Semiconductor run-to-run control system with state and model parameter estimation

• US 6,748,280 B1
• Filed: 06/14/2002
• Issued: 06/08/2004
• Est. Priority Date: 10/23/2001
• Status: Expired due to Fees

• Current Assignee: Applied Materials Incorporated
• Original Assignee: Brooks Automation Incorporated (Azenta, Inc.)
• Inventors: Zou, Jianping, Edwards, Keith A., Mullins, James A.
• Primary Examiner(s): Picard, Leo
• Assistant Examiner(s): JARRETT, RYAN A
• Application Number:
  US10/171,758
• Time in Patent Office: 725 Days
• Field of Search: 700/29-31, 700/44, 700/45, 700/52, 700/121
• US Class Current: 700/31
• CPC Class Codes: G05B 13/042 in which a parameter or coefficient is automatically adjusted to optimise the performance
• CPC Class Codes: G05B 13/048 using a predictor
• CPC Class Codes: G05B 15/02 electric
• CPC Class Codes: G05B 17/02 electric
• CPC Class Codes: H01L 22/20 Sequence of activities consisting of a plurality of measurements, corrections, marking or sorting steps

Abstract

A method for a run-to-run (R2R) control system includes processing materials using a process input and producing a process output, storing the process input in a database, the storing including using a timestamp, and storing at least one measurement of the process output in the database aligned with each process input using the timestamp. The method further includes iterating over the data in the database to estimate one or more coefficients for a model, and, if one or more measurements is missing, replacing the missing measurements based on a prediction from said model. The model is updated with said coefficient estimates. The method additionally includes iterating over the data from the database to estimate a process state, and, if one or more of the measurements is missing from the database, replacing the missing measurements based on prediction from the model. The model is updated with said process state estimate. A controller may receive the updated model and utilize the model to produce the next process input. The updated model may also be utilized to generate an estimate for a measurable process variable, wherein the estimate can be compared to an actual measurement to determine if the estimate is within confidence limits. If the estimate is not within confidence limits, a fault is indicated.

Claims

1. A method for controlling a manufacturing process, includes:

processing using a process input and producing a process output;

storing the process input in a database;

storing at least one measurement of the process output in the database associated with each process input;

iterating over data from the database to estimate one or more coefficients for a model;

if one or more measurements is missing from the database during model coefficient estimation, replacing the missing measurements for the database based on a prediction from said model;

updating said model with said coefficient estimates;

iterating over data from the database to estimate a process state;

if one or more measurements is missing from the database during process state estimation, replacing the missing measurements for the database based on a prediction from said model; and

updating said model with said process state estimate.

3. The method of claim 1 wherein a controller receives the updated model and utilizes said updated model to produce the next process input.

4. The method of claim 3 wherein the controller comprises one or more of a model-predictive control (MPC) type controller, a direct model inverter type controller, and a simple integral controller.

5. The method of claim 3 wherein the controller minimizes the error in each run of the process using the estimate of the process state.

6. The method of claim 1 wherein the process operates on either a lot or a part of a lot of semiconductor devices.

7. The method of claim 1 wherein the database is coupled to one or more modules that operate on the database.

8. The method of claim 7 wherein the modules include an estimation module, a compute module, a sort module, a pairing module and a data storage module.

9. The method of claim 1 wherein said model coefficient estimation and said process state estimation includes providing a confidence limit.

10. The method of claim 9 further including utilizing said updated model to generate an estimate for a measurable process variable, and comparing an actual measurement to said estimate to determine if said estimate is within said confidence limits, wherein if said estimate is not within said confidence limits, a fault is indicated.

12. The method of claim 1 wherein the replacing of missing measurements for the database is accomplished either explicitly or implicitly.

13. The method of claim 1 further including solving a constrained quadratic optimization at each iteration of said data.

14. The method of claim 1 wherein said storing of the process input includes providing a timestamp, and wherein said at least one measurement of the process output is associated with each process input using the timestamp.

35. A program storage media readable by a machine and containing instructions for performing the method contained in claim 1.

2. A method for controlling a manufacturing process, includes:

processing using a process input and producing a process output;

storing the process input in a database;

storing at least one measurement of the process output in the database associated with each process input;

iterating over data from the database to estimate one or more coefficients for a model;

if one or more measurements is missing from the database during model coefficient estimation, replacing the missing measurements for the database based on a prediction from said model;

updating said model with said coefficient estimates;

iterating over data from the database to estimate a process state;

if one or more measurements is missing from the database during process state estimation, replacing the missing measurements for the database based on a prediction from said model; and

updating said model with said process state estimate; and

wherein if actual measurements become available a time period after model coefficient estimation or process state estimation, for said one or more measurements which was missing, replacing the predicted measurements for the previously missing measurements with said actual measurements, storing said actual measurements in said database, and utilizing said actual measurements in subsequent model coefficient estimation and process state estimation.

