Minimization of spectral phase errors in spectrally resolved white light interferometry by the iterative least-squared phase-shifting method

Abstract

We introduce an iterative least-squared phase-shifting method, reported as inherently insensitive to any types of phase-shifting errors, to calculate the spectral phase in phase-shifting spectrally resolved white light interferometry (PS-SRI). The actual phase shifts corresponding to all wavelengths can be reversely determined from the phase-shifted spectral interference fringes through least-squares fitting and the spectral phase is calculated by using the actual phase shifts in an iterative numerical manner. Because this method reduces the phase-shifting errors for calculating the spectral phase effectively, it can contribute to improving the accuracy of measuring topographic surface profiles. Moreover, it leads to accomplishing thin film thickness measurements less than 100 nm, in which most white light scanning interferometry and spectrally resolved white light interferometry (SRI) have difficulty because of mismatching measured spectral phase with the theoretical model by the spectral phase error. In this paper, a short description of the iterative least-squared phase-shifting method is presented and verified with simulations for calculating the topographic surface and thin film thickness profiles in PS-SRI.