Inspection with previous step subtraction

1. An inspection system comprising:

a controller including one or more processors configured to execute program instructions causing the one or more processors to implement an inspection recipe by:

receiving first-step images of a plurality of sample regions after a first process step;

receiving second-step images of the plurality of sample regions after a second process step, wherein the second process step modifies the sample in at least one of the plurality of sample regions;

identifying one of the plurality of sample regions as a test region and at least some of the remaining sample regions as comparison regions, wherein the second-step image of the test region is a test image and the second-step images of the comparison regions are second-step comparison images;

generating a multi-step difference image by a weighted subtraction of a combination of at least one of the second-step comparison images and at least two of the first-step images from the test image; and

identifying defects in the test region associated with the second process step based on the multi-step difference image.

2. The inspection system of claim 1 , wherein a number of the first-step images used to generate the multi-step difference image is one greater than a number of the second-step comparison images to provide that the multi-step difference image is generated based on a same number of the first-step and the second-step images.

3. The inspection system of claim 2 , wherein the number of first-step images used to generate the multi-step difference image is at least three.

4. The inspection system of claim 1 , further comprising:

classifying at least some of the defects as at least one of a nuisance or a defect of interest.

5. The inspection system of claim 1 , wherein weights associated with the weighted subtraction are determined by a fitting technique.

6. The inspection system of claim 1 , wherein weights associated with the weighted subtraction are determined by a regression technique.

7. The inspection system of claim 1 , wherein weights associated with the weighted subtraction are determined by a multi-color adaptive threshold (MCAT) technique.

8. The inspection system of claim 1 , wherein weights associated with the weighted subtraction are determined by a machine learning technique.

9. The inspection system of claim 1 , wherein generating the multi-step difference image by the weighted subtraction of the combination of at least one of the second-step comparison images and at least two of the first-step images from the test image comprises:

generating the multi-step difference image (I MSDIFF,i ) based on the equation I MSDIFF,i =I SecondStep,i −Σ j≠i α j I SecondStep,j −Σ j β j I FirstStep,j , wherein I SecondStep,i corresponds to the test image, Σ j≠i α j I SecondStep,j corresponds to the at least one of the second-step comparison images, and Σ j β j I FirstStep,j corresponds to the at least two of the first-step images, wherein subscripts i and j correspond to the sample regions, wherein α and β correspond to weights.

10. The inspection system of claim 1 , wherein generating the multi-step difference image by the weighted subtraction of the combination of at least one of the second-step comparison images and at least two of the first-step images from the test image comprises:

generating the multi-step difference image (I MSDIFF,i ) based on the equation I MSDIFF,i =I SecondStep,i −Σ j≠i α j I SecondStep,j −[I FirstStep,i −Σ j≠i β j I FirstStep,j ], where I SecondStep,i corresponds to the test image, Σ j≠i α j I SecondStep,j corresponds to the at least one of the second-step comparison images, and [I FirstStep,i −Σ j≠i β j I FirstStep,j ] corresponds to the at least one of the first-step images, wherein subscripts i and j correspond to the sample regions, wherein α, β, and γ correspond to weights.

11. The inspection system of claim 1 , wherein generating the multi-step difference image by the weighted subtraction of the combination of at least one of the second-step comparison images and at least two of the first-step images from the test image comprises:

generating the multi-step difference image (I MSDIFF,i ) based on the equation I MSDIFF,i =I SecondStep,i −Σ j≠i α j I SecondStep,j −γ[I FirstStep,i −Σ j≠i β j I FirstStep,j ], where I SecondStep,i corresponds to the test image, Σ j≠i α j I SecondStep,j corresponds to the at least one of the second-step comparison images, and [I FirstStep,i −Σ j≠i β j I FirstStep,j ] corresponds to the at least one of the first-step images, wherein subscripts i and j correspond to the sample regions, wherein α, β, and γ correspond to weights.

12. The inspection system of claim 1 , wherein detecting defects on the sample based on the multi-step difference image comprises:

providing the inspection reference image to a machine learning algorithm; and

detecting defects on the sample based on an output of the machine learning algorithm.

13. The inspection system of claim 1 , wherein detecting defects on the sample based on the multi-step difference image comprises:

detecting defects on the sample based on the multi-step difference image using a multi-die adaptive threshold (MDAT) technique.

14. The inspection system of claim 1 , wherein the defects comprise:

a deviation of at least one of a shape, size, or orientation of a feature fabricated by the second process step.

15. The inspection system of claim 1 , wherein the defects comprise:

an absence of a feature intended to be fabricated by the second process step.

16. The inspection system of claim 1 , wherein the defects comprise:

at least one of a scratch, a pit, or residual material in the inspection region after the second process step.

17. The inspection system of claim 1 , wherein the defects comprise:

at least one of an unintended bridge between two features fabricated by the second process step or an unintended break in a feature fabricated by the second process step.

18. An inspection method comprising:

generating first-step images of a plurality of sample regions after a first process step;

generating second-step images of the plurality of sample regions after a second process step, wherein the second process step modifies the sample in at least one of the plurality of sample regions;

identifying one of the plurality of sample regions as a test region and at least some of the remaining sample regions as comparison regions, wherein the second-step image of the test region is a test image and the second-step images of the comparison regions are comparison images;

generating a multi-step difference image by a weighted subtraction of a combination of at least one of the second-step comparison images and at least two of the first-step images from the test image; and

identifying defects in the test region associated with the second process step based on the multi-step difference image.