11. The method of claim 2 wherein said time period can be variable.

15. A method for controlling a manufacturing process, includes:

processing using a process input and producing a process output;

storing the process input in a database;

storing at least one measurement of the process output in the database associated with each process input;

iterating over data from the database to estimate one or more coefficients for a model, wherein a constrained quadratic optimization is solved at each iteration of said data;

if one or more measurements is missing from the database during model coefficient estimation, replacing the missing measurements for the database based on a prediction from said model;

updating said model with said coefficient estimates;

iterating over data from the database to estimate a process state, wherein a constrained quadratic optimization is solved at each iteration of said data;

if one or more measurements is missing from the database during process state estimation, replacing the missing measurements for the database based on a prediction from said model; and

updating said model with said process state estimate.

17. The method of claim 15 wherein a controller receives the updated model and utilizes said updated model to produce the next process input.

18. The method of claim 17 wherein the controller comprises one or more of a model-predictive control (MPC) type controller, a direct model inverter type controller, and a simple integral controller.

19. The method of claim 17 wherein the controller minimizes the error in each run of the process using the estimate of the process state.

20. The method of claim 15 wherein the process operates on either a lot or a part of a lot of semiconductor devices.

21. The method of claim 15 wherein the database is coupled to one or more modules that operate on the database.

22. The method of claim 21 wherein the modules include an estimation module, a compute module, a sort module, a pairing module and a data storage module.

23. The method of claim 15 wherein said model coefficient estimation and said process state estimation includes providing a confidence limit.

24. The method of claim 23 further including utilizing said updated model to generate an estimate for a measurable process variable, and comparing an actual measurement to said estimate to determine if said estimate is within said confidence limits, wherein if said estimate is not within said confidence limits, a fault is indicated.

26. The method of claim 15 wherein the replacing of missing measurements for the database is accomplished either explicitly or implicitly.

27. The method of claim 15 wherein said storing of the process input includes providing a timestamp, and wherein said at least one measurement of the process output is associated with each process input using the timestamp.

36. A program storage media readable by a machine and containing instructions for performing the method contained in claim 15.

16. A method for controlling a manufacturing process, includes:

processing using a process input and producing a process output;

storing the process input in a database;

storing at least one measurement of the process output in the database associated with each process input;

iterating over data from the database to estimate one or more coefficients for a model, wherein a constrained quadratic optimization is solved at each iteration of said data;

if one or more measurements is missing from the database during model coefficient estimation, replacing the missing measurements for the database based on a prediction from said model;

updating said model with said coefficient estimates;

iterating over data from the database to estimate a process state, wherein a constrained quadratic optimization is solved at each iteration of said data;

if one or more measurements is missing from the database during process state estimation, replacing the missing measurements for the database based on a prediction from said model; and

updating said model with said process state estimate; and

wherein if actual measurements become available a time period after model coefficient estimation or process state estimation, for said one or more measurements which was missing, replacing the predicted measurements for the previously missing measurements with said actual measurements, storing said actual measurements in said database, and utilizing said actual measurements in subsequent model coefficient estimation and process state estimation.

25. The method of claim 16 wherein said time period can be variable.

28. A method for controlling a manufacturing process, includes:

providing a database with process input data and associated process output data;

iterating over data from the database to estimate one or more coefficients for a model;

if one or more measurements is missing from the database during model coefficient estimation, replacing the missing measurements for the database based on a prediction from said model;

updating said model with said coefficient estimates;

iterating over data from the database to estimate a process state;

if one or more measurements is missing from the database during process state estimation, replacing the missing measurements for the database based on a prediction from said model; and

updating said model with said process state estimate.

30. The method of claim 28 wherein said model coefficient estimation and said process state estimation includes providing a confidence limit.

31. The method of claim 30 further including utilizing said updated model to generate an estimate for a measurable process variable, and comparing an actual measurement to said estimate to determine if said estimate is within said confidence limits, wherein if said estimate is not within said confidence limits, a fault is indicated.

33. The method of claim 28 wherein the replacing of missing measurements for the database is accomplished either explicitly or implicitly.

34. The method of claim 28 further including solving a constrained quadratic optimization at each iteration of said data.

37. A program storage media readable by a machine and containing instructions for performing the method contained in claim 28.

29. A method for controlling a manufacturing process, includes:

providing a database with process input data and associated process output data;

iterating over data from the database to estimate one or more coefficients for a model;

if one or more measurements is missing from the database during model coefficient estimation, replacing the missing measurements for the database based on a prediction from said model;

updating said model with said coefficient estimates;

iterating over data from the database to estimate a process state;

if one or more measurements is missing from the database during process state estimation, replacing the missing measurements for the database based on a prediction from said model; and

updating said model with said process state estimate; and

wherein if actual measurements become available a time period after model coefficient estimation or process state estimation, for said one or more measurements which was missing, replacing the predicted measurements for the previously missing measurements with said actual measurements, storing said actual measurements in said database, and utilizing said actual measurements in subsequent model coefficient estimation and process state estimation.

32. The method of claim 29 wherein said time period can be variable.