On Thu, 31 Jan 2008, Li, Jingjing wrote: > I'm learning MEEP and here is a very simple problem which gave me a > surprise. Considering a structure composed of 3 planar layers of > dielectrics, say, epsilon=3, 2, 3, respectively, and we want to study > the plane wave transmission/reflection at normal incidence. The layers > are inifinite in y-z plane thus the problem is basically a 1-D problem. > Of course we can simulate it in MEEP using the 1D feature. We can also > construct an equivalent 2D problem by considering a planar waveguide of > certain height (say, sy) composed of two perfect electric conductors > placed at +/-sy, and place the 3 planar dielectric layers inside the > waveguide. A plane wave can be launched and propogate in this > waveguide. We can also construct a 3D equivalent problem, by > considering a waveguide with PEC walls at y dirrection and PMC walls at > z direction.
Jingjing, The right way to set up this problem in 2d and 3d is to use periodic boundary conditions in the yz directions, since this is the structure you actually have. You are correct that the use of PEC/PMC boundaries will theoretically give you the same solution for this particular polarization of input light, but periodic boundaries are better because they are independent of polarization (and are more conceptually accurate). Apparently, you're running into some numerical instability for the combination of PEC/PMC boundaries; I haven't seen this before, but the periodic boundaries should work. Steven _______________________________________________ meep-discuss mailing list [email protected] http://ab-initio.mit.edu/cgi-bin/mailman/listinfo/meep-discuss
