Context-based automated defect classification system using multiple morphological masks

• US 6,456,899 B1
• Filed: 12/07/1999
• Issued: 09/24/2002
• Est. Priority Date: 12/07/1999
• Status: Expired due to Fees

• Current Assignee: Lockheed Martin Energy Research Corporation (Martin Marietta Corporation)
• Original Assignee: UT-Battelle LLC
• Inventors: Hunt, Martin A., Gleason, Shaun S., Sari-Sarraf, Hamed
• Primary Examiner(s): Picard, Leo
• Assistant Examiner(s): Bahta, Kidest
• Application Number:
  US09/454,568
• Time in Patent Office: 1,022 Days
• Field of Search: 700/121, 700/165, 700/166, 700/164, 700/108, 700/109, 700/110, 700/115, 700/204, 382/149, 382/159, 706/900, 250/310
• US Class Current: 700/212
• CPC Class Codes: G03F 7/7065 Defects, e.g. optical inspection of patterned layer for defects
• CPC Class Codes: G05B 19/41875 characterised by quality surveillance of production
• CPC Class Codes: G05B 2219/32186 Teaching inspection data, pictures and criteria and apply them for inspection
• CPC Class Codes: G05B 2219/32196 Store audit, history of inspection, control and workpiece data into database
• CPC Class Codes: G05B 2219/45031 Manufacturing semiconductor wafers
• CPC Class Codes: Y02P 90/02 Total factory control, e.g. smart factories, flexible manufacturing systems [FMS] or integrated manufacturing systems [IMS]
• CPC Class Codes: Y10S 706/90 Fuzzy logic

Abstract

Automatic detection of defects during the fabrication of semiconductor wafers is largely automated, but the classification of those defects is still performed manually by technicians. This invention includes novel digital image analysis techniques that generate unique feature vector descriptions of semiconductor defects as well as classifiers that use these descriptions to automatically categorize the defects into one of a set of pre-defined classes. Feature extraction techniques based on multiple-focus images, multiple-defect mask images, and segmented semiconductor wafer images are used to create unique feature-based descriptions of the semiconductor defects. These feature-based defect descriptions are subsequently classified by a defect classifier into categories that depend on defect characteristics and defect contextual information, that is, the semiconductor process layer(s) with which the defect comes in contact. At the heart of the system is a knowledge database that stores and distributes historical semiconductor wafer and defect data to guide the feature extraction and classification processes. In summary, this invention takes as its input a set of images containing semiconductor defect information, and generates as its output a classification for the defect that describes not only the defect itself, but also the location of that defect with respect to the semiconductor process layers.

Claims

1. A digital computer for automatically classifying defects in a semiconductor wafer comprising:

means for identifying the background layers in a single reference semiconductor wafer image resulting in a layer identification image, said background layer identifying means including a background layer segmenter for dividing said reference semiconductor wafer image into different physical regions, said background layer identifying means further including a background layer classifier for classifying said physical regions of said reference semiconductor wafer image;

means for registering a defect semiconductor wafer image to said reference semiconductor image and subsequently detecting a defect in said defect semiconductor image;

means for classifying said detected defect using features of said defect;

a database for storing precomputed information and sharing information between said background layer classifier and said detected defect classifying means; and

means for associating said classified detected defect with said layer identification image resulting in a context-based classification of said defect.

2. The system of claim 1 wherein said means for identifying the background layers in a reference semiconductor image includes means for segmenting the background layers of said reference image and means for classifying said segmented background layers.

3. The system of claim 1 wherein said means for identifying the background layers in a reference semiconductor image includes means for extracting the hue, saturation and value levels from each pixel in said reference semiconductor image and a neural network for classifying said extracted hue, saturation and value levels.

4. The system of claim 1 wherein said background layer segmenter includes:

means for locating the edges of the semiconductor layer structures using a difference-of-Gaussians intensity gradient;

means for performing as colorspace conversion from the original red, green, blue colorspace to the excess red, green and blue colorspace; and

means for assigning all pixels within a connected edge boundary region a unique tag.

