Method of measuring pattern dimension and method of controlling semiconductor device process
View Patent ↗This invention provides a method of measuring semiconductor pattern dimensions capable of realizing a stable and highly precise pattern dimension measurement technique even when the pattern cross-sectional shapes are changed and making the calculation amount relatively small to reduce the calculation time. More specifically, the relationship between cross-sectional shapes of a pattern and measurement errors in a specified image processing technique is evaluated in advance by the electron beam simulation in a pattern measurement system in a length measuring SEM, and in the actual dimension measurement, dimensions of an evaluation objective pattern are measured from image signals of a scanning electron microscope, and errors of the dimensional measurement of the evaluation objective pattern are estimated and revised based on the relationship between cross-sectional shapes of a pattern and measurement errors evaluated in advance, thereby realizing highly precise measurement where dimensional errors depending on pattern solid shapes are eliminated.
1. A method of measuring pattern dimensions, comprising:
a first step in which a relationship between cross-sectional shapes of a pattern and measurement errors of a pattern is evaluated in a specified image processing technique; and
a second step in an actual measurement in which dimension measurement of an evaluation objective pattern from image signals of a scanning electron microscope is carried out, and errors of the dimension measurement of said evaluation objective pattern is revised based on said relationship between the cross-sectional shapes of a pattern and the measurement errors of a pattern evaluated in advance.
2. The method of measuring pattern dimensions according to claim 1 ,
wherein a database in which the relationship between measurement errors between a pattern position detected by the specific image processing technique and actual position of the pattern and the cross-sectional shapes of said pattern is evaluated and recorded in advance is established in said first step, and cross-sectional shapes of said evaluation objective pattern are evaluated and the position of said evaluation objective pattern is detected by said specified image processing technique, and measurement errors in the case of measuring the evaluation objective patterns having said cross-sectional shapes are estimated and said measurement errors are revised based on the relationship between said pattern cross-sectional shapes and said measurement errors recorded in advance into said database in said second step.
3. The method of measuring pattern dimensions according to claim 2 ,
wherein said cross-sectional shape includes one of a tilt angle of a sidewall, roundness of a corner of pattern top portion, and roundness of a corner of pattern bottom portion, or a combination thereof.
4. The method of measuring pattern dimensions according to claim 2 ,
wherein said database is established by an electron beam simulation.
5. The method of measuring pattern dimensions according to claim 2 ,
wherein said database is established by a cross section measurement, an AFM measurement, or a measurement by scatterometry.
6. The method of measuring pattern dimensions according to claim 2 ,
wherein said evaluation of cross-sectional shapes in said actual dimension measurement is carried out by the use of feature quantity of an image calculated from SEM images.
7. The method of measuring pattern dimensions according to claim 2 ,
wherein said evaluation of cross-sectional shapes in said actual dimension measurement is carried out by scatterometry.
8. The method of measuring pattern dimensions according to claim 2 ,
wherein cross-sectional structure information of an objective to be measured and electron beam images and/or waveforms thereof obtained from an SEM observation of said cross-sectional structure or a simulation that simulates the SEM observation are displayed side by side, and a position of the pattern detected by designated image processing conditions is displayed on said cross-sectional structure and said electron beam images or said waveforms, thereby adjusting image processing conditions.
9. The method of measuring pattern dimensions according to claim 2 ,
wherein said specified image processing technique uses dimension measurement image processing algorithms or image processing parameters in which variations due to noises or SEM device parameters are small.
10. The method of measuring pattern dimensions according to claim 2 ,
wherein the relationship between said pattern cross-sectional shapes and said dimension measurement errors is recorded in the form of functions into the database.
11. The method of measuring pattern dimensions according to claim 2 ,
wherein, with regard to the establishment of said database, the relationship between said pattern cross-sectional shapes and signal waveforms obtained from actual images or an electron beam simulation is recorded, and after determining image processing algorithms and parameters of dimension measurement, the relationship between said cross-sectional shapes and said dimension measurement errors are calculated based on a combination of said cross-sectional shapes and said signal waveforms, and recorded into said database.
12. The method of measuring pattern dimensions according to claim 2 ,
wherein cross-sectional shapes of said evaluation objective pattern are evaluated by the use of plural electron beam signals in which angles between an electron beam emitted from said scanning electron microscope and a surface of the measurement sample are different.
13. A method of controlling semiconductor device manufacturing process,
wherein defects in process are detected and the process is controlled based on pattern measurement results obtained by using said measuring method of pattern dimensions described in claim 1 .