CN-115516293-B - Mode selection and defect detection training

Abstract

A system may be configured for joint defect discovery and optical mode selection. Defects are detected during the defect discovery step. The discovered defects are accumulated into a mode selection dataset. Mode selection is performed using the mode selection dataset to determine a mode combination. The pattern combination may then be used to train the defect detection model. Additional defects may then be detected by the defect detection model. The additional defects may then be provided to the mode selection dataset to further perform mode selection and train the defect detection model. One or more runtime patterns may then be determined. The system may be configured for mode selection and defect detection at the image pixel level.

Inventors

• ZHANG JING
• DONG YUJIE
• BHATIA VIKRAM
• P. McBride
• K. Bahaskar
• DUFFY BRIAN

Assignees

• 科磊股份有限公司

Dates

Publication Date

20260508

Application Date

20210517

Priority Date

20201221

Claims (20)

1. 1. A system, comprising: A controller communicatively coupled to an inspection subsystem configured to image at least one sample when configured with any of a plurality of candidate optical modes, the controller including one or more processors configured to execute program instructions that cause the one or more processors to repeatedly perform optical mode selection and defect detection training in combination: Receiving defect data for at least one defect on at least a portion of the at least one sample; Receiving at least one image from the inspection subsystem and storing the at least one image in a dataset, wherein the at least one image is associated with the at least one defect detected on the at least the portion of the at least one sample by the inspection subsystem configured to have a candidate optical mode of the plurality of candidate optical modes; Selecting one or more optical modes from the plurality of candidate optical modes by executing a mode selection model, and Training a defect detection model using images associated with the one or more selected optical modes; Wherein the one or more processors are further configured to determine at least one run-time optical mode from the plurality of candidate optical modes.
2. 2. The system of claim 1, wherein the mode selection model comprises: a sparse vector comprising a plurality of indices, each of the plurality of indices comprising a mode selection weight between zero and a first value.
3. 3. The system of claim 2, wherein the one or more optical modes selected by the mode selection model are selected by applying a threshold to the sparse vector.
4. 4. The system of claim 2, wherein the one or more optical modes selected by the mode selection model are selected by providing the plurality of indices as weight vectors to the defect detection model.
5. 5. The system of claim 1, wherein the mode selection model comprises: A random channel discard vector comprising a plurality of indices, each of the plurality of indices comprising a zero or non-zero mode selection weight.
6. 6. The system of claim 5, wherein the plurality of indexes are randomly set to zero or a non-zero value during each iteration.
7. 7. The system of claim 6, wherein the one or more optical modes of the dataset used to train the defect detection model are determined by the plurality of indices having the non-zero values.
8. 8. The system of claim 1, wherein the pattern selection model includes a model agnostic meta-learning algorithm.
9. 9. The system of claim 1, wherein the mode selection model comprises at least one of a forward selection or a backward selection algorithm.
10. 10. The system of claim 1, wherein determining the at least one runtime pattern includes generating a ranking table including at least one of a signal-to-noise ratio, a receiver operating characteristic, a capture rate, a nuisance rate, or a computational cost.
11. 11. The system of claim 1, wherein the plurality of candidate optical modes are determined by dimension reduction, the dimension reduction including at least one of correlation analysis or principal component analysis.
12. 12. The system of claim 1, wherein each of the plurality of candidate optical modes includes a wavelength, a focal length, an aperture, and a bandwidth.
13. 13. The system of claim 1, wherein the defect detection model includes at least one of a depth generation model, a convolutional neural network, a generation countermeasure network, a condition generation countermeasure network, a variational self-encoder, a representation learning network, or a transformer model.
14. 14. The system of claim 1, wherein the inspection subsystem comprises a broadband plasma inspection tool.
15. 15. The system of claim 1, further comprising performing a defect inspection test.
16. 16. The system of claim 15, further comprising performing reasoning using the defect detection model to evaluate at least one of stability or sensitivity of the defect detection model.
17. 17. The system of claim 1, wherein, when the one or more optical modes do not include an associated image in the dataset, the inspection subsystem is configured to image the at least one sample when configured to have the one or more optical modes selected from the plurality of candidate modes by the mode selection model.
18. 18. A method for performing optical mode selection and defect detection training, comprising: receiving defect data for at least one defect on at least a portion of at least one sample; Receiving at least one image from an inspection subsystem and storing the at least one image in a dataset, wherein the at least one image is associated with the at least one defect detected on the at least the portion of the at least one sample by the inspection subsystem configured to have a candidate optical mode of a plurality of candidate optical modes; selecting one or more optical modes from the plurality of candidate optical modes by executing a mode selection model; training a defect detection model using the images associated with the one or more optical modes selected by the mode selection model, and A defect inspection test is performed.
19. 19. The method of claim 18, wherein the mode selection model includes one or more of a random channel drop vector, a sparse vector, a model agnostic meta-learning algorithm, a forward selection algorithm, or a backward selection algorithm.
20. 20. A system, comprising: an inspection subsystem configured to image at least one sample when configured with a plurality of candidate optical modes; a controller communicatively coupled to the inspection subsystem, the controller including one or more processors configured to execute program instructions that cause the one or more processors to repeatedly perform optical mode selection and defect detection training in combination: Receiving defect data for at least one defect on at least a portion of the at least one sample; Receiving at least one image from the inspection subsystem and storing the at least one image in a dataset, wherein the at least one image is associated with the at least one defect detected on the at least the portion of the at least one sample by the inspection subsystem configured to have a candidate optical mode of the plurality of candidate optical modes; Selecting one or more optical modes from the plurality of candidate optical modes by executing a mode selection model, and Training a defect detection model using images associated with the one or more selected optical modes; Wherein the one or more processors are further configured to determine at least one run-time optical mode from the plurality of candidate optical modes.

