nano-particle tutorial; unfortunately I see much more peaks as expected and I think that I'm missing something about the far field calculation.

It's worth noting that if I remove the lithium niobate sample and I leave air instead, the computed result shows, as expected, the bare gaussian beam (normE) BUT the far field is not a gaussian (as it should be because the Fourier transform of a gaussian is again a gaussian)! This strange behaviour suggests that I'm not using the far-field tool in a proper way...

Thanks in advance for any suggestions.

Nicola


3 Replies Last Post 13 févr. 2014, 13:04 UTC−5
Sergei Yushanov Certified Consultant

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 13 févr. 2014, 09:25 UTC−5
Nicola,

A couple notes on your model:

- PML layer is too close to the sample – increase radius of the PML domain
- Far-field domain should include all air domains around the sample
- It is better to mesh PML domain with swept mesh and set 5-6 elements through the PML thickness

Best regards,
Sergei
Nicola, A couple notes on your model: - PML layer is too close to the sample – increase radius of the PML domain - Far-field domain should include all air domains around the sample - It is better to mesh PML domain with swept mesh and set 5-6 elements through the PML thickness Best regards, Sergei

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 13 févr. 2014, 12:26 UTC−5
Dear Sergei,

thanks for you suggestions; I attach a new model (again after stripping meshes and solutions). Far field is changed but again when I remove the sample the gaussian beam has not a gaussian far field. I have to say that I had also some problems to fully apply your suggestions; in particular:

- PML layer is too close to the sample – increase radius of the PML domain
Ok, I did it but with care because otherwise the model grows too much for my machine.

- Far-field domain should include all air domains around the sample
Ok, easy. In the nano-particle tutorial the air domain was only one, so I miss to understand that I have to select all the air domains;

- It is better to mesh PML domain with swept mesh and set 5-6 elements through the PML thickness
For what I know, swept can only be applied to 3d models; after trying many ways, actually I put a free triangular mesh with smaller "maximum element size";

In conclusion, I still not like the results with the bare gaussian beam model (without sample), but it seems that the model geometry is the most critical factor.

Best regards,

Nicola
Dear Sergei, thanks for you suggestions; I attach a new model (again after stripping meshes and solutions). Far field is changed but again when I remove the sample the gaussian beam has not a gaussian far field. I have to say that I had also some problems to fully apply your suggestions; in particular: - PML layer is too close to the sample – increase radius of the PML domain Ok, I did it but with care because otherwise the model grows too much for my machine. - Far-field domain should include all air domains around the sample Ok, easy. In the nano-particle tutorial the air domain was only one, so I miss to understand that I have to select all the air domains; - It is better to mesh PML domain with swept mesh and set 5-6 elements through the PML thickness For what I know, swept can only be applied to 3d models; after trying many ways, actually I put a free triangular mesh with smaller "maximum element size"; In conclusion, I still not like the results with the bare gaussian beam model (without sample), but it seems that the model geometry is the most critical factor. Best regards, Nicola


Sergei Yushanov Certified Consultant

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 13 févr. 2014, 13:04 UTC−5
Nicola,

Yes, for 2D geometry Mapped mesh should be used for PML (not swept). Attached is your model with mapped mesh in PML domain.

To understand what is going on, I would suggest to increase wavelength and play around with the model parameters and setup until you get physically sound solution. This would speed up debugging process.

Best regards,
Sergei
Nicola, Yes, for 2D geometry Mapped mesh should be used for PML (not swept). Attached is your model with mapped mesh in PML domain. To understand what is going on, I would suggest to increase wavelength and play around with the model parameters and setup until you get physically sound solution. This would speed up debugging process. Best regards, Sergei

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.