System and method for classifying an anomaly

• US 6,483,938 B1
• Filed: 03/06/2000
• Issued: 11/19/2002
• Est. Priority Date: 05/31/1996
• Status: Expired due to Term

• Current Assignee: Rudolph Technologies Incorporated (Onto Innovation Inc.)
• Original Assignee: Texas Instruments, Inc.
• Inventors: Cleavelin, C. Rinn, Hennessey, A. Kathleen, Lin, YouLing, Hastings, Howard V. II, Wong, Wan S., Kinikoglu, Pinar, Reddy, Rajasekar
• Primary Examiner(s): Patel, Jayanti K.
• Application Number:
  US09/519,678
• Time in Patent Office: 988 Days
• Field of Search: 382/100, 382/143, 382/144-159, 382/181, 382/218, 382/224, 382/112, 382/141, 382/199, 382/305, 348/125, 348/128, 348/133, 356/237.1, 700/110, 700/117, 702/35, 702/118, 706/14, 706/15, 706/46, 706/1-2, 706/20, 716/1-4
• US Class Current: 382/149
• CPC Class Codes: G01N 21/95607 using a comparative method
• CPC Class Codes: G06T 2207/10056 Microscopic image
• CPC Class Codes: G06T 2207/30148 Semiconductor; IC; Wafer
• CPC Class Codes: G06T 7/001 using an image reference approach
• CPC Class Codes: H01L 22/12 for structural parameters, e.g. thickness, line width, refractive index, temperature, warp, bond strength, defects, optical inspection, electrical measurement of structural dimensions, metallurgic measurement of diffusions electrical measurement of diffusions H01L22/14
• CPC Class Codes: H01L 22/20 Sequence of activities consisting of a plurality of measurements, corrections, marking or sorting steps
• CPC Class Codes: H01L 2924/014 Solder alloys

Abstract

A method and system for generating and managing a knowledgebase for use in identifying anomalies on a manufactured object, such as a semiconductor wafer, includes measures for adding, deleting, and organizing data from the knowledgebase.

Claims

1. A method for classifying an anomaly of an object, the method comprising:

preparing a primitive-based representation of an object having an anomaly;

comparing the primitive-based representation of the object to a primitive-based reference image to detect and locate the anomaly;

selecting a descriptor for the anomaly, the descriptor associated with a plurality of descriptor values;

determining a first descriptor value of the anomaly using the primitive-based representation;

comparing the first descriptor value with a second descriptor value associated with a defect class;

determining a match between the first descriptor value and the second descriptor value; and

classifying the anomaly as a defect according to the defect class.

2. The method of claim 1, wherein the second descriptor value is associated with a rule stored in a knowledgebase.

3. The method of claim 1, further comprising:

associating a confidence level with the defect class; and

determining the match of the first descriptor value with the second descriptor value using the confidence level.

4. The method of claim 1, further comprising:

associating a weight with the second descriptor value, the weight indicating a significance of the second descriptor value; and

determining the match of the first descriptor value with the second descriptor value using the weight.

5. The method of claim 1, wherein the defect class is a first defect class, and further comprising:

associating a secondary validation table with the first defect class, the secondary validation table operable to indicate an attribute of the first defect class operable to distinguish the first defect class from a second defect class; and

determining the match of the first descriptor value with the second descriptor value using the secondary validation table.

6. The method of claim 1, further comprising:

determining whether the defect class is correct for the anomaly; and

if the defect class is not correct;

determining that the anomaly is associated with a new defect class;

creating a new rule for the new defect class; and

storing the new rule in a knowledgebase.

7. A method of indexing an anomaly, the method comprising:

preparing a primitive-based representation of an object having an anomaly;

comparing the primitive-based representation of the object to a primitive-based reference image to detect and locate the anomaly;

creating a file for the anomaly;

selecting a descriptor having a descriptor value associated with the anomaly;

creating a subdirectory for the file, the subdirectory associated with the descriptor; and

using the descriptor value to name the subdirectory to produce a file name for the file of the anomaly.

8. The method of claim 7, further comprising creating a pointer from the file name to an image file comprising an image of the anomaly.

9. The method of claim 7, wherein the subdirectory is a first subdirectory, the defect descriptor is a first defect descriptor, the descriptor value is a first descriptor value, the file name is a first file name, and further comprising:

creating a second subdirectory for the file, the second subdirectory associated with a second defect descriptor; and

using a second descriptor value to name the second subdirectory to produce a second file name for the file of the anomaly.