Semi-supervised anomaly detection in scanning electron microscope images

Patent metadata

Field	Value
Publication number	US-10789703-B2
Application number	US-201816106341-A
Country	US
Kind code	B2
Filing date	Aug 21, 2018
Priority date	Mar 19, 2018
Publication date	Sep 29, 2020
Grant date	Sep 29, 2020

Abstract

Autoencoder-based, semi-supervised approaches are used for anomaly detection. Defects on semiconductor wafers can be discovered using these approaches. The model can include a variational autoencoder, such as a one that includes ladder networks. Defect-free or clean images can be used to train the model that is later used to discover defects or other anomalies.

First claim

What is claimed is:

1. A system comprising:
a wafer inspection tool configured to generate images of a wafer, wherein the wafer inspection tool includes an electron beam source and a detector; and
a processor in electronic communication with the wafer inspection tool, wherein the processor operates a model configured to find one or more anomalies in the images, wherein the model is trained using semi-supervised machine learning based on only defect-free training images of semiconductor devices.

2. The system of claim 1 , wherein the wafer inspection tool is a scanning electron microscope.

3. The system of claim 1 , wherein the model includes a variational autoencoder.

4. The system of claim 3 , wherein the variational autoencoder includes ladder networks.

5. A method comprising:
receiving an image of a wafer at a processor, wherein the processor operates a model configured to find one or more anomalies in the image, wherein the model is trained using semi-supervised machine learning based on only defect-free training images of semiconductor devices; and
determining presence of one or more anomalies in the image using the model.

6. The method of claim 5 , wherein the image is a scanning electron microscope image.

7. The method of claim 5 , wherein the training uses nominal patterns.

8. The method of claim 5 , wherein the model includes a variational autoencoder.

9. The method of claim 8 , wherein the variational autoencoder includes ladder networks.

10. The method of claim 5 , further comprising obtaining the image using a wafer inspection tool.

11. The method of claim 10 , wherein the wafer inspection tool is a scanning electron microscope.

12. The method of claim 5 , wherein the one or more anomalies are each one of an anomaly patch or an anomaly region.

13. The method of claim 5 , further comprising:
determining, using the processor, a distance between the image and the defect-free training images in a feature space; and
determining, using the processor, if the image is an outlier based on the distance.

14. The method of claim 5 , further comprising determining, using the processor, if the image is an outlier using a generative adversarial network with an autoencoder as a generator and a convolutional neural network as a discriminator.

15. A non-transitory computer-readable storage medium, comprising one or more programs for executing a model on one or more computing devices, wherein the model is trained using semi-supervised machine learning based on only defect-free training images of semiconductor devices, and wherein the model is configured to:
receive an image of a wafer; and
determine presence of one or more anomalies in the image.

16. The defect detection model of claim 15 , wherein the image is a scanning electron microscope image.

17. The defect detection model of claim 15 , wherein the model is trained using nominal patterns.

18. The defect detection model of claim 15 , wherein the model includes a variational autoencoder.

19. The defect detection model of claim 18 , wherein the variational autoencoder includes ladder networks.

20. The defect detection model of claim 15 , wherein the model is configured to perform outlier detection thereby detecting anomalies.

Assignees

• Kla Tencor Corp

Inventors

• Lu Shaoyu
• He Li
• Venkataraman Sankar

Classifications

• G06T7/0004 Primary

  Industrial image inspection · CPC title

• G06V10/764

  using classification, e.g. of video objects · CPC title

• G06T7/001 Primary

  using an image reference approach · CPC title

• G06N3/047

  Probabilistic or stochastic networks · CPC title

• G06N3/045

  Combinations of networks · CPC title