Logo PTB

A new method to derive best-fit parameters and their uncertainties from depolarizing Mueller-matrices.

19.09.2019

We show a novel method to derive best-fitting sample parameters and their uncertainties from measured Mueller matrices.
The nonoptimal treatment of depolarization in current approaches to evaluate Mueller matrix measurements may lead to systematic measurement errors and inadequate uncertainty estimations, particularly for surfaces with substantial depolarization. To overcome these issues, we develop improved analysis methods providing realistic uncertainty estimations and more reliable measurement results.
Commonly used merit functions treat depolarization of Mueller-ellipsometry measurements as a residual error in the optimization process. The best fit solution has the smallest overall depolarization. We propose to use depolarization to derive a weight-factor for each residual contribution to the merit-function.
Exploiting raw data provided by our Sentech SE 850 ellipsometer we include also measurement statistics in the merit function. With it, embedded into a Bayesian approach the best-fit values and their uncertainties are determined much more reliably and realistically.
We demonstrate the advantages of this new method analyzing both virtual and real measurement data.
          
Grunewald, Tobias; 4.2, Bild- und Wellenoptik, PTB-Braunschweig
Wurm, Matthias; 4.2, Bild- und Wellenoptik, PTB-Braunschweig
Teichert, Sven; 4.2, Bild- und Wellenoptik, PTB-Braunschweig
Bodermann, Bernd; 4.2, Bild- und Wellenoptik, PTB-Braunschweig
Reck, Johanna; Sentech Instruments GmbH, Berlin, GERMANY
Richter, Uwe; Sentech Instruments GmbH, Berlin, GERMANY