Abstract
The use of the biased estimator in the fitting of spectroscopic ellipsometry data is examined and applied to data from two-channel polarization modulation ellipsometry experiments. It is pointed out that the use of the biased estimator, as opposed to the unbiased estimator that is usually found in the literature, allows the experimentalist to weight properly the more accurate parts of the spectrum, to switch among different representations of the data, and to calculate a goodness of fit. The fit to data taken on a 59-nm SiO2 film on Si is examined with both the biased and the unbiased estimators.
Keywords