CN-121995689-A - SEM image contour extraction method, system and terminal

CN121995689ACN 121995689 ACN121995689 ACN 121995689ACN-121995689-A

Abstract

The application provides an SEM image contour extraction method, a system and a terminal, which are used for respectively extracting an outer contour line and an inner contour line of a to-be-measured graph in a wafer to be measured according to an SEM image, selecting the outer contour line or the inner contour line as an initial contour line of the to-be-measured graph according to different properties of the to-be-measured graph, correcting the initial contour line according to a key dimension machine measurement value output by an electronic scanning microscope machine to obtain a target contour line of the to-be-measured graph, thereby improving the precision and the reliability of the target contour line, retaining the contour morphology feature of the to-be-measured graph, being beneficial to verifying and calibrating an OPC model, and solving the technical problems of low measurement precision, poor reliability and deviation from the key dimension of the contour output by the electronic scanning microscope machine when the two-dimensional contour of the wafer to-be-measured graph is extracted from the SEM image.

Inventors

Request for anonymity

Assignees

华芯程(上海)科技有限公司

Dates

Publication Date: 20260508
Application Date: 20260119

Claims (10)

1. A SEM image contour extraction method, comprising: acquiring a wafer to be measured, and determining a measurement position and a measurement direction of a pattern to be measured in the wafer to be measured according to the measurement requirement of the wafer to be measured; focusing an electron beam on a measurement position of the wafer to be measured by adopting an electron scanning microscope, and scanning the wafer to be measured along the measurement direction to obtain an SEM image of the pattern to be measured; Performing contour extraction operation on the SEM image to obtain an initial contour line of the graph to be measured; And acquiring a critical dimension machine measurement value of the graph to be measured output by an electronic scanning microscope machine, correcting the initial contour line according to the critical dimension machine measurement value, and obtaining a target contour line of the graph to be measured.
2. The SEM image contour extraction method according to claim 1, wherein the manner in which the contour extraction operation is performed on the SEM image includes: Arranging a plurality of sampling lines parallel to the measuring direction along the graph to be measured in the SEM image, and obtaining a plurality of gray scale waveform curves along the directions of the sampling lines; Positioning one or more contour interested areas of the graph to be measured based on the SEM image, and acquiring peak points and trough points of each gray scale waveform curve in each contour interested area so as to draw an outer contour line of the graph to be measured according to each peak point and draw an inner contour line of the graph to be measured according to each trough point; And selecting the outer contour line or the inner contour line as an initial contour line of the graph to be measured according to the property of the graph to be measured.
3. The SEM image contour extraction method according to claim 2, wherein the acquiring of each gray scale waveform curve further includes: Flattening the gray scale waveform curves.
4. The SEM image contour extraction method according to claim 2, wherein the positioning of the regions of interest of each contour of the graph to be measured comprises: And acquiring one or more contour datum lines of the graph to be measured, and extending preset measurement sizes to two sides along the measurement direction based on the contour datum lines to obtain one or more contour interested areas of the graph to be measured.
5. The SEM image contour extraction method according to claim 2, wherein the selecting of the outer contour line or the inner contour line as the initial contour line of the pattern to be measured according to the property of the pattern to be measured includes: If the graph to be measured is solid, selecting the outer contour line as an initial contour line of the graph to be measured; And if the graph to be measured is hollow, selecting the inner contour line as an initial contour line of the graph to be measured.
6. The SEM image contour extraction method according to claim 1, wherein correcting the initial contour line according to the cd machine measurement value comprises: Acquiring a critical dimension image measurement value of the graph to be measured according to the initial contour line, and calculating a difference value between the critical dimension image measurement value and the critical dimension machine measurement value and a critical dimension correction coefficient; And acquiring contour edge point measurement values of a plurality of sampling positions of the graph to be measured according to the initial contour line, and calculating contour edge point correction values of all the sampling positions according to the critical dimension correction coefficient so as to draw a target contour line of the graph to be measured.
7. The SEM image contour extraction method according to claim 6, wherein the critical dimension correction factor has a calculation formula: ; Wherein, the Correcting the coefficient for the critical dimension of the graph to be measured; a difference value between the critical dimension image measurement value and the critical dimension machine measurement value; Measuring a value of the critical dimension image of the graph to be measured; and measuring the value of the key size machine for the graph to be measured.
8. The SEM image contour extraction method according to claim 6, wherein the calculation formula of the contour edge point correction value is: ; Wherein, the Correcting the coefficient for the critical dimension of the graph to be measured; The first pattern to be measured Contour edge point measurement values of the sampling positions; The first pattern to be measured Contour edge point correction values for the individual sampling locations.
9. An SEM image contour extraction system, comprising: The measuring parameter determining module is used for obtaining a wafer to be measured and determining the measuring position and the measuring direction of a pattern to be measured in the wafer to be measured according to the measuring requirement of the wafer to be measured; The SEM image acquisition module is connected with the measurement parameter determination module and is used for focusing an electron beam at a measurement position of the wafer to be measured by adopting an electron scanning microscope and scanning the wafer to be measured along the measurement direction to obtain an SEM image of the pattern to be measured; The profile extraction module is connected with the SEM image acquisition module and is used for executing profile extraction operation on the SEM image to obtain an initial profile line of the graph to be measured; and the contour correction module is connected with the contour extraction module and is used for acquiring the key dimension machine measurement value of the graph to be measured output by the electronic scanning microscope machine, correcting the initial contour line according to the key dimension machine measurement value and obtaining the target contour line of the graph to be measured.
10. The SEM image contour extraction terminal is characterized by comprising a memory and a processor; The memory is used for storing a computer program; The processor is configured to execute the computer program stored in the memory, so that the terminal executes the SEM image contour extraction method according to any one of claims 1 to 8.

