Method and apparatus for identifying root cause of defect using composite defect map

• US 9,665,795 B2
• Filed: 12/31/2013
• Issued: 05/30/2017
• Est. Priority Date: 05/30/2013
• Status: Active Grant

• Current Assignee: Samsung SDS (China) Co., Ltd. (Samsung Group)
• Original Assignee: Samsung SDS (China) Co., Ltd. (Samsung Group)
• Inventors: Shin, Kae Young, Doo, Min Kyun
• Primary Examiner(s): Alhija, Saif
• Application Number:
  US14/144,881
• Publication Number:
  US 20140355866A1
• Time in Patent Office: 1,246 Days
• Field of Search: 703 6
• CPC Class Codes: G06T 2207/30164 Workpiece; Machine component
• CPC Class Codes: G06T 7/0004 Industrial image inspection
• CPC Class Codes: G06V 10/751 Comparing pixel values or logical combinations thereof, or feature values having positional relevance, e.g. template matching
• CPC Class Codes: G06V 2201/06 Recognition of objects for industrial automation

Abstract

Provided are a method and apparatus for automatically determining defective equipment by using a sample defect map showing defect distribution in each cell of a defective sample and production history information of each product, wherein the defective sample is a set of products, each being divided into a plurality of cells. According to this invention, the method of determining defective equipment is provided. The method comprises, generating a sample defect map which shows defect distribution in each cell of a defective sample, generating an equipment defect map for at least one of pieces of equipment through which each product of the defective sample passed, calculating, for each piece of equipment whose equipment defect map has been generated, map similarity between the sample defect map and the equipment defect map, and determining one or more defective pieces of equipment, which are the cause of the defective sample, based on the map similarity for each piece of equipment whose equipment defect map has been generated, wherein the defective sample is a set of products, each being divided into a plurality of cells, and the equipment defect map for a specific piece of equipment shows defect distribution in each cell of products that passed through the specific piece of equipment among the products of the defective sample.

Claims

1. A method of determining defective equipment, the method comprising:

generating a sample defect map which shows defect distribution in cells of a defective sample;

generating equipment defect maps of a plurality of pieces of equipment through which products of the defective sample passed;

calculating map similarities between the sample defect map and the equipment defect maps; and

determining one or more defective pieces of equipment of the plurality of pieces of equipment, which are the cause of the defective sample, based on the map similarities,

wherein the defective sample comprises the products, each of the products being divided into a plurality of cells, and one of the equipment defect maps for one of the plurality of pieces of equipment, shows a defect distribution in cells of the products that passed through the one of the plurality of pieces of equipment,

wherein the calculating of the map similarity for each piece of equipment whose equipment defect map has been generated comprises;

calculating cell similarity of each cell position of each product, for each piece of equipment that each product passed through, and whose equipment defect map has been generated; and

calculating map similarity for each piece of equipment whose equipment defect map has been generated,

wherein the cell similarity of a specific cell position is a ratio of the number of defective cells in the specific cell position of the equipment defect map to the number of defective cells in the specific cell position of the sample defect map, and the map similarity is the average of cell similarity values of at least some of the cell positions.

2. The method of claim 1, wherein the generating of the equipment defect maps comprises:

calculating defect rates for the plurality of pieces of equipment through which the products of the defective sample passed; and

generating the equipment defect maps for the plurality of pieces of equipment whose defect rates exceed a predetermined reference value,

wherein one of the defect rates for the one of the plurality of pieces of equipment is a ratio of a number of defective cells included in the products of the defective sample that passed through the one of the plurality of pieces of equipment to a total number of defective cells included in all the products of the defective sample.

3. The method of claim 1, wherein the calculating of the map similarities comprises calculating a map similarity, which is an average of cell similarity values of cell positions of interest selected based on the number of defective cells, for one of the plurality of pieces of equipment.

