Learning Forum

[agranmoe]

 

Hi there,

I am trying to design an evanescent coupler consisting of a linearly tapered GaAs waveguide resting on top of a high-aspect ratio SiN waveguide all buried in SiO2. I would like to use the EME solver so that I can efficiently vary the taper length and save lots of computation time. To verify my EME results, I ran the same setup in FDTD and receive very different transmission results (transmission from GaAs fundamental TE mode to SiN fundamental TE mode). In FDTD, I measure this using the expansion in the output port and in EME I measure this using the S21 scattering matrix element. I will list the various things I have checked in both my simulations below. Please help me figure out why these methods give me such drastically different results, thanks!!

Thanks in advance,

Austin

FDTD:

• Doubled simulation time and increased mesh accuracy from 2 to 3 and get same results.
• Mesh overrides for both waveguides in the Y and Z directions were applied to ensure accuracy.
• No modes are cutoff by PML boundary when looking at mode profiles via ports

EME: 

• Ran script to verify number of cells in tapered region is sufficient.
• Mode convergence sweep looks normal.
• Checked all diagnostics Lumerical suggests and found no issues apparent to me.

Simulation parameters of interest:

Target wavelength: 950nm

GaAs waveguide dimensions: 250nm wide tapered down to 100nm over 40um, 150nm thick. Refractive index 3.6. Mesh override: dy = 0.04um, dz = 0.02um

SiN waveguide dimensions: 2um wide, 40nm thick. Refractive index 1.943. Mesh override: dz = 0.005um

SiO2 refractive index: 1.4511

 

 

 

 

[Dev]

Hello, 

Please refer to the following:

Why am I seeing a significant difference in coupling efficiency from EME to FDTD? (ansys.com)

Difference in results between FDTD and EME (ansys.com)