PML zmin and zmax setup for gaussian source

emh711emh711 Member Posts: 12

For a gaussian source simulation, is it recommended to have zmin and zmax be at lease one wavelength away from the simulation object?

Best Answers

  • chris_kopetskichris_kopetski Posts: 59Ansys Employee
    Accepted Answer

    Is Z the direction of propagation of your source, as shown here?


    If so, the main consideration of the Z location of the boundary condition is the distance from the structures. They should be at least 1/2 a wavelength from the structure.

    For X/Y, the 1/2 wavelength rule also applies. Additionally, it's important to make sure the XY span of the simulation is large enough to 'fit' the entire guassian beam. Having the boundary conditions truncate the beam will lead to errors.


  • kjohnsonkjohnson Posts: 219Ansys Employee
    Accepted Answer

    Hello @emh711 ,

    There are a couple of considerations for the simulation span in a simulation with a Gaussian source. First, the boundary should be at least a half wavelength away from edges of geometry objects. This is to prevent evanescent fields from interacting with the boundaries. Note that this does not include objects that pass all the way through the boundary, for example substrate or cladding objects.

    Second, the span of the simulation should be large enough that the edges of the Gaussian source are not truncated by the boundaries. The amplitude of the Gaussian source should reach a very small value, for example 10^-4, before it reaches the boundary (note that this also applies to the span of the source object itself). As long as this is true, the edges of the Gaussian source object itself doesn't need to be any distance away from the boundary. In other words, the span of the source can match the span of the simulation region.

Sign In or Register to comment.