US-12626354-B2 - Learnable defect detection for semiconductor applications

US12626354B2US 12626354 B2US12626354 B2US 12626354B2US-12626354-B2

Abstract

Methods and systems for learnable defect detection for semiconductor applications are provided. One system includes a deep metric learning defect detection model configured for projecting a test image for a specimen and a corresponding reference image into latent space, determining a distance in the latent space between one or more different portions of the test image and corresponding portion(s) of the corresponding reference image, and detecting defects in the one or more different portions of the test image based on the determined distances. Another system includes a learnable low-rank reference image generator configured for removing noise from one or more test images for a specimen thereby generating one or more reference images corresponding to the one or more test images.

Inventors

Jing Zhang
Zhuoning Yuan
Yujie Dong
Kris Bhaskar

Assignees

KLA CORPORATION

Dates

Publication Date: 20260512
Application Date: 20221211

Claims (20)

1 . A system configured to generate a reference image for a specimen, comprising: one or more computer systems; and one or more components executed by the one or more computer systems, wherein the one or more components comprise a learnable low-rank reference image generator, wherein the one or more computer systems are configured for inputting one or more test images for a specimen into the learnable low-rank reference image generator, wherein the one or more test images are generated for different locations on the specimen corresponding to the same location in a design for the specimen, and wherein the learnable low-rank reference image generator is configured for removing noise from the one or more test images thereby generating one or more reference images corresponding to the one or more test images; and wherein a defect detection component detects defects on the specimen based on differences between the one or more test images and their corresponding one or more reference images.
2 . The system of claim 1 , wherein the defect detection component comprises a deep learning defect detection component.
3 . The system of claim 1 , wherein the defect detection component comprises a deep metric learning defect detection model.
4 . The system of claim 1 , wherein the specimen is a wafer on which a layer of patterned features has been formed using multiple lithography exposure steps.
5 . The system of claim 1 , wherein the specimen is a wafer on which a layer of patterned features has been formed using extreme ultraviolet lithography.
6 . The system of claim 1 , wherein the learnable low-rank reference image generator comprises a learnable principle component analysis model.
7 . The system of claim 1 , wherein the learnable low-rank reference image generator comprises a learnable independent component analysis model or a learnable canonical correlation analysis model.
8 . The system of claim 1 , wherein the learnable low-rank reference image generator comprises a linear or non-linear regression model, a spatial low-rank neural network model, or a spatial low-rank probabilistic model.
9 . The system of claim 1 , wherein the defect detection component is included in the one or more components executed by the one or more computer systems, and wherein the one or more computer systems are further configured for jointly training the learnable low-rank reference image generator and the defect detection component with one or more training images and pixel-level ground truth information for the one or more training images.
10 . The system of claim 9 , wherein the one or more training images comprise images for one or more defect classes selected by a user, images for one or more hot spots on the specimen selected by the user, or images for one or more weak patterns in a design for the specimen selected by the user.
11 . The system of claim 9 , wherein the pixel-level ground truth information comprises information for the one or more training images generated from results of physics simulation performed with the one or more training images.
12 . The system of claim 9 , wherein the pixel-level ground truth information comprises information converted into a first format from known defect locations in a second format different than the first format.
13 . The system of claim 1 , wherein the learnable low-rank reference image generator and the defect detection component are further configured for inline defect detection.
14 . The system of claim 1 , wherein the one or more components further comprise a defect classification component configured for separating the detected defects into two or more types, and wherein the defect classification component is a deep learning defect classification component.
15 . The system of claim 1 , wherein the different locations comprise locations in different dies on the specimen.
16 . The system of claim 1 , wherein the different locations comprise multiple locations in only one die on the specimen.
17 . The system of claim 1 , wherein the one or more test images correspond to a job frame generated for the specimen by an imaging system, and wherein the one or more computer systems are further configured for repeating the inputting for one or more other test images corresponding to a different job frame generated for the specimen by the imaging system such that the learnable low-rank reference image generator separately generates the one or more reference images for the job frame and the different job frame.
18 . The system of claim 1 , wherein the one or more test images are generated for the specimen by an imaging system using only a single mode of the imaging system, and wherein the one or more computer systems are further configured for repeating the inputting for one or more other test images generated for the specimen by the imaging system using a different mode of the imaging system such that the learnable low-rank reference image generator generates the one or more reference images for the one or more other test images.
19 . A computer-implemented method for generating a reference image for a specimen, comprising: inputting one or more test images for a specimen into a learnable low-rank reference image generator, wherein the learnable low-rank reference image generator is included in one or more components executed by one or more computer systems, wherein the one or more test images are generated for different locations on the specimen corresponding to the same location in a design for the specimen, and wherein the learnable low-rank reference image generator is configured for removing noise from the one or more test images thereby generating one or more reference images corresponding to the one or more test images; and detecting defects on the specimen based on differences between the one or more test images and their corresponding one or more reference images.
20 . A non-transitory computer-readable medium storing program instructions executable on one or more computer systems for performing a computer-implemented method for generating a reference image for a specimen, wherein the computer-implemented method comprises: inputting one or more test images for a specimen into a learnable low-rank reference image generator, wherein the learnable low-rank reference image generator is included in one or more components executed by the one or more computer systems, wherein the one or more test images are generated for different locations on the specimen corresponding to the same location in a design for the specimen, and wherein the learnable low-rank reference image generator is configured for removing noise from the one or more test images thereby generating one or more reference images corresponding to the one or more test images; and detecting defects on the specimen based on differences between the one or more test images and their corresponding one or more reference images.

