Method for determining parameters of to-be-measured sample from theoretical spectrum library and measurement equipment

Abstract

The invention provides a method for determining morphology parameters of a to-be-measured sample based on spectrum library matching and measuring equipment. The method comprises the following steps: constructing a theoretical spectrum library according to a structure model of the to-be-measured sample, acquiring a measurement spectrum, and determining an initial approximate optimal matching spectrum and an initial approximate optimal matching parameter vector of the measurement spectrum from the theoretical spectrum library; obtaining at least three sampling points of each to-be-measured parameter and a mean square error of the theoretical spectrum and the measurement spectrum corresponding to each sampling point; and performing parabola fitting on the distribution curve of the mean squareerrors of the theoretical spectrum and the measurement spectrum corresponding to the at least three sampling points of each to-be-measured parameter relative to the parameter values corresponding tothe sampling points to obtain a global optimal matching value of each to-be-measured parameter. According to the method, the matching precision between the measurement spectrum and the theoretical spectrum is remarkably improved, and the problem of inaccurate interpolation caused by the adoption of a traditional method when morphology parameters are coupled is solved.

Description

Technical field [0001] The invention relates to the technical field of semiconductor testing equipment, in particular to a method and measuring equipment for determining the parameters of a sample to be tested from a theoretical spectral library based on OCD measurement technology. technical background [0002] With the development of semiconductor manufacturing industry, the size of semiconductor devices is shrinking, and the design of device structure is becoming more and more complicated, which puts forward more stringent control requirements for semiconductor manufacturing process. Therefore, it is necessary to perform wafer measurement or inspection after many process steps of semiconductor manufacturing to obtain higher yield. OpticalCritical-Dimension (OCD) measurement technology is a commonly used critical dimension measurement technology in semiconductor manufacturing process. [0003] The basic working principle of OCD measurement technology is: by establishing a theore...

AI Technical Summary

Problems solved by technology

However, in many samples to be tested, the multiple shape parameters of the sample to be tested are not independent but have strong coupling; The change of the combination is sensitive, but it is difficult to determine the contribution of each parameter to the change of the spectrum. This situation corresponds to the distribution of the MSE as a non-orthogonal hypersurface near the approximate best matching point (such as figure 1 )

At this time, if the traditional method is used, that is, each parameter dimension is independently corrected, and only along a certain parameter coordinate axis near the initial approximate best matching point (ie figure 1 The coordinate axis where the shape parameter HT is located or the coordinate axis where the shape parameter MCD is located) will deviate from the bottom area of ​​the hypersurface, or the information contained in the MSE curve change in a single parameter dimension is not sufficient, which makes the follow-up The MSE curve fitting method is unreasonable or even invalid, which eventually leads to insufficient measurement accuracy of the shape parameters

Images