Description

Mode selection and defect detection training Cross-reference to related applications The present application claims the rights of U.S. provisional application No. 63/027,975, filed on 21, 5/21, 2020, in accordance with the 35 u.s.c. ≡119 (e), the entire contents of which are incorporated herein by reference. Technical Field The present invention relates generally to semiconductor inspection and, more particularly, to classifying defects detected by semiconductor inspection. Background The semiconductor manufacturing environment is often highly controlled to inhibit contamination of the wafer with foreign materials that may interfere with the manufacturing process or degrade the performance of the finished device. Inspection systems are commonly used to locate defects, such as, but not limited to, foreign particles, on a substrate for screening and avoidance measures. The sensitivity of defect inspection may vary based on factors such as, but not limited to, defect type, measurement parameters, or defect detection model. Thus, identification of suitable measurement parameters and defect detection models can present challenges. It would therefore be advantageous to provide a system and method that addresses the above-described drawbacks. Disclosure of Invention A system in accordance with one or more illustrative embodiments of the present disclosure is disclosed. In one illustrative embodiment, the system includes a controller. In another illustrative embodiment, the controller is communicatively coupled to a verification subsystem. In another illustrative embodiment, the inspection subsystem is configured to image at least one sample when configured with any of a plurality of candidate optical modes. In another illustrative embodiment, the controller includes one or more processors configured to execute program instructions that cause the one or more processors to jointly perform optical mode selection and defect detection training. In another illustrative embodiment, the processor receives defect data for at least one defect on at least a portion of the at least one sample. In another illustrative embodiment, the processor receives at least one image from the inspection subsystem and stores the at least one image in a dataset. In another illustrative embodiment, the processor selects one or more optical modes from the plurality of candidate modes by executing a mode selection model. In another illustrative embodiment, the processor trains a defect detection model using the images associated with the one or more optical modes selected by the mode selection model. In another illustrative embodiment, the processor is further configured to determine at least one run-time optical mode from the plurality of candidate optical modes. A method according to one or more illustrative embodiments of the present disclosure is disclosed. The method may include performing optical mode selection and defect detection training. In one illustrative embodiment, the method includes receiving defect data for at least one defect on at least a portion of at least one sample. In another illustrative embodiment, the method includes receiving at least one image from an inspection subsystem and storing the at least one image in a dataset. In another illustrative embodiment, the method includes selecting one or more optical modes from the plurality of candidate optical modes by executing a mode selection model. In another illustrative embodiment, the method includes training a defect detection model using the images associated with the one or more optical modes selected by the mode selection model. In another illustrative embodiment, the method includes performing a defect inspection test. A system in accordance with one or more illustrative embodiments of the present disclosure is disclosed. In one illustrative embodiment, the system includes an inspection subsystem configured to image at least one sample when configured with a plurality of candidate optical modes. In another illustrative embodiment, the system includes a controller communicatively coupled to the inspection subsystem. In another illustrative embodiment, the controller includes one or more processors configured to execute program instructions that cause the one or more processors to jointly perform optical mode selection and defect detection training. In another illustrative embodiment, the processor receives defect data for at least one defect on at least a portion of the at least one sample. In another illustrative embodiment, the processor receives at least one image from the inspection subsystem and stores the at least one image in a dataset. In another illustrative embodiment, the processor selects one or more optical modes from the plurality of candidate modes by executing a mode selection model. In another illustrative embodiment, the processor trains a defect detection model using the images associated with the one or more optical modes selected by the mode sel