Description

SEM image contour extraction method, system and terminal Technical Field The present application relates to the field of semiconductor technologies, and in particular, to a method, a system, and a terminal for extracting an SEM image contour. Background In the chip manufacturing process, when a pattern on a mask is projected onto a wafer through a photoetching machine, an actual imaging pattern can deviate from a design pattern due to the diffraction effect of light, and the phenomena of line end shortening, line width narrowing, right angle rounding and the like are shown. OPC (Optical Proximity Correction ) techniques, by pre-adjusting the mask pattern, compensates for these optical errors in a reverse direction, ensuring that the actual imaged pattern after exposure is as close as possible to the design pattern. In order to verify and calibrate an OPC model, a scanning electron microscope (Scanning Electron Microscope, SEM) is generally used to obtain an SEM image of a wafer pattern to be measured, and an actual profile of the wafer pattern to be measured is extracted from the SEM image, and line width (CD) differences between the actual profile and a design pattern are compared, and line end shortening (Line End Shortening) and Edge Placement Error (EPE) are calculated to optimize parameters of the OPC model and improve accuracy of an OPC model prediction result. Therefore, the two-dimensional profile of the wafer to be measured is extracted from the SEM image as the basis for performing key parameter calculation and analysis. The accuracy and reliability of contour extraction have important influence on the improvement strategy and prediction accuracy of an OPC model, so that the accuracy of the actual imaging graph of the photoetching process can be influenced, and the quality and performance of a final semiconductor product are influenced. In the prior art, extracting a two-dimensional outline of a pattern to be measured of a wafer from an SEM image mainly depends on a digital image processing algorithm, specifically, an edge detection algorithm, such as an edge detection operator based on threshold segmentation, is adopted to identify edge pixel points in the SEM image, and post-processing such as connection, fitting, smoothing and the like is performed on a plurality of detected edge pixel points to generate a continuous vector outline. However, the existing contour extraction technology has significant disadvantages: ① SEM images themselves suffer from severe noise, edge blurring, and contrast non-uniformity problems, resulting in broken, rough, and filled false edges of the extracted contours based on conventional edge detection algorithms. ② The edge detection operator based on threshold segmentation is extremely sensitive to parameter selection, and morphological post-processing can introduce geometric distortion, affecting the measurement accuracy of the wafer pattern to be measured. ③ The profile measurement value obtained based on the SEM image has deviation from the profile critical dimension output by the electronic scanning microscope machine, and cannot be matched accurately. Currently, with the wide application of deep learning models, there are also attempts to extract two-dimensional contours of a pattern to be measured of a wafer from SEM images based on the deep learning model. However, this method relies heavily on a large amount of labeling data and has limited generalization ability. More importantly, the output of the deep learning model driven by pure data may violate physical knowledge, such as the phenomenon of contour self-intersection, or the non-smooth saw tooth shape of the contour, etc., so that the obtained contour measurement value lacks in interpretability, and the reliability under a severe measurement scene is difficult to ensure. Disclosure of Invention In view of the above-mentioned drawbacks of the prior art, the present application is directed to a method, a system and a terminal for extracting an SEM image profile, which are used for solving the technical problems of low measurement accuracy, poor reliability and deviation from the critical dimension of the profile output by an electronic scanning microscope (SEM) platform in the existing two-dimensional profile for extracting the wafer to be measured from the SEM image. In order to achieve the above and other related objects, a first aspect of the present application provides an SEM image contour extraction method, which includes obtaining a wafer to be measured, determining a measurement position and a measurement direction of a pattern to be measured in the wafer to be measured according to a measurement requirement of the wafer to be measured, focusing an electron beam on the measurement position of the wafer to be measured by using an electron scanning microscope, scanning the wafer to be measured along the measurement direction to obtain an SEM image of the pattern to be measured, performing a contour ext