Discussion Forum

Hi,

I am trying to simulate using the Rf module TM in 2D an infinite structure in the x dimension. (comsol 3.3)
Therefore I'm getting used to the periodic boundary conditions and the floquet bloch phase shift.

I simplified my problem to a very simple one: a plane wave propagating with wavevector k=kx*ex + ky*ey in an infinitely long space in the x dimension.

My domain is L wide and H high.

I've set scattering boundary condition on top (source of plane wave side) with a field exp(i*kx*x).

I've set scattering boundary condtions at bottom, no field.

I've set right and left boundraies to PEC.

Then I've put left and right boundaries as source and destination in the periodic boundary conditions, setting 2 contraints on Hz and psi, namely Hz => Hz*exp(-i*kx*L) and psi => psi*exp(-i*kx*L).

I've put the 4 corner vertex as sources and destinations as well.

Finally I've set the "Use hermitian transpose option..." in the solve advanced parameters.

Then I solve for 1 wavelength and several angles and can get a correct plane wave propagating as if the domain was infinite...

Is my floquet phaseshift wrong? Did I set a wrong boundary condition somewhere?

Any help greatly appreciated.

G.