Learning Forum

[iqnlo]

Hi All

I'm designing a Fiber-chip coupler and find a wired thing.

From this example, https://support.lumerical.com/hc/en-us/articles/360042305354-Edge-coupler

The definition of total coupling efficiency is mode overlap between fiber mode and the coupler mode times the adiabatic mode transfer efficiency from coupler tip mode to the waveguide mode.

In the example project, there is no problem, the mode over lap is about 90%(or higher) and the total coupling efficiency is about 85%. So we can know that the adiabatic transfer efficiency is also higher than 90%.

When I make some change on the structure, something wired happened. For example, deleting the three silicon nitride layers. The mode overlap will drop greatly, to about 50% (without fiber position optimization). However, the transmission efficiency (S21^2) is 70%. This is wired because follow the principle of coupler, the 70% should come from 50% times the adiabatic transfer efficiency. Even if the second efficiency is 100%, the result should not be 70%.

Do someone know where this problem comes from? I'm really confused.....?

[Guilin Sun]

This is a more theoretical question than simulation. The published result is optimized to excite only one mode, so using the above definition is correct. When you have higher transmission efficiency, it may indicate that high modes are excited and somehow the higher order modes convert to the fundamental modes, during their propagation through the taper. I would suggest that you do more mode analysis, and also using higher order modes to test the taper its self to see its conversion. Please make sure the mode source is not truncated by the applied boundaries.

[iqnlo]

Hi
Thank you so much for your reply. May I ask some further questions?
As you said, I need to do some mode analysis and use the higher-order mode to see the conversion.
May I know how to do these tests? Could you please show me some examples so that I can make sure I understand them correctly? Thanks a lot!

[Guilin Sun]

You could use mode expansion monitor, or port to check if higher order modes are excited and converted to the desired mode. You can use additional two sets of ports in different locations along the propagation after coupling, and check if the higher order modes are excited and where they go. In addition, overlap does not equal coupling, even though in many cases they are almost equal.