19. The inspection method of claim 18 , wherein a number of the first-step images used to generate the multi-step difference image is one greater than a number of the second-step comparison images to provide that the multi-step difference image is generated based on a same number of the first-step and the second-step images.

20. The inspection method of claim 19 , wherein the number of first-step images used to generate the multi-step difference image is at least three.

21. The inspection method of claim 18 , further comprising:

classifying at least some of the defects as at least one of a nuisance or a defect of interest.

22. The inspection method of claim 18 , wherein weights associated with the weighted subtraction are determined by a fitting technique.

23. The inspection method of claim 18 , wherein weights associated with the weighted subtraction are determined by a regression technique.

24. The inspection method of claim 18 , wherein weights associated with the weighted subtraction are determined by a multi-color adaptive threshold (MCAT) technique.

25. The inspection method of claim 18 , wherein weights associated with the weighted subtraction are determined by a machine learning technique.

26. The inspection method of claim 18 , wherein generating the multi-step difference image by the weighted subtraction of the combination of at least one of the second-step comparison images and at least two of the first-step images from the test image comprises:

generating the multi-step difference image (I MSDIFF,i ) based on the equation I MSDIFF,i =I SecondStep,i −Σ j≠i α j I SecondStep,j −Σ j β j I FirstStep,j , wherein I SecondStep,i corresponds to the test image, Σ j≠i α j I SecondStep,j corresponds to the at least one of the second-step comparison images, and Σ j≠i α j I FirstStep,j corresponds to the at least two of the first-step images, wherein subscripts i and j correspond to the sample regions, wherein α and β correspond to weights.

27. The inspection method of claim 18 , wherein generating the multi-step difference image by the weighted subtraction of the combination of at least one of the second-step comparison images and at least two of the first-step images from the test image comprises:

generating the multi-step difference image (I MSDIFF,i ) based on the equation I MSDIFF,i =I SecondStep,i −Σ j≠i α j I SecondStep,j −[I FirstStep,i −Σ j≠i β j I FirstStep,j ], where I SecondStep,i corresponds to the test image, Σ j≠i α j I SecondStep,j corresponds to the at least one of the second-step comparison images, and [I FirstStep,i −Σ j≠i β j I FirstStep,j ] corresponds to the at least one of the first-step images, wherein subscripts i and j correspond to the sample regions, wherein α and β correspond to weights.

28. The inspection method of claim 18 , wherein generating the multi-step difference image by the weighted subtraction of the combination of at least one of the second-step comparison images and at least two of the first-step images from the test image comprises:

generating the multi-step difference image (I MSDIFF,i ) based on the equation I MSDIFF,i =I SecondStep,i −Σ j≠i α j I SecondStep,j −γ[I FirstStep,i −Σ j≠i β j I FirstStep,j ], where I SecondStep,i corresponds to the test image, Σ j≠i α j I SecondStep,j corresponds to the at least one of the second-step comparison images, and [I FirstStep,i −Σ j≠i β j I FirstStep,j ] corresponds to the at least one of the first-step images, wherein subscripts i and j correspond to the sample regions, wherein α, β, and γ correspond to weights.

29. The inspection method of claim 18 , wherein detecting defects on the sample based on the multi-step difference image comprises:

providing the inspection reference image to a machine learning algorithm; and

detecting defects on the sample based on an output of the machine learning algorithm.

30. The inspection method of claim 18 , wherein detecting defects on the sample based on the multi-step difference image comprises:

detecting defects on the sample based on the multi-step difference image using a multi-die adaptive threshold (MDAT) technique.

31. The inspection method of claim 18 , wherein the defects comprise:

a deviation of at least one of a shape, size, or orientation of a feature fabricated by the second process step.

32. The inspection method of claim 18 , wherein the defects comprise:

an absence of a feature intended to be fabricated by the second process step.

33. The inspection method of claim 18 , wherein the defects comprise:

at least one of a scratch, a pit, or residual material in the inspection region after the second process step.

34. The inspection method of claim 18 , wherein the defects comprise:

at least one of an unintended bridge between two features fabricated by the second process step or an unintended break in a feature fabricated by the second process step.

35. An inspection system comprising:

an imaging sub-system including an illumination source and a detector configured to image a sample; and

a controller communicatively coupled to the imaging sub-system, the controller including one or more processors configured to execute program instructions causing the one or more processors to implement an inspection recipe by:

receiving first-step images of a plurality of sample regions after a first process step from the imaging sub-system;

receiving second-step images of the plurality of sample regions after a second process step, wherein the second process step modifies the sample in at least one of the plurality of sample regions from the imaging sub-system;

identifying one of the plurality of sample regions as a test region and at least some of the remaining sample regions as comparison regions, wherein the second-step image of the test region is a test image and the second-step images of the comparison regions are comparison images;

generating a multi-step difference image by a weighted subtraction of a combination of at least one of the second-step comparison images and at least two of the first-step images from the test image; and

identifying defects in the test region associated with the second process step based on the multi-step difference image.

36. The inspection system of claim 35 , wherein a number of the first-step images used to generate the multi-step difference image is one greater than a number of the second-step comparison images to provide that the multi-step difference image is generated based on a same number of the first-step and the second-step images.

37. The inspection system of claim 35 , wherein the imaging sub-system comprises:

a bright-field imaging sub-system.

38. The inspection system of claim 35 , wherein the imaging sub-system comprises:

a dark-field imaging sub-system.