Photonics

Photonics

TE and TM Modes in a Photonic Crystal

    • schinta3
      Subscriber

      Hello,

      In the following example of a photonic crystal slab, setting a Z-min BC of symmetric and Electric Dipoles is said to excite the TE modes: https://support.lumerical.com/hc/en-us/articles/360041567454

      However in another example: https://support.lumerical.com/hc/en-us/articles/360041567754

      TE modes are exited by vertically aligned Magnetic dipoles and a symmetric boundary condition.

      While we aren't calculating the bandstructure in both cases, in the case where we do calculate the bandstructure, so we need to 'cast' randomly oriented magnetic dipoles/TE and electric dipoles/TM?

    • Guilin Sun
      Ansys Employee
      yes, you are right that E- and M-dipoles use different symmetry BCs, which is the root-cause of symmetric and anti-symmetric BCs. The structure is symmetric to the axes but it is the polarization that determines which symmetry BC it should be used. There is theory behind, so please do not "randomly" set them.
      For some details please refer this page: Symmetric and anti-symmetric BCs in FDTD and MODE
    • schinta3
      Subscriber
      Sorry, I was unclear.
      For TE modes -> Always use Magnetic Dipoles and Symmetric conditions
      For TM modes -> Always use Electric Dipoles and Anti-symmetric conditions
      The random casting I was referring to is EITHER of the above options, but the orientation/position of the dipoles is random within the slab unit cell to ensure all modes are excited.
      Is this Correct?
    • Guilin Sun
      Ansys Employee
      Usually it is ok, as long as the dipole is in the symmetry plane.
      if not, simulating band structure is ok.
      If you want to simulate other quantities, such as dipole radiated power from a structure, Purcell factors etc, it will not be ok. This is because, when symmetry BC is used, it flips the simulation region. So if the dipole is not on the symmetry plane, say off set, then you will have TWO or even FOUR dipoles at the simulation volume. and the result is not additive, so the result can be wring.
      In summary, when the intensity/power is required in the result, only one dipole should be simulated w/o symmetry. Bandstructure simulation does not need accurate intensity so it is ok.
    • schinta3
      Subscriber
      Hello One more question. We are simulating bandstructures for rectangular Photonic Crystal slabs. Would Bloch BC in x-y and PML for both z BCs with electric dipoles ensure TM modes, and likewise with electric dipoles ensure TM Modes? Or is a symmetry or antisymmetry condition needed?
    • Guilin Sun
      Ansys Employee
      For such PC slab, please follow up the online examples. Both e- and m- dipoles can excited modes, but one may excite one polarization more than the other; and the symmetry BCs are used to separate/isolate the band structures for TE and TM modes. As long as you can find proper band structure, it does not really matter what dipole type you are using. But the symmetry BCs matter.
      In summary, it is the symmetry BCs that separate the the TE/TM modes or band structure.
Viewing 5 reply threads
  • You must be logged in to reply to this topic.