Method and apparatus for classifying manufacturing outputs

• US 7,313,454 B2
• Filed: 12/02/2005
• Issued: 12/25/2007
• Est. Priority Date: 12/02/2005
• Status: Expired due to Fees

• Current Assignee: MKS Instruments Incorporated
• Original Assignee: MKS Instruments Incorporated
• Inventors: Hendler, Lawrence, Richter, Matthew
• Primary Examiner(s): Picard, Leo
• Assistant Examiner(s): Lee, Douglas S
• Application Number:
  US11/292,485
• Publication Number:
  US 20070129836A1
• Time in Patent Office: 753 Days
• Field of Search: 700/73, 700/101, 700/110, 702/179, 702/182, 702/183, 702/185
• US Class Current: 700/110
• CPC Class Codes: G05B 23/0281 Quantitative, e.g. mathematical distance; Clustering; Neural networks; Statistical analysis

Abstract

A method for process monitoring involves acquiring data samples associated with a plurality of manufacturing related variables for a plurality of outputs of a manufacturing process. The distance of each data sample relative to every other data sample is then calculated. The outputs are then grouped based on the Euclidean distances that satisfy a boundary determining criterion.

Claims

1. A method for process monitoring, comprising:

a) acquiring data samples associated with a plurality of manufacturing related variables for outputs of a manufacturing process;

b) calculating the distance of each data sample relative to every other data sample; and

c) grouping outputs based on the distances that satisfy a boundary determining criterion.

2. The method of claim 1 wherein the distance is selected from the group consisting of Euclidean distance, Mahalanobis distance and Manhattan distance.

3. The method of claim 1 wherein the boundary determining criterion comprises grouping data samples located within a predefined distance relative to each other.

4. The method of claim 1, comprising associating a classification with a group.

5. The method of claim 4 wherein the classification is a manufacturing defect.

6. The method of claim 1, comprising plotting the plurality of data samples as points in a phase space.

7. The method of claim 6 wherein calculating the distance comprises calculating the Euclidean distance of each point in phase space relative to every other point in phase space.

8. The method of claim 1 wherein the plurality of manufacturing related variables is one or more states of the manufacturing process.

9. The method of claim 1, comprising computing mathematical transformations of the data samples acquired for each output.

10. The method of claim 1 wherein computing mathematical transformations comprises computing principal components and loadings associated with the data samples acquired for each output.

11. The method of claim 9, comprising calculating the Euclidean distance of each data sample from every other data sample based on the mathematical transformations.

12. The method of claim 1 wherein the outputs of the manufacturing process are wafers produced by a semiconductor manufacturing process.

13. The method of claim 1 wherein the boundary defining criterion is selected by a user.

14. The method of claim 3 wherein the predefined distance is based on an analytical, experimental, or empirical model.

15. The method of claim 1, comprising identifying classification of outputs of a manufacturing process based on the Euclidean distance of a data sample relative to the groups.

16. The method of claim 1, comprising further grouping the groups based on whether the groups have common characteristics.

17. The method of claim 1, comprising developing a model for predicting output class of a new manufacturing output based on calculating the Euclidean distance of the data samples of the new manufacturing output relative to the groups.

18. The method of claim 1, comprising modifying the boundary determining criterion based on new manufacturing outputs.

19. A method for semiconductor process monitoring, comprising:

a) acquiring data samples associated with a plurality of manufacturing related variables for outputs of a semiconductor manufacturing process;

b) calculating the distance of each data sample relative to every other data sample; and

c) grouping outputs based on the distances that satisfy a boundary determining criterion.

20. A system for process monitoring, comprising:

a) means for acquiring data samples associated with a plurality of manufacturing related variables for outputs of a manufacturing process;

b) means for calculating the distance of each data sample relative to every other data sample; and

c) means for grouping outputs based on the distances that satisfy a boundary determining criterion.