[MODE EME] Results are different with different number of cell groups

[ke-chun_yeh]

Thank you for contacting Lumerical support. Before proceeding, please note that this is a public forum and anything you post will be visible to the general public. For users residing in the United States, please ensure that the files and information contained within your posts are not International Traffic in Arms Regulations (ITAR) or export controlled.

To help us solve your issue as quickly as possible, please provide the following information:

1. What device are you trying to simulate? Include diagrams if available.
Laser with a taper Si waveguide underneath to couple light from laser to silicon platform

2. What results are you trying to obtain? Be as specific as possible.
Transmission efficiency (S21^2)

3. Description of the problem or issue.
The tapered Si waveguide can be divided into three part: width (1) increases linearly at the input, (2) keeps the same in the middle, and (3) decreases linearly at the output. And I used to set one cell group for the whole structure, as below.

set(“number of cell groups”,1);
set(“group spans”,xw2); #xw2=the total length of the structure
set(“cells”,[10]);
set(“subcell method”,1);

I’m wondering if the S21^2 I got is wrong for my setup, because the S21^2 is different from the one obtained from 3 cell groups. Is the setup of 3 cell groups necessary to get the correct S21^2?

Thank you for your help!

4. Lumerical product and software version.
Lumerical 2019b MODE

5. Please attach your simulation files, including any logs.

[kjohnson]

Hello @ke-chun_yeh,

@ke-chun_yeh

I’m wondering if the S21^2 I got is wrong for my setup, because the S21^2 is different from the one obtained from 3 cell groups. Is the setup of 3 cell groups necessary to get the correct S21^2?

Setting multiple cell groups is not required, however in this case it would be a good idea. This is because multiple cells will be required along the tapered regions, however only one cell is required in the middle where the cross-section does not change. With only one cell group the spacing of the cells would be uniform across the whole structure. Using cell groups you can place multiple cells along the tapers and apply the CVCS subcell method, while placing only one cell over the middle region. By placing multiple cells only where they are required your simulation will be more efficient.

Please let me know if you have any questions.

[ke-chun_yeh]

Hi,

Got it. Thanks for your help!