Description

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention generally relates to methods and systems for learnable defect detection for semiconductor applications. Certain embodiments relate to systems and methods for detecting defects on a specimen using a deep metric learning defect detection model and/or a learnable low-rank reference image generator. 2. Description of the Related Art The following description and examples are not admitted to be prior art by virtue of their inclusion in this section. Fabricating semiconductor devices such as logic and memory devices typically includes processing a substrate such as a semiconductor wafer using a large number of semiconductor fabrication processes to form various features and multiple levels of the semiconductor devices. For example, lithography is a semiconductor fabrication process that involves transferring a pattern from a reticle to a resist arranged on a semiconductor wafer. Additional examples of semiconductor fabrication processes include, but are not limited to, chemical-mechanical polishing (CMP), etch, deposition, and ion implantation. Multiple semiconductor devices may be fabricated in an arrangement on a single semiconductor wafer and then separated into individual semiconductor devices. Inspection processes are used at various steps during a semiconductor manufacturing process to detect defects on wafers to drive higher yield in the manufacturing process and thus higher profits. Inspection has always been an important part of fabricating semiconductor devices. However, as the dimensions of semiconductor devices decrease, inspection becomes even more important to the successful manufacture of acceptable semiconductor devices because smaller defects can cause the devices to fail. Most inspection methods include two major steps: generating a reference image followed by performing defect detection. There are many different ways to perform each step. For example, a reference image may be generated by determining a median or average of multiple images. In another example, the reference image may be a constructed reference and alternatives based on low-rank approximation. Defect detection may also be performed in a number of different ways. For example, defect detection may be unsupervised using a subtraction-based detection algorithm (e.g., MDAT, LCAT, etc.). Alternatively, defect detection may be supervised using a pixel-level detection algorithm (e.g., single image detection performed using a deep learning (DL) model and electron beam images). There are, however, a number of disadvantages to the various defect detection methods that are currently used. For example, generating a reference image by calculating a median or average is generally insufficient for dealing with die-to-die process variation although constructed reference and alternatives based on low-rank approximation partially addresses this issue. However, constructed reference and alternatives are potentially ineffective for sub-regions on a wafer and possibly destroy defect signals. Currently used unsupervised defect detection methods are disadvantageous because detection depends on the quality of the reference images and test images. Such unsupervised defect detection methods also do not provide selectivity on sensitivity enhancement for targeted defect types or relatively small defects. Supervised defect detection methods require a lot of labeled defect candidates for training, which in practice can be time consuming for recipe setup. Accordingly, it would be advantageous to develop systems and methods for learnable defect detection for semiconductor applications that do not have one or more of the disadvantages described above. SUMMARY OF THE INVENTION The following description of various embodiments is not to be construed in any way as limiting the subject matter of the appended claims. One embodiment relates to a system configured to detect defects on a specimen. The system includes one or more computer systems and one or more components executed by the one or more computer systems. The one or more components include a deep metric learning (DML) defect detection model configured for projecting a test image generated for a specimen and a corresponding reference image into latent space. For one or more different portions of the test image, the DML defect detection model is also configured for determining a distance in the latent space between the one or more different portions and corresponding one or more portions of the corresponding reference image. In addition, the DML defect detection model is configured for detecting defects in the one or more different portions of the test image based on the distances determined for the one or more different portions of the test image. The system may be further configured as described herein. Another embodiment relates to a computer-implemented method for detecting defects on a specimen. The method includes projecting a test