Statistical character analysis for through-focus scanning optical microscopy in double floating variables measurement applications

Abstract

Through-focus scanning optical microscopy (TSOM) is a model-based optical metrology for nano-scaled geometries with abilities of high efficiency, low cost, and non-destruction. The state-of-the-art studies on TSOM have been devoted to improving accuracies of single variable measurements or multiple variable measurements in limited degree of freedom, while the statistical significance of such applications has been generally ignored. Therefore, in this paper, a method requiring only one simulated model with multiple floating variables is proposed to analyze the statistical significance of TSOM. Given the advantage of ease of adjusting the simulated inputs, this method possesses abilities to qualitatively and quantitatively analyze variable correlations, repeatability accuracies, sensitivity of each variable as affected by the others, and the propagation of uncertainties in experimental parameters by first generating the simulated model for the TSOM system using finite difference time domain and Fourier optics; then, introducing subtle changes to the output - TSOM image. The proposed method is feasible for generalization to various nano-scaled structures and analyses of various experimental parameters, which helps improve the traceability of TSOM and its industrial applications.