treating quasi-TE mode in overlap with the mode of an optical fiber about femwell HOT 1 CLOSED

On the first glance it makes sense to me what you're doing.
But we should first benchmark this example to some literature and reproduce their results to make sure we didn't miss something in the implementation. (I'm wondering right now if we need to square the overlap)
https://doi.org/10.1109/JPHOT.2016.2600037 would probably be a good paper for comparison.

The assumption that the field is only in E_x/y is also something to double check, but according to https://apps.dtic.mil/sti/pdfs/AD1124366.pdf it should be fine. Whatever, eventually we should also consider having some distance between the edge coupler and the fiber (as in reality we will for sure have some gap there), so we'd need to add z in the equation for the Gaussian and I'd guess we'd need to calculate for that full-vectorial.

from femwell.