I've been working to model NIR waveguiding of a thin film on silicon. To implement this, I've created 2D models of an air gap followed by a thin dielectric film on a silicon wafer in the wave optics module and the RF module. I've placed an electric dipole point source in the thin dielectric film.
As I change the meshing, the magnitude of the electric field in the device changes significantly. Playing around with meshing, it seems to exclusivley change due to the meshing around the dipole. Currently, the meshing is much less than the wavelength of light in any media (~50x less), so I don't believe that the mesh frequency is the problem. Indeed, the problem continues to get worse the finer I mesh my system.
I've tried using a PML and perfect electrical conductor layers as boundary conditions and implemented the model in wave optics and RF modules.
Is there something obvious that I'm missing?
Edit: For whatever reason, using an out-of-plane line sources seemed to significanlty fix my problems. When I used an electric field dipole, there were wild reflections in the PML layer and I'm not entirely sure why. Does anybody have advice? If possible, I'd like to use the electric field dipole so that I can test different polarizations.