Guilin Sun
Ansys Employee
Please note that all the parameters you mentioned may affect the result, so even if you sweep them it does not guarantee the accuracy for another device simulation. In addition, those parameters will need to be nested-sweeps . However there are parameter combinations which are wasted" for example, you use higher mesh accuracy but fewer PML layers and smaller PML distances. They are knowingly not effective for PML performance. Considering varFDTD is a variational method and its accuracy should be verified by 3D FDTD, I would suggest that you do the following:
1: choose PML at least half wavelength
2: roughly measure the PML thickness about half wavelength. You can estimate the thickness by definition of mesh accuracy: mesh accuracy 2 means 1/10 of the smallest wavelength INSIDE the material;
3: sweep mesh accuracy only (2,3,4 should be enough)
Please try those modification for varFDTD, and then use 3D FDTD to verify if the result is acceptable from varFDTD.