Defect examination on a semiconductor specimen
There is provided a system and method for defect examination on a semiconductor specimen. The method comprises obtaining a runtime image of the semiconductor specimen, generating a reference image based on the runtime image using a machine learning (ML) model, and performing defect examination on the runtime image using the generated reference image. The ML model is previously trained during setup using a training set comprising one or more pairs of training images, each pair including a defective image and a corresponding defect-free image. The training comprises, for each pair, processing the defective image by the ML model to obtain a predicted image, and optimizing the ML model to minimize a difference between the predicted image and the defect-free image.
1. A computerized system for defect examination on a semiconductor specimen, the system comprising a processing and memory circuitry (PMC) configured to:
obtain a runtime image of the semiconductor specimen;
generate a reference image based on the runtime image using a machine learning (ML) model, wherein the ML model is previously trained during setup using a training set comprising one or more pairs of training images, each pair including a defective image and a corresponding defect-free image, the training comprising, for each pair:
processing the defective image by the ML model to obtain a predicted image; and
optimizing the ML model to minimize a difference between the predicted image and the defect-free image,
wherein the defective image and the defect-free image in each pair are pre-processed before the training by performing image augmentation on the defective image and the defect-free image using inpainting, giving rise to multiple defective sub-images corresponding to respective portions of the defect-free image, and wherein each defective sub-image has a removed subset of sub-regions, and the ML model works as an inpainting network to reconstruct the removed subset of sub-regions in the defective sub-image; and
perform defect examination on the runtime image using the generated reference image.
2. The computerized system according to claim 1 , wherein the ML model is a deep neural network (DNN) configured to learn non-linear mapping between two populations of defective images and defect-free images.
3. The computerized system according to claim 1 , wherein the defective image in each pair is one of: an actual image of the semiconductor specimen acquired by an examination tool and verified as containing a defective feature, or a simulated image by synthetically implanting a defective feature on a defect-free image, and wherein the corresponding defect-free image in each pair is one of: an actual image of the semiconductor specimen acquired by an examination tool and verified as not containing a defective feature, or a simulated image generated based on design data of the semiconductor specimen.
4. The computerized system according to claim 1 , wherein the defective image and the defect-free image in each pair are pre-processed before the training by registering the defective image and the defect-free image to correct misalignment therebetween.
5. The computerized system according to claim 1 , wherein the defective image and the defect-free image in each pair are pre-processed before the training by performing noise filtration on the defective image and the defect-free image to reduce noises representative of variations caused by one or more physical processes of the specimen.
6. The computerized system according to claim 1 , wherein the defective image and the defect-free image in each pair are pre-processed before the training by performing image registration, noise filtration and image augmentation on the defective image and the defect-free image.
7. The computerized system according to claim 1 , wherein the ML model comprises at least one learning module preceding a DNN, the at least one learning module trained to perform at least one of image registration and noise filtration on the defective image and the defect-free image in each pair before being processed by the DNN.
8. The computerized system according to claim 1 , wherein the ML model is optimized using a loss function selected from: Mean squared error (MSE), Sum of absolute difference (SAD), structural similarity index measure (SSIM), and an edge-preserving loss function.
9. The computerized system according to claim 1 , wherein the defect examination is one of: defect detection, defect review, and defect classification.
10. The computerized system according to claim 1 , wherein the generated reference image has improved alignment with the runtime image and shares similar noise levels as of the runtime image, and the PMC is configured to perform defect examination by directly comparing the runtime image with the generated reference image to obtain a defect examination result indicative of defect distribution on the semiconductor specimen.
11. A computerized method of defect examination on a semiconductor specimen, the method comprising:
obtaining a runtime image of the semiconductor specimen;
generating a reference image based on the runtime image using a machine learning (ML) model, wherein the ML model is previously trained during setup using a training set comprising one or more pairs of training images, each pair including a defective image and a corresponding defect-free image, the training comprising, for each pair:
processing the defective image by the ML model to obtain a predicted image; and
optimizing the ML model to minimize a difference between the predicted image and the defect-free image,
wherein the defective image and the defect-free image in each pair are pre-processed before the training by performing image augmentation on the defective image and the defect-free image using inpainting, giving rise to multiple defective sub-images corresponding to respective portions of the defect-free image, and wherein each defective sub-image has a removed subset of sub-regions, and the ML model works as an inpainting network to reconstruct the removed subset of sub-regions in the defective sub-image; and
performing defect examination on the runtime image using the generated reference image.
12. The computerized method according to claim 11 , wherein the defective image in each pair is one of: an actual image of the semiconductor specimen acquired by an examination tool and verified as containing a defective feature, or a simulated image by synthetically implanting a defective feature on a defect-free image, and wherein the corresponding defect-free image in each pair is one of: an actual image of the semiconductor specimen acquired by an examination tool and verified as not containing a defective feature, or a simulated image generated based on design data of the semiconductor specimen.
13. The computerized method according to claim 11 , wherein the defective image and the defect-free image in each pair are pre-processed before the training by registering the defective image and the defect-free image to correct misalignment therebetween.
14. The computerized method according to claim 11 , wherein the defective image and the defect-free image in each pair are pre-processed before the training by performing noise filtration on the defective image and the defect-free image to reduce noises representative of variations caused by one or more physical processes of the specimen.
15. The computerized method according to claim 11 , wherein the ML model comprises at least one learning module preceding a DNN, the at least one learning module trained to perform at least one of image registration and noise filtration on the defective image and the defect-free image in each pair before being processed by the DNN.
16. The computerized method according to claim 11 , wherein the defect examination is one of: defect detection, defect review, and defect classification.
17. The computerized method according to claim 11 , wherein the generated reference image has improved alignment with the runtime image and shares similar noise levels as of the runtime image, and the performing defect examination comprises directly comparing the runtime image with the generated reference image to obtain a defect examination result indicative of defect distribution on the semiconductor specimen.
18. A non-transitory computer readable storage medium tangibly embodying a program of instructions that, when executed by a computer, cause the computer to perform a method of defect examination on a semiconductor specimen, the method comprising:
obtaining a runtime image of the semiconductor specimen;
generating a reference image based on the runtime image using a machine learning (ML) model, wherein the ML model is previously trained during setup using a training set comprising one or more pairs of training images, each pair including a defective image and a corresponding defect-free image, the training comprising, for each pair:
processing the defective image by the ML model to obtain a predicted image; and
optimizing the ML model to minimize a difference between the predicted image and the defect-free image,
wherein the defective image and the defect-free image in each pair are pre-processed before the training by performing image augmentation on the defective image and the defect-free image using inpainting, giving rise to multiple defective sub-images corresponding to respective portions of the defect-free image, and wherein each defective sub-image has a removed subset of sub-regions, and the ML model works as an inpainting network to reconstruct the removed subset of sub-regions in the defective sub-image; and
performing defect examination on the runtime image using the generated reference image.