Learning Forum

[Saha]

I have been using the material from the database "SiO2 (Glass) - Palik" at wavelength 2um. I noticed the imaginary n has  value of -1.11022e-16. It has resulted in some discrepency like negative absorption in RTA analysis. I always assumed Lumerical considers index or permittivity as n= real(n)+i*imag(n) and hence negative imag(n) would be considered gain for the system. So my query is-

1. How to correctly model the material in such case to consider loss in stead of gain ( which I believe should be the case for any real dielectric material)?
2. If I want to exclude the imaginary part altogether while keep using the database value (real parts only), is there a way to set it up in the simulation?

[Guilin Sun]

In your case this material SiO2 can be practivally considered as pure dielectric with a given refractive index directly at the geometry tab, instead of using the mutil-coefficent fitting model.

Secondly, this small negative value is due to fitting error. If you use much smaller "fit tolerance" , or larger weight for the imaginery in the "show advanced" tab, it can be corrected.

Thirdly, if such small "gain" does not cause simulation diverging or wrong result, it is ok.

Speaking to  negative absorption in RTA analysis, which one is larger than 1? if R or T is larger than one, then the simulation should be modified: https://optics.ansys.com/hc/en-us/articles/12614264530323-Transmission-Results-Greater-Than-One

If both R and T are normal then it is the marial "gain" that causes the issue. Choice of #1 is the simpest solution.