Learning Forum

[jasper]

Hi,

I am trying to simulate the two following bragg gratings, the first being a corrugated ('standard') bragg grating and the second being a 'post' bragg grating:

After a few attempts, I have managed to correctly simulate the corrugated grating with the characteristic results. However I am stuck simulating the second ('post') grating. One of the things troubling me is that the mode convergence sweep doesn't converge:

It seems to me that since the structure is different, multiple modes are present in the post section of the 'post' bragg grating, in contrast to the corrugated grating which is engineered to only carry a single mode in both sections (waveguide and corrugated). Could the EME solver have trouble with that? I have tried limiting the # of modes searched in the EME solver settings, but this doesn't fix it (also using 50 instead of 6 modes):

I have tried using PML (to avoid the unphysical modes being found with metal boundaries), but that didn't change much.

I made sure to select the correct (Fundamenta TE) Modes at both ports. 

 

I'd be very grateful for some piece of advice!

Thanks in advance,

Jasper

 

[Guilin Sun]

EME is based on the continuity of the tangential fields. Thus a proper number of modes at each cell is needed. If the modes do not match well then such "inconvergence" can happen. Unfortunately we do not know exactly how many modes are "proper". This may depend on your understanding of the device with more testings, and check the mode profiles, diagnosis results to modify the settings. For higher order modes in the combined waveguide (posts and center waveguide), modes are more easily degenerated and can appear in different locations. This increases the challenges. If you know such device will not induce TM modes, you may try to use symmetry BCs (Symmetric or anti-Symmetric, depending on the device): https://optics.ansys.com/hc/en-us/articles/360034382694-Symmetric-and-anti-symmetric-BCs-in-FDTD-and-MODE  

[jasper]

So if I understand this correctly, I must manually set the modes in the EME cell group 2 (multi-mode) settings so that all the proper modes (and no unproper modes) are used? I have tried that and still get erroneous results. I have also looked at the error diagnostics tool and found that the E2/H2 tangential field discontinuities have the following values:

In this post (Understanding EME error diagnostics – Ansys Optics), this is supposedly to check if enough modes are being used. But as shown in the screenshot above, when using more modes the sweep doesn't converge. What can you recommend to fix this issue?

[Guilin Sun]

As said previously, EME is based on the continuity of the tangential fields. If obvious discontinuity exists it indicates that it needs more number modes, propagating and non-propagating modes. When the waveguide has discontinuity, eg sharp change of the cross section, it will needs non-propagating modes to keep the tangential fields constant. If the non-poropagating modes are continious such as the radiating modes, you may need more number of modes. It is very challenge in some cases to choose proper number of modes. While in theory infinite number of modes will work, in reality due to several reasons more tests on FINITE number of modes are necessary. It is true when the cross section changes significantly.