Method of automatic defect classification

Patent metadata

Field	Value
Publication number	US-9727799-B2
Application number	US-201514819566-A
Country	US
Kind code	B2
Filing date	Aug 6, 2015
Priority date	Dec 22, 2014
Publication date	Aug 8, 2017
Grant date	Aug 8, 2017

Abstract

A method of automatic defect classification (ADC) includes detecting defective parts from a substrate wherein at least one unit process is performed; and classifying defect types of the respective defective parts, wherein the classifying includes obtaining a scanning electron microscope (SEM) image of each of the defective parts; registering information about the substrate in a graphic data system (GDS) image corresponding to each SEM image; defining a plurality of defects of interest (DOIs) categorizing defects of the respective defective parts; defining a DOI rule that is a criterion for determining which defects of the respective defective parts correspond to which DOI from among the DOIs; and analyzing the image to classify which defects of the respective defective parts correspond to which DOI from among the DOIs according to the DOI rule.

First claim

What is claimed is:

1. An automatic defect classification (ADC) method, comprising:
detecting defective parts on a substrate upon which at least one semiconductor manufacturing process has been performed; and
classifying defect types of the respective defective parts, wherein classifying comprises:
obtaining a scanning electron microscope (SEM) image of each of the defective parts;
registering information about the substrate in a graphic data system (GDS) image corresponding to each SEM image;
defining a plurality of defects of interest (DOIs);
defining a DOI rule for determining which defects of the defective parts comprise a DOI from among the plurality of DOIs; and
analyzing the SEM image and the GDS image according to the DOI rule to classify which of the defects of the respective defective parts correspond to which of the DOIs.

2. The method of claim 1 , wherein the information about the substrate comprises information about one of an insulating line pattern, a conductive line pattern, and a conductive dot pattern.

3. The method of claim 2 , wherein the information about the substrate comprises material information about one of the insulating line pattern, the conductive line pattern, and the conductive dot pattern.

4. The method of claim 3 , wherein analyzing comprises:
obtaining a matched region and an unmatched region by matching the SEM image with the GDS image; and
analyzing the unmatched region.

5. The method of claim 4 , wherein analyzing the unmatched region comprises:
comparing and analyzing material information about the unmatched region with material information registered in the GDS image in a region corresponding to the unmatched region.

6. The method of claim 3 , wherein the information about the substrate comprises material information about the conductive line pattern,
wherein at least one DOI from among the plurality of DOIs comprises an insulating line bridge, and
wherein the DOI rule is defined as when the material information registered in the GDS image of the region corresponding to the unmatched region is the same as the material information about the conductive line pattern, the DOI is classified as the insulating line bridge.

7. The method of claim 3 , wherein the information about the substrate comprises material information about the insulating line pattern,
wherein at least one DOI from among the plurality of DOIs comprises a conductive line bridge, and
wherein the DOI rule is defined as when the material information registered in the GDS image of the region corresponding to the unmatched region is the same as the material information about the insulating line pattern, the DOI is classified as the conductive line bridge.

8. The method of claim 3 , wherein the information about the substrate comprises material information about the conductive dot pattern and material information about a barrier layer surrounding the conductive dot pattern,
wherein at least one DOI from among the plurality of DOIs comprises a conductive dot void, and
wherein in order to classify the DOI into the conductive dot void, the DOI rule is defined as when the unmatched region is positioned in the barrier layer while the material information registered in the GDS image of the region corresponding to the unmatched region is the same as the material information about the conductive dot pattern, the DOI is classified as the conductive dot void.

9. The method of claim 1 , wherein defining the DOI rule comprises:
defining a temporary DOI rule;
determining reliability of the temporary DOI rule; and
defining the temporary DOI rule as a fixed DOI rule when it is determined that a defect classification result by the temporary DOI rule reaches or exceeds a criteria reliability value.

10. The method of claim 1 , further comprising:
analyzing a cause of a defect, based on a result of classifying defect types of the respective defective parts; and
correcting a process recipe of the at least one semiconductor manufacturing process based on the cause of the defect.

11. An automatic defect classification (ADC) method, comprising:
detecting defective parts on a substrate upon which at least one semiconductor manufacturing process has been performed;
a first classifying defect types of the respective defective parts by using a feature vector, and
a second classifying the defect types of the respective defective parts by using a scanning electron microscope (SEM) image, wherein the second classifying comprises:
defining a plurality of defects of interest (DOIs);
defining a DOI rule for determining which defects of the defective parts comprise a DOI from among the plurality of DOIs; and
analyzing the SEM image and a graphic data system (GDS) image according to the DOI rule and classifying which of the defects correspond to which of the DOIs.

12. The method of claim 11 , wherein detecting defective parts comprises obtaining an optical image of the substrate via an optical inspection apparatus and generating a substrate map,
the method further comprising:
a third classifying the defect types on the substrate map.

13. The method of claim 12 , wherein classifying the defect types on the substrate map comprises:
defining a plurality of map DOIs categorizing the defect types on the substrate map; and
defining a map DOI rule for determining which defects on the substrate map comprise a DOI from among the plurality of DOIs.

14. The method of claim 11 , wherein analyzing comprises:
obtaining respective SEM images of the defective parts; and
registering substrate information in respective GDS images corresponding to the SEM images.

15. The method of claim 11 , wherein the first classifying precedes the second classifying.

16. An automatic defect classification method (ADC), comprising:
detecting defective parts on a substrate upon which at least one semiconductor manufacturing process has been performed, the detecting comprising obtaining an optical image of the substrate via an optical inspection apparatus and generating a substrate map; and
classifying defect types of the respective defective parts by simultaneously using a feature vector and a scanning electron microscope (SEM) image of each of the defective parts.

17. The method of claim 16 , wherein classifying the defect types of the respective defective parts by using an SEM image comprises:
defining a plurality of defects of interest (DOIs);
defining a DOI rule for determining which defects of the defective parts comprise a DOI from among the plurality of DOIs; and
analyzing the SEM image and a graphic data system (GDS) image according to the DOI rule to classify which of the defects of the respective defective parts correspond to which of the DOIs, wherein the GDS image comprises information about the substrate.

18. The method of claim 17 , wherein the information about the substrate comprises information about one of an insulating line pattern, a conductive line pattern, and a conductive dot pattern.

19. The method of claim 16 , wherein classifying defect types of the respective defective parts comprises sequentially using the substrate map, the feature vector and the SEM image of each of the defective parts.

20. The method of claim 16 , wherein classifying defect types of the respective defective parts comprises simultaneously using the feature vector, the SEM image of each of the defective parts, and the substrate map.

Assignees

• Samsung Electronics Co Ltd

Inventors

• Park Il-Suk
• Ahn Jeong-Ho
• Cho Hyung-Suk

Classifications

• G06T7/001 Primary

  using an image reference approach · CPC title

• G06V10/75

  Organisation of the matching processes, e.g. simultaneous or sequential comparisons of image or video features; Coarse-fine approaches, e.g. multi-scale approaches; using context analysis; Selection of dictionaries · CPC title

• G06F18/22

  Matching criteria, e.g. proximity measures · CPC title

• G06V10/759

  Region-based matching · CPC title

• G06T2207/30148

  Semiconductor; IC; Wafer · CPC title