4. The method of claim 3, wherein in the calculating the map similarities comprises:

selecting a predetermined number of cell positions from all cell positions of the products in the defective sample in an order of a largest number of defective cells;

calculating a linear regression equation of the selected predetermined number of cell positions by using an order of the numbers of defective cells in each of predetermined number of cell positions as X axis variable, and the number of defective cells as Y variable;

calculating differences between actual numbers of defective cells of the selected predetermined number of cell positions and estimated numbers of defective cells obtained by the linear regression equation and calculating a standard deviation of the differences;

calculating a modified regression equation by adding or subtracting a predetermined multiple of the standard deviation to or from the linear regression equation;

selecting a high defective reference cell position from cell positions excluding the selected cell positions by using the modified regression equation; and

selecting all cell positions having numbers of defective cells equal to or greater than a number of defective cells of the high defective reference cell position, as the cell positions of interest.

5. The method of claim 3, wherein the calculating of the map similarities comprises:

selecting a predetermined number of cell positions having numbers of defective cells which rank at about a middle in an order of numbers of defective cells from all cell positions of the products in the defective sample;

calculating a linear regression equation of the selected predetermined number of cell positions by using an order of the numbers of defective cells in each of predetermined number of cell positions as the X axis variable, and the number of defective cells as the Y axis variable;

calculating differences between actual numbers of defective cells of the selected predetermined number of cell positions and estimated numbers of defective cells obtained by the linear regression equation and calculating a standard deviation of the differences;

calculating a first modified regression equation by adding a predetermined multiple of the standard deviation to the linear regression equation;

calculating a second modified regression equation by subtracting the predetermined multiple of the standard deviation from the linear regression equation;

selecting a second low defective reference cell position from cell positions excluding the selected cell positions by using the first modified regression equation;

selecting a first low defective reference cell position from the cell positions excluding the selected cell positions by using the second modified regression equation; and

selecting all cell positions having numbers of defective cells between the number of defective cells of the first low defective reference cell position and the second low defective reference cell position, as the cell positions of interest.

6. The method of claim 3, wherein the calculating of the map similarities comprises:

selecting a predetermined number of cell positions from all cell positions of the products in the defective sample in an order of largest number of defective cells;

calculating a linear regression equation of the selected predetermined number of cell positions by using an order of the numbers of defective cells in each of predetermined number of cell positions as the X axis variable, and the number of defective cells as the Y axis variable;

calculating differences between actual numbers of defective cells of the selected predetermined number of cell positions and estimated numbers of defective cells obtained by the linear regression equation and calculating a standard deviation of the differences;

calculating a modified regression equation by adding or subtracting a predetermined multiple of the standard deviation to or from the linear regression equation;

selecting a high defective reference cell position from cell positions excluding the selected cell positions by using the modified regression equation;

selecting all cell positions having numbers of defective cells equal to or greater than a number of defective cells of the high defective reference cell position, as high defective cell positions;

selecting a predetermined number of cell positions having numbers of defective cells which rank at about a middle in an order of the numbers of defective cells from all cell positions of the products in the defective sample;

calculating a linear regression equation of the selected predetermined number of cell positions by using an order of the numbers of defective cells in each of predetermined number of cell positions as the X axis variable, and the number of defective cells as the Y axis variable;

calculating differences between actual numbers of defective cells of the selected predetermined number of cell positions and estimated numbers of defective cells obtained by the linear regression equation and calculating a standard deviation of the differences;

calculating a first modified regression equation by adding a predetermined multiple of the standard deviation to the linear regression equation;

calculating a second modified regression equation by subtracting the predetermined multiple of the standard deviation from the linear regression equation;

selecting a second low defective reference cell position from cell positions excluding the selected cell positions by using the first modified regression equation;

selecting a first low defective reference cell position from the cell positions excluding the selected cell positions by using the second modified regression equation;

selecting all cell positions having numbers of defective cells between a number of defective cells of the first low defective reference cell position and a number of defective cells of the second low defective reference cell position as low defective cell positions; and

selecting all cell positions excluding the high defective cell positions and the low defective cell positions, as medium defective cell positions.

