AUTOMATIC FAULT DETECTION AND CLASSIFICATION IN A PLASMA PROCESSING SYSTEM AND METHODS THEREOF

• US 20120101622A1
• Filed: 06/29/2010
• Published: 04/26/2012
• Est. Priority Date: 06/30/2009
• Status: Active Grant

• Current Assignee: Lam Research Corporation
• Original Assignee: Lam Research Corporation
• Inventors: Yun, Gunsu, Venugopal, Vijayakumar
• Granted Patent: US 8,989,888 B2
• Time in Patent Office: Days
• US Class Current: 700/110
• CPC Class Codes: G06F 17/00 Digital computing or data processing equipment or methods, specially adapted for specific functions information retrieval, database structures or file system structures therefor G06F16/00
• CPC Class Codes: H01J 37/32935 Monitoring and controlling tubes by information coming from the object and/or discharge
• CPC Class Codes: H01J 37/3299 Feedback systems

Abstract

A method for automatically detecting fault conditions and classifying the fault conditions during substrate processing is provided. The method includes collecting processing data by a set of sensors during the substrate processing. The method also includes sending the processing data to a fault detection/classification component. The method further includes performing data manipulation of the processing data by the fault detection/classification component. The method yet also includes executing a comparison between the processing data and a plurality of fault models stored within a fault library. Each fault model of the plurality of fault models represents a set of data characterizing a specific fault condition. Each fault model includes at least a fault signature, a fault boundary, and a set of principal component analysis (PCA) parameters.

Claims

1. A method for automatically detecting fault conditions and classifying said fault conditions during substrate processing, comprising:

collecting processing data by a set of sensors during said substrate processing;

sending said processing data to a fault detection/classification component;

performing data manipulation of said processing data by said fault detection/classification component; and

executing a comparison between said processing data and a plurality of fault models stored within a fault library, wherein each fault model of said plurality of fault models representing a set of data characterizing a specific fault condition, wherein said each fault model including at least a fault signature, a fault boundary, and a set of principal component analysis (PCA) parameters.

7. The method of claim 1 wherein said each fault model is generated by

collecting a plurality of sensor data for a plurality of substrates,

assigning a designation to each set of sensor data of said plurality of sensor data, wherein said designation including one of a bad set of sensor data, a good set of sensor data, and an unknown set of sensor data, and

performing pre-filtering on a plurality of sensor channels to exclude a first set of sensor channels, wherein said first set of sensor channels does not include sensor data that contributes to fault detection, wherein said pre-filtering includes calculating summary statistical data for each channel of said plurality of sensor channels across said plurality of substrates.

8. The method of claim 7 wherein said designation of said each set of sensor data is based on characteristics associated with at least one of condition of components with said processing chamber and condition of a substrate associated with said each set of sensor data.

9. The method of claim 7 wherein said calculating of said summary statistical data includes applying local linear fit to sensor data for said each channel.

10. The method of claim 7 wherein said pre-filtering further includes calculating a variation in said summary statistical data.

2. The method of claim I wherein said data manipulation of said processing data by said fault detection/classification component includes performing summary statistic calculations for each sensor channel, wherein said performing of said summary statistic calculations includes at least performing pre-filtering on said set of sensor data to exclude a first set of sensor channels that is invariant to said fault conditions.

3. The method of claim 2 wherein said data manipulation further includes performing weighted PCA on a second set of sensor channels, wherein said second set of sensor channels does not include said first set of sensor channels.

4. The method of claim 3 wherein said data manipulation further includes generating a set of data samples represented in a multi-dimensional PCA space based on said set of PCA parameters.

5. The method of claim 4 wherein said comparison against said plurality of faults models includes a correlation of said processing data computed against a set of fault vectors associated with a first fault model.

6. The method of claim 5 wherein said comparison further includes determining if said processing data is within a fault boundary of said first fault model, wherein

if said processing data is within said fault boundary of said first fault model, generating a warning indicating a possible fault condition, and

if said processing data is outside of said fault boundary of said first fault model, identifying another fault model for performing comparison against said processing data.

11. An arrangement for automatically detecting fault conditions and classifying fault conditions during substrate processing within a processing chamber of a plasma processing system, comprising:

a set of sensors configured for monitoring said processing chamber and collecting processing data during substrate processing;

a fault library configured at least for storing a plurality of fault models, wherein each fault model of said plurality of fault models representing a set of data characterizing a specific fault condition; and

a fault detection/classification component configured for performing analysis on said processing data, wherein said analysis includes comparing said processing data against at least one fault model of said plurality of fault models, wherein each fault model including at least a fault signature, a fault boundary, and a set of principal component analysis (PCA) parameters.

12. The arrangement of claim 11 further including a controller configured for controlling said plasma processing system based on outputs of said fault detection/classification component.

13. The arrangement of claim 11 wherein said analysis performed by said fault detection/classification component includes performing summary statistic calculations for each data channel, wherein said performing of said summary statistic calculations includes at least performing pre-filtering on said set of sensor data to exclude a first set of sensor channels that is invariant to said fault conditions.

14. The arrangement of claim 13 wherein said analysis performed by said fault detection/classification component includes performing weighted PCA on a second set of sensor channels, wherein said second set of sensor channels does not include said first set of sensor channels.

15. The arrangement of claim 14 wherein said analysis performed by said fault detection/classification component includes generating a set of data samples represented in a multi-dimensional PCA space based on PCA parameters.

16. The arrangement of claim 15 wherein said analysis performed by said fault detection/classification component includes performing a correlation of said processing data computed against a set of fault vectors associated with a first fault model.

17. The arrangement of claim 16 wherein said analysis performed by said fault detection/classification component includes determining if said processing data is within a fault boundary of said first fault model, wherein

if said processing data is within said fault boundary of said first fault model, generating a warning indicating a possible fault condition, and

if said processing data is outside of said fault boundary of said first fault model, identifying another fault model for performing comparison against said processing data.

18. The arrangement of claim 11 wherein said each fault model is generated by

collecting a plurality of sensor data for a plurality of substrates,

assigning a designation to each set of sensor data of said plurality of sensor data, wherein said designation including one of a bad set of sensor data, a good set of sensor data, and an unknown set of sensor data, and

performing pre-filtering on a plurality of sensor channels to exclude a first set of sensor channels, wherein said first set of sensor channels does not include sensor data that contributes to fault detection, wherein said pre-filtering includes calculating summary statistical data for each channel of said plurality of sensor channels across said plurality of substrates.

19. The arrangement of claim 18 wherein said designation of said each set of sensor data is based on characteristics associated with at least one of condition of components with said processing chamber and condition of a substrate associated with said each set of sensor data.

20. The arrangement of claim 17 wherein said calculating of said summary statistical data includes applying local linear fit to sensor data for said each channel.