http://ab-initio.mit.edu/pipermail/meep-discuss/2006-April/000075.htmlIn my opinion, the computation cell acts as a cavity due to a small reflection of PML.The problem is that you didn't use the PML correctly: it needs to *overlap* your structure in order to absorb fields propagating in the dielectric material.
I also found that the PML affects the field inside the simulation box. In this figure http://www.cus.cam.ac.uk/~sis24/tmp/test.png (also attached), the evanescent field of a waveguide mode (extracted with harminv) decays the slower the closer the upper boundary to the waveguide surface. Since the region above the waveguide is occupied by air (default material) the PML should overlap it as mentioned above. The green lines correspond to a resolution of 10, and the blue ones arefor a resolution of 30. The (artificial) width of the interface with the air also has an effect of the field above the waveguide. It would be
nice to be able to vary the mesh size to resolve sharp features like this better. Any plans to implement this in Meep in the future?
test.png
Description: Binary data
_______________________________________________ meep-discuss mailing list [email protected] http://ab-initio.mit.edu/cgi-bin/mailman/listinfo/meep-discuss
