Learning Forum

[Joel Sved]

Hi, 

I've been running a simulation of a topology-based 1x10 power splitter with 1um air cladding, 220 nm Silicon slab (with air holes) and SiO2 substrate. (operating wavelength 1550 nm, TE mode source)

I am importing the silicon metasurface as an (n,k) material, the 2D effective index is approx 2.8 given the cladding and substrate materials. 

After optimisation, I import the gds to verify the 2D simulation result, which looks fine. 

When I then run the 3D simulation with the updated Si effective index of approx 3.48, the E-field result looks completely different. 

I'm not sure what I am doing wrong, the only difference is I am changing the effective index of the Silicon slab and waveguides. 

Thanks very much for your help, 

Joel 

[Guilin Sun]

You should try to inverse design in 3D, since 2D means infinitely long in 3rd dimension whereas the actual device has very small 3rd dimension size. You can use the 2D result as a starting point.