Learning Forum

[Jeff]

Hello:

For simulating scattering (side diffraction) efficiency of a 2D photonic crystal with limited X/Y dimention, would it be possible to use a Gaussian beam source instead of a plane wave and if yes, do we have to set the similar source structure as the TFSF, i.e., a source enclosing the whole object XY area (here the 2D photonic crystal)? Thank you!

[Kirill]

Hello Jeff,

It’s definitely possible to achieve, but the setup details will depend on whether you want to eliminate the non-diffracted light or not, as both options can be valid.
To speed up your progress, I suggest browsing through the Scattering - list of examples on our website.
Also you might want to start with the following resources, as they could have some relevant discussions:
- DVD surface scattering (DGTD)
- Bidirectional scattering distribution function (BSDF)
- Photonic crystal cavity
- Far field projections from a box of monitors

Best regards,
Kirill

[Jeff]

Thank you very much Kirill! I'll browse the sections as you suggested. While you mentioned that it's valid to have either plane wave or Gaussian beam as the box source (TFSF) for modeling side scattering/diffraction, I seem not able to find out how to replace a plane wave with a Gaussian beam for a TFSF box source. Could you guide me on how to set up a Gaussian beam box source like a plane wave TFSF? Also, could you elaborate a bit more on the meaning of "...the setup details will depend on whether you want to eliminate the non-diffracted light or not..."? Thank you again! 

[Kirill]

Dear Jeff,

• The article Tips and best practices when using the FDTD TFSF source states that “The TFSF source is an advanced version of the plane-wave source designed primarily for particle scattering simulations.” Therefore, you cannot simply replace a plane-wave source with a Gaussian beam inside TFSF. It has to be custom simulation.
• The main advantage of the TFSF source is the ability to analyze the scattered field without the presence of the excitation field. However, not every scattering problem requires this approach. For example, in the case of diffraction gratings, all diffracted light is analyzed without needing this feature. Please check if custom TFSF is relevant to your specific simulation.
• In most cases, a plane-wave source should suffice to approximate a slowly varying Gaussian beam profile, but this may not apply to your specific scenario.
• If you still wish to implement a custom TFSF source, the main idea is to run two simulations: one without the scattering object and another with it. By subtracting the fields, you can isolate the scattered field. Note that it is crucial to ensure both simulations use the same mesh. Please also check the following threads: Ansys Insight: Custom TFSF source, scattering calculation with gaussian beam in FDTD.

Best regards,
Kirill

 

 

 

[Jeff]

Hi Kirill,

Thank you so much for the detailed explanation! Then is this understanding correct: we need to consider using a TFSF source in modeling only if we are to calculate the scattering in the forward (transmission) direction? In my current case, I need to calculate the scattering/diffraction in the side/transverse (not transmitting) direction from a 2D volume crystal structure, then a TFSF source is not needed? Thank you again!

[Kirill]

Hello Jeff,

Yes, you mostly need the TFSF approach when your signal overlaps with a propagating (or reflecting) excitation wave in the near field. If you run a simulation with just the excitation Gaussian beam (with the structure disabled) and get near-zero signal in the direction of interest, then you can obtain meaningful results without using a TFSF-like source.

Best regards,
Kirill