5. The system of claim 4 wherein said means for assigning all pixels within a connected edge boundary region is a means for extracting first-order statistics from the excess color values of said reference image.

6. The system of claim 4 wherein said means for assigning all pixels within a connected edge boundary region is a means for extracting fractal dimensions from the gray scale intensity of said reference image.

7. The system of claim 1 additionally including means for generating a full defect mask of said detected defect for use by said detected defect classifying means.

8. The system of claim 1 additionally including means for generating an edge defect mask of said detected defect for use by said detected defect classifying means.

9. The system of claim 1 additionally including means for generating a surround defect mask of said detected defect for use by said detected defect classifying means.

10. The system of claim 1 wherein said reference semiconductor image and said defect semiconductor images are captured in the focal plane of the semiconductor wafer surface.

11. The system of claim 1 wherein said reference semiconductor image and said defect semiconductor images are captured above the focal plane of the semiconductor wafer surface.

12. The system of claim 1 wherein said reference semiconductor image and said defect semiconductor images are captured below the focal plane of the semiconductor wafer surface.

13. The system of claim 1 wherein said means for classifying said detected defect using features of said defect includes means for extracting features using three different masks.

14. The system of claim 1 wherein said means for classifying said detected defect using features of said defect includes extracting features using three different masks at three different focus offsets.

15. A method for automatically classifying defects in a semiconductor wafer comprising the steps of:

identifying the background layers in a single reference semiconductor wafer image using a background layer identifying means resulting in a layer identification image, said step of identifying the background layers including the steps of segmenting said reference semiconductor wafer image into different physical regions, and classifying said physical regions of said reference semiconductor wafer image;

registering a defect semiconductor wafer image with respect to said reference semiconductor image and subsequently detecting a defect in said defect semiconductor image;

classifying said detected defect using features of said defect in a detected defect classifying means;

storing precomputed information in a database;

sharing information between said background layer identifying means and said detected defect classifying means via said database; and

associating said classified detected defect with said layer identification image resulting in a context-based classification of said defect.

16. The method of claim 15 wherein said step of identifying the background layers in a reference semiconductor image includes the steps of segmenting the background layers of said reference image and classifying said segmented background layers.

17. The method of claim 15 wherein said step of identifying the background layers in a reference semiconductor image includes the steps of extracting the hue, saturation and value levels from each pixel in said reference semiconductor image and classifying said extracted hue, saturation and value levels in a neural network.

18. The method of claim 15 wherein said step of segmenting said reference semiconductor wafer image into different physical regions includes the steps of:

locating the edges of the semiconductor layer structures using a difference-of-Gaussians intensity gradient;

performing a colorspace conversion from the original red, green, blue colorspace to the excess red, green and blue colorspace; and

assigning all pixels within a connected edge boundary region a unique tag.

19. The method of claim 18 wherein said step of assigning all pixels within a connected edge boundary region comprises the step of extracting first-order statistics from the excess color values of said reference image.

20. The method of claim 18 wherein said step of assigning all pixels within a connected edge boundary region comprises the step of extracting fractal dimensions from the gray scale intensity of said reference image.

21. The method of claim 15 additionally including the step of generating a full defect mask of said detected defect for use by said detected defect classifying means.

22. The method of claim 15 additionally including the step of generating an edge defect mask of said detected defect for use by said detected defect classifying means.

23. The method of claim 15 additionally including the step of generating a surround defect mask of said detected defect for use by said detected defect classifying means.

24. The method of claim 15 wherein said reference semiconductor image and said defect semiconductor images are captured in the focal plane of the semiconductor wafer surface.

25. The method of claim 15 wherein said reference semiconductor image and said defect semiconductor images are captured above the focal plane of the semiconductor wafer surface.

26. The method of claim 15 wherein said reference semiconductor image and said defect semiconductor images are captured below the focal plane of the semiconductor wafer surface.

27. The method of claim 15 wherein said step of classifying said detected defect using features of said defect includes means for extracting features using three different masks.

28. The method of claim 15 wherein said step of classifying said detected defect using features of said defect includes extracting features using three different masks at three different focus offsets.