To-be-measured sample piece parameter measuring method based on ellipsometer

Abstract

The invention provides a to-be-measured sample piece parameter measuring method based on an ellipsometer. The method comprises the following steps: building a t distribution model based on a characteristic that random noise in an ellipsometer measurement system meets Gaussian distribution; collecting a finite-order optical periodic light beam reflected by the to-be-measured sample piece to obtaina light intensity harmonic signal of the sample piece; eliminating random noise contained in the light intensity harmonic signal through the established t distribution model, estimating a light intensity true value reflected by the sample piece, and further calculating to obtain a corresponding Fourier coefficient true value; and finally, according to calculation formulas such as a specific physical model, solving parameters such as the thickness and the optical constant of the sample piece. Based on the fact that light intensity random noise meets Gaussian distribution, the collected light intensity is subjected to data processing through the t distribution theory in statistics, the true value of the light intensity is estimated, then the parameters of the to-be-detected sample piece areestimated, and the repeatability precision of the ellipsometer detection system is improved while the detection efficiency is not reduced.

Description

technical field [0001] The technical field of instrument measurement of the present invention, more specifically, relates to a method for measuring parameters of a sample to be tested based on an ellipsometer. Background technique [0002] An ellipsometer is an optical measurement instrument used to probe film thickness, optical constants, and the microstructure of materials. Due to the high measurement accuracy, suitable for ultra-thin films, non-contact with the sample, no damage to the sample and no need for a vacuum environment, the ellipsometer is an extremely attractive measuring instrument. In recent years, with the development and application of computers, the fitting analysis of ellipsometric data has become easier, prompting the application of ellipsometer in more fields. The automation of hardware and the maturity of software have greatly improved the speed of calculation, and mature software has provided new methods to solve problems. Therefore, ellipsometers a...

AI Technical Summary

Problems solved by technology

However, in the case of small samples, the randomness of the data is strong, and its variance is different from the overall sample variance

If the simple arithmetic mean is used, there may be a large difference between the small sample mean and the overall sample mean.

Images