7. The method of claim 1, wherein the determining of the one or more defective pieces of equipment comprises:

calculating similarity ratios for the plurality of pieces of equipment;

determining one or more pieces of the plurality of pieces of equipment, which are the cause of the defective sample, based on the calculated similarity ratios,

wherein one of the similarity ratios is a ratio of one of the map similarities to a defect occupancy rate, and the defect occupancy rate for a specific piece of equipment of the plurality pieces of equipment is a proportion of the number of defective cells included in the products of the defective sample that passed through the specific piece of equipment in the total number of defective cells included in the sample defect map.

8. The method of claim 7, wherein the defect occupancy rate for a specific piece of equipment of the plurality of pieces of equipment is a proportion of the number of defective cells included in cell positions of interest of the products in the defective sample that passed through the specific piece of equipment in the number of defective cells included in the cell positions of interest of the products of the defective sample,

wherein the cell positions of interest are cell positions which are used in the calculation of the map similarities between the sample defect map and the equipment defect maps.

9. The method of claim 1, further comprising receiving criteria for selecting the defective sample and selecting products to be included in the defective sample from inspection result information of the products which are included in product inspection data.

10. A non-transitory machine readable storage medium on which a computer program is recorded, the computer program performing:

a process of generating a sample defect map which shows defect distribution in cells of a defective sample;

a process of generating equipment defect maps of a plurality of pieces of equipment through which the products of the defective sample passed;

a process of calculating map similarities between the sample defect map and the equipment defect maps; and

a process of determining one or more defective pieces of equipment of the plurality of pieces of equipment, which are the cause of the defective sample, based on the map similarities,

wherein the defective sample comprises the products, each of the products being divided into a plurality of cells, and one of the equipment defect maps for one of the plurality of pieces of equipment, shows a defect distribution in cells of the products that passed through the one of the plurality of pieces of equipment,

wherein the process of calculating of the map similarity for each piece of equipment whose equipment defect map has been generated comprises;

a process of calculating cell similarity of each cell position of each product, for each piece of equipment that each product passed through, and whose equipment defect map has been generated; and

a process of calculating map similarity for each piece of equipment whose equipment defect map has been generated,

wherein the cell similarity of a specific cell position is a ratio of the number of defective cells in the specific cell position of the equipment defect map to the number of defective cells in the specific cell position of the sample defect map, and the map similarity is the average of cell similarity values of at least some of the cell positions.

11. An apparatus for determining defective equipment, the apparatus comprising:

a sample defect map generator which is configured to generate a sample defect map which shows defect distribution in cells of a defective sample;

an equipment defect map generator which is configured to generate equipment defect maps of a plurality of pieces of equipment through which the products of the defective sample passed; and

a determiner which is configured to calculate map similarities between the sample defect map and the equipment defect maps and determining one or more defective pieces of equipment of the plurality of pieces of equipment, which are the cause of the defective sample, based on the map similarities,

wherein the defective sample comprises the products, each of the products being divided into a plurality of cells, and one of the equipment defect maps for one of the plurality of pieces of equipment, shows a defect distribution in cells of the products that passed through the one of the plurality of pieces of equipment,

wherein the determination unit calculating cell similarity of each cell position of each product, for each piece of equipment that each product passed through, and whose equipment defect map has been generated; and

calculating map similarity for each piece of equipment whose equipment defect map has been generated,

wherein the cell similarity of a specific cell position is a ratio of the number of defective cells in the specific cell position of the equipment defect map to the number of defective cells in the specific cell position of the sample defect map, and the map similarity is the average of cell similarity values of at least some of the cell positions.

12. The apparatus of claim 11, wherein each piece of equipment is identified by a combination of a corresponding process identifier and a corresponding equipment identifier.