Photonics

Photonics

How to set chi3 in the form of third-order tensor in the nonlinear simulation?

    • yli9003
      Subscriber

      Hello, 

      How to set chi3 in the form of third-order tensor in the nonlinear simulation?

    • Guilin Sun
      Ansys Employee
      Currently the full tensor is not officially supported. But the diagonalized tensor should work :

      Please try this if you have diagonal tensor Chi3.

      If you want off-diagonal tensor, you will need to develop your own plugin dll, and it is little supported from our side: Flexible Material Plugin Framework

    • yli9003
      Subscriber
      Thank you for your response. Third-order nonlinear susceptibility, as far as I know, are third-order tensors with 27 tensor elements. But why does FDTD default to a second-order tensor? For instance, to which tensor element does ¤çxx, ¤çyy and ¤çzz particularly refer?
      Furthermore, the FDTD website's tutorial discusses utilizing the attribute function to set the refractive index tensor with non-diagonal elements. Is this function useful for determining the ¤ç3 tensor?
    • Guilin Sun
      Ansys Employee
      You are right that the general 3rd order tensor has 27 elements. Lumerical's plugin material model can model 3by3 so it is simplified to xx,yy,and zz, meaning only ExEx,EyEy and EzEz terms are included.
      I do not know which tutorial you are referring. For anisotropy material, yes you can use Matrix Transformation attribute to set such materials. Basically it diagonizes the full matrix into diagonal matrix and transformation matrix. However, not all 3by3 matrix can be diagnozed, in particular for broadband. This attribute is only used for anisotropy materials, not for non-linear materials.
    • WenheJia
      Subscriber
      Hello In the third-order nonlinear process, I believe three electric field components must be considered. If only ExEx, EyEy, and EzEz terms are included, for example, SHG rather than THG or FWM can be obtained. Furthermore, I'm confused as to why ¤ç2 and ¤ç3 are in the same form in FDTD.
    • Guilin Sun
      Ansys Employee
      Yes you are right. However the current plugin material model only support diagonal case.
      I am not sure what do you mean "¤ç2 and ¤ç3 are in the same form in FDTD". They have different units and act differently with the fields. If you still have questions regarding this, please write a new post.
      https://support.lumerical.com/hc/en-us/articles/360034394734-Advanced-and-custom-optical-material-models-in-FDTD-and-MODE
Viewing 5 reply threads
  • You must be logged in to reply to this topic.