Learning Forum

[100521665]

I am trying to implement my own (n,k) sampled data, and while doing so, Lumerical's fitting does this:

 

 

Which I've been using through my simulations because it wasn't improving doing whatever I tried to do to recover the actual curve, but now that I am seeing some unconsistencies I want to ensure that it isn't the fitting the one making them. Any tip on how would you fit better this curve? Maybe subtracting points, choosing a better range? Anything is appreciated. 

[Guilin Sun]

The fitting mechanism is to ensure the causality. Thus it has special requirement for the fitting. Lumerical multi-coefficient material mode works the best for most natural materials. However in some cases it can be challenge such as your case. 

You may reduce the number of coefficients (as 20 does not work now),  try the advanced fitting feature to use smaller weight factor to enhance the real part fitting. Some tips can be found here:

Tips for improving the quality of optical material fits

Using the Material Explorer to view and adjust optical material models