Discussion Forum

It's a 3D simulation, there are two Bragg mirrors on each side of the cavity, the geometry is in the attach file[3d geometry].[ Bragg mirror is just alternative silicon and air with thickness of λ/4n , and the cavity in the center has a thickness of λ/2, air] , here the λ is the wavelength of f=100GHz, which I want to resonate.
Actually, I did the simulation in 2D first, and get a very close result as I expected, which is also included in attach file.[2D simulation result]

But I just do the same simulation in 3D, here comes the problem.
As in 2D I set the boundaries of the "waveguide" to be PMC,
in 3D, I set all the boundary surface to be PMC,
Also, in both simulation, I use numeric port in the front and back surface.

Since I just use the same method, it's wierd to see a simulation result in 3D is totally different as 2D.

So my question is
1. Can I set numeric port in this problem
2. Can I make all surface to be PMC(except front and back port) in 3D, if not, what should I set
3. Should I add air sourronding the waveguide? Actually I don't think I need because in 2D simulation, I didn't do that and get a pretty good result
4. Since from 2D to 3D, the length of the "waveguide" changes from infinite to a number I set(2λ), should I increase it?
5. Because it takes so much time, I'm wondering if I can take the advantage of symmetry:
I just cut the waveguide into 1/4, then set the cut plane to be PMC( because I saw somewhere saying if it's a symmetric Electric field, I can set PMC), but if I do this, still, all the boundaries are PMC