Reflectivity from an imported structure

guillem_maguillem_ma Member Posts: 1

Hi,

I am simulating the reflectivity from an SEM imported image. I can see some peaks that should not appear between 300-500nm . I defined the simulation as PML on Y and periodic on X. Any suggestion to get rid off those peaks?


Thanks,



Comments

  • greg_baethgegreg_baethge Posts: 148Ansys Employee

    Hi @guillem_ma,

    Thank you for posting your question on the forum. The first thing to check would be if the simulation reaches the auto shutoff min value. The auto shutoff represents the fraction of energy left in the simulation. Since FDTD is a time domain method, we use a Fourier transform to get the frequency domain fields. The auto shutoff allows to check if the fields have decayed enough by the end of the simulation so the Fourier transform doesn't show numerical artifacts. Typically, we want it to be below 1e-5. You can check this in the simulation log file.

    If it hasn't reached this value, you may need to increase the simulation time in the FDTD properties. Depending on the structure, the default value of 1000fs might not be enough.

  • guillem_maguillem_ma Posts: 3Member, Unconfirmed

    Hi @greg_baethge ,

    For that result I tried with 2000fs and I even reduced the auto shutoff to 1E-07. Any other suggestions? I can attach the file if needed

    Thanks,

  • greg_baethgegreg_baethge Posts: 148Ansys Employee

    Thanks for the update, @guillem_ma. Unfortunately, we're not allowed to download any file (see this post). Could you show the auto shutoff values from the log file? We're expecting something like:

    All processes are communicating and ready to proceed to simulation...
    Starting 734301 total iterations
    0.00653683% complete. Max time remaining: 1 hr, 5 mins, 39 secs. Auto Shutoff: 1
    100% complete. Max time remaining: 0 mins, 0 secs. Auto Shutoff: 6.87589e-06
    Early termination of simulation, the autoshutoff criteria are satisfied.
    

    Obviously, you can have much more lines that in this example.

  • guillem_maguillem_ma Posts: 3Member, Unconfirmed

    Hi @greg_baethge ,

    As you can see the auto shutoff is reached. Any other suggestions?


    Regards,

  • greg_baethgegreg_baethge Posts: 148Ansys Employee

    Thanks for checking, @guillem_ma. A couple of other things you can try:

    • In the FDTD setting, you can try the "steep angle" PML profile to increase their efficiency considering light can be scattered by the structure
    • From the screen copy. it seems the bottom boundary of the FDTD region is on the limit of the structure. It is usually better to have the structure going through the PML to make sure there is no artificial interface object/vacuum that would create some reflection

    Additionally, you can run some convergence testing to make sure the simulation settings are correct.

  • greg_baethgegreg_baethge Posts: 148Ansys Employee

    I forgot to mention, I assume you are using periodic boundary conditions. This can add some artificial feature in the spectra if the structure is not periodic (due the discontinuity at the edge of the simulation region). You can try and increase the width of the simulation region to reduce this.

Sign In or Register to comment.