Diverging FDTD for high power in nonlinear GaAs photonic crystal unsolvable

- November 17, 2023 at 9:56 pm
  
  Jordi Lopez Leyva
  Subscriber
  
  I'm simulating a photonic crystal with GaAs of Chi2, surrounded by a material with n=1.6. I use a dipole source at the center with total amplitude = 1.83e-18. My mesh is uniform with dx,dy,dz=0.03um and dt stability factor = 0.1. Boundaries are PML stabilized with 100 layers.
  Lumerical suggests to check the type of instability by converting boundaries to metal, which I did, and the instability still presists, meaning it should be dt instability. However, my simulation is already very heavy (running with HPC). How can I solve the instability?
- November 27, 2023 at 8:41 pm
  
  Amrita Pati
  Ansys Employee
  
  Hi Jordi,
  I see that you are already using a quite low dt stability factor and the issue is not due to PML. Is this a broadband simulation? Sometimes the fitting of some materials cause the instability. If it is a brodband simulation, can you try doing the simulation just for a single frequecy point and check if you still see the divergence?
  Regards,
  Amrita
  - November 28, 2023 at 9:01 am
    
    Jordi Lopez Leyva
    Subscriber
    
    Hi Amrita,
    I've tried to use single frequency, I think simulation has progressed slightly further than before, but still diverges. Here's my source: