Different results when simulation rerun

- March 1, 2024 at 2:00 pm
  
  sicli21
  Subscriber
  
  Hi I am simulating a Lithium Niobate device using CHARGE and FEEM. I modified the TFLN modulator scripts a bit but every time I get different effective index change as a function of applied voltage.
  
  My geometry looks like
  xspan of monitor is 20 um and y span is 3 um.
- March 1, 2024 at 5:30 pm
  
  Amrita Pati
  Ansys Employee
  
  Hi,
  That is a bit strange. The simulation should give the same results if nothing has been changed. I will test it for the modified file.
  Also, I would like to add that we had performed a lot of convergence testing on the span of the simulation in both CHARGE and FEEM to confirm the validity of the results. I believe, you will also need to perform convergence testing for this modified geometry. Also, the script for mode tracking could be finding different modes which could be behind the different results. The script for mode tracking is not very sophisticated so, we had to check individually for each voltage to make sure that the right mode was selected. Anyway, I will have a look and get back to you.
  - March 4, 2024 at 10:18 am
    
    sicli21
    Subscriber
    
    Hi, how can I perform convergence testing for the simulation region ? In FEEM, I can look at neff and loss. What parameters should I look for in CHARGE ? For the FEEM, I increased number of modes and reduced TE pol fraction a little.
    Also, I would like to know if I am interested in only electrostatics what material properties are the most important ones ? I want to use different semiconductor materials that is not in the material list.