Some problems about simulating etched diffraction grating(EDG) by using varFDTD


Hi all,
I try to simulate etched diffraction grating(EDG) by use of varFDTD. The EDG structure is based on the paper Planar Concave Grating Demultiplexer Fabricated on a nanophotonics SOI platform.pdf (355.7 KB)


However, the propagation pattern is not correct. Only a small fraction of light is transmitted into the output waveguide.


I am not sure where the problem is. Could anyone help me solve the problem?

The varFDTD file is attached here. Echelle_grating.lms (342.7 KB)


Hi @lishifeng

Sorry for a late reply.

I have checked your simulation file and it is set properly. I did not check the geometry on the mentioned references, but you may double check the geometry to make sure that it is correct. By looking at the profile monitor, I can see some coupling to the waveguides, but there are also scattering in the -y direction both from gratings and facet of waveguides. Does reflection from grating make sense?

You can also put a monitor in front of the first waveguide and use apodization to measure the power of the input waveguide. You can then normalize your results to this power to avoid the reflections that occur in the facets.


Thanks for your reply.
EDG is based on the diffraction of grating. The problem here is that only a small fraction of light (2%-3%) is couped into the output waveguide. But there is about 20-30% of the light coupled into the output waveguide in the reference.

I am not sure how to set the apodization of the monitor. Do you have an example?


Hi @lishifeng

In your case, we are interested to measure the light that is injected from the source at the output of Input_wg in the area the is shown by green light in the figure below:


We cannot measure the entire power that passes through the monitor as there will be some reflected power from gratings if we include all the simulation time. Thus, apodization can be used to measure the results at a specific time to remove these reflected power. This is important because light is reflected at the facet of input waveguides, and authors may have reported the power results right before entering the output waveguides.

You can choose “end apodization” (as you are only interested in the input power and want to remove the reflected power) for the input waveguide and “start apodization” for the output waveguides and then normalize the results w.r.t. input power obtained from the monitor.

The link below might be also useful for you to review:

I hope this was helpful.


Thanks. I will try as you suggested