Far Field Pattern


#1

Good afternoon;

We are performing 3D FDTD simulations of a SOI grating antenna to support our experimental measurements. From both theoretical analysis and experimental characterization, we know that the maximum radiation angle should be between 16 and 19 degrees but, when simulating it ithrough the farfield3d function, we get a farfield pattern whose direction of maximum radiation is 40 degrees. Is this due to simulation size, monitor size , or mesh size? Or that the monitor is too far away from the source?. Annex an image of the near-field, far-field and perspective view simulation.




#2

@jlpitar

From the near field image, the minimum value of the colour bar doesn’t go down to 0, so this suggests that the near field monitor span is not quite large enough to capture all of the fields, and this could cause some artifacts in the far field projection as described here:
https://kb.lumerical.com/en/index.html?solvers_far_field_projections_far_field_filtering.html

To remedy this you can increase the simulation region and monitor’s x and y spans to increase the collection angles of the monitor. Moving the monitor closer to the structure has the same effect of increasing the collection angle.

However, as you also mentioned, the mesh step size could be too coarse and this could also cause the results to be inaccurate. I would also suggest using a mesh override region over the grating pattern to specify a finer mesh over the structure and performing some convergence testing of the mesh step size to see if using a finer mesh gives better results.

Some additional things you could double check are to make sure that the simulation time is long enough so that the auto shutoff threshold is reached which is discussed here:

And you can also check the mode source to make sure you are injecting the expected waveguide mode.

Let me know if you are able to resolve the issue!


#3

Hi,

attending the tips, we increase the simulation region and monitor’s x and y spans to increase the collection angles of the monitor and, we moved the monitor closer to the structure. Also, we are using a mesh override region over the grating pattern to specify a finer mesh over the structure. and, in adittion, we modified the internally check the auto shutoff conditions every 50 time steps of the simulation, but the results are inaccurate.

Any other suggestions?, we really don’t understand why the results are so different.


#4

@jlpitar

It could be possible that there are some differences between the expected structure and the simulated one such as the material data, period or geometry of the structure, or the wavelength which is projected compared to the wavelength of the source in experiment. It’s difficult to say without having a copy of the simulation file and description of the intended device.

One thing that may be helpful is to first simulate a simpler 2D grating structure similarly to what’s done here:
https://kb.lumerical.com/en/index.html?pic_passive_grating_coupler_2d.html

This would be faster to simulate making it faster to test for troubleshooting the cause of the discrepancy, and you should be able to calculate an expected propagation angle using the equation on slide 10 in the presentation for this video about grating coupler design:
https://www.lumerical.com/solutions/communications/video/silicon_photonics_grating_coupler_design_video.html

Hopefully that will help. If you need further assistance you could also attach a copy of your simulation file.


#5

From both theoretical analysis, comsol simulation and experimental characterization, we know that the maximum radiation angle should be between 16 and 19 degrees. Annex the simulation file.grating_antenna_lumerical.fsp (2.8 MB)


#6

@jlpitar

I checked your simulation file and didn’t find any errors with the setup. I then tried to calculate the expected radiation angle based on equation 2 of the following paper:
http://www.mina.ubc.ca/files/UGC_v3.pdf

This is actually the same method as used in the grating coupler design video here:
https://www.lumerical.com/solutions/communications/video/silicon_photonics_grating_coupler_design_video.html

In the calculation I used the effective index of the grating teeth which is calculated using a mode source in a 2D simulation region (neff ~ 2.8), and I used the effective index of the grating slots as 1 since there is no Si layer in the slot region to support a guided mode. The resulting angle that I calculated was about 38 degrees which is close to the angle from the far field projection.

The file to calculate the neff of the tooth is here:
neff_tooth_calculation.fsp (241.0 KB)

I’m not sure if my calculation method is completely correct, so if you had a different method to calculate the theoretical radiation angle of the device which generates a different result please let me know!