Run-to-run controller for use in microelectronic fabrication

• US 6,587,744 B1
• Filed: 06/20/2000
• Issued: 07/01/2003
• Est. Priority Date: 06/22/1999
• Status: Expired due to Fees

• Current Assignee: Applied Materials Incorporated
• Original Assignee: Brooks Automation Incorporated (Azenta, Inc.)
• Inventors: Tsakalis, Konstantinos, Stoddard, Kevin D., Schulze, Bradley D.
• Primary Examiner(s): Follansbee, John
• Assistant Examiner(s): PHAM, THOMAS K
• Application Number:
  US09/599,357
• Time in Patent Office: 1,106 Days
• Field of Search: 700/29, 700/30, 700/31, 700/121, 700/95, 700/7, 700/17, 700/45, 438/5, 438/14, 438/17, 714/51
• US Class Current: 700/121
• CPC Class Codes: G05B 19/41885 characterised by modeling, simulation of the manufacturing system
• CPC Class Codes: G05B 2219/31086 Communication of carriage, agv data, workpiece data at each station
• CPC Class Codes: G05B 2219/31104 Remote configuration of parameters of controlled devices
• CPC Class Codes: G05B 2219/31422 Upload, download programs, parameters from, to station to, from server
• CPC Class Codes: G05B 2219/32388 Autonomous flexible system, cells and agv autonomous
• CPC Class Codes: Y02P 90/02 Total factory control, e.g. smart factories, flexible manufacturing systems [FMS] or integrated manufacturing systems [IMS]

Abstract

A automated run-to-run controller for controlling manufacturing processes comprises set of processing tools, a set of metrology tools for taking metrology measurements from the processing tools, and a supervising station for managing and controlling the processing tools. The supervising station comprises an interface for receiving metrology data from the metrology tools and a number of variable parameter tables, one for each of the processing tools, collectively associated with a manufacturing process recipe. The supervising station also includes one or more internal models which relate received metrology data to one or more variables for a processing tool, and which can modify variables stored in the variable parameter table to control the process tools using feedback and/or feed-forward control algorithms. Feed-forward control algorithms may, in certain embodiments, be used to adjust process targets for closed loop feedback control. The supervising station may have a user interface by which different feedback or feed-forward model formats (single or multi-variate) may be interactively selected based upon experimental or predicted behavior of the system, and may also permit users to utilize their own models for run-to-run control.

Claims

1. A run-to-run control system for controlling manufacturing processes, comprising:

a plurality of processing tools;

a plurality of metrology tools for monitoring operation of said processing tools each metrology tool adapted to obtain metrology data from a corresponding processing tool; and

a supervising station, said supervising station comprising;

an interface for receiving metrology data from each of said metrology tools;

a memory;

a plurality of variable parameter tables, one for each of said processing tools, stored in said memory, said variable parameter tables collectively associated with a manufacturing process or recipe, each variable parameter table being downloaded to a respective processing tool prior to operation of the respective processing tool; and

at least one model structure relating received metrology data received from a prior metrology tool to a target set-point for a subsequent processing tool, and another model structure adapted to control operation of the subsequent processing tool in a feed-back control loop, wherein one or more variables of a subsequent variable parameter tables associated with the subsequent processing tool are modified in response to application of said received metrology data according to said at least one model structure.

2. The run-to-run controller of claim 1, wherein said at least one model structure comprises a feedback control model.

3. The run-to-run controller of claim 1, wherein said at least one model structure comprises a feed-forward control model.

4. The run-to-run controller of claim 1, wherein said at least one model structure comprises both a feedback control model and a feed-forward control model.

5. The run-to-run controller of claim 1, further comprising a user interface for selecting one or more model formats from which said at least one model structure is generated.

6. The run-to-run controller of claim 1, further comprising an interface for receiving said at least one model structure from an external plug-in unit.

7. The run-to-run controller of claim 1, wherein all or part of each variable parameter table is downloaded to its corresponding processing tool prior to operation of the processing tool, and wherein said variables modified in response to said received metrology data are automatically downloaded to said processing tools without user intervention being required.

8. The run-to-run controller of claim 4, wherein said supervising station further comprises a target set-point associated with said feedback control model, said supervisory station adjusting said target set-point based upon an output of said feed-forward control model.

9. The run-to-run controller of claim 5, wherein said one or more model formats are interactively selectable from a plurality of predefined model formats, said predefined model formats comprising a linear model format, a quadratic model format and a cubic model format.

10. The run-to-run controller of claim 9, wherein a fading-memory least-squares algorithm is used to determine model parameters for said at least one model structure from experimental data.

11. The run-to-run controller of claim 9, wherein said at least one model structure comprises a multi-input, multi-output model.

12. The run-to-run controller of claim 10, wherein said at least one model structure is user-adjustable with a dead-zone non-linear gain adjustment.

13. The run-to-run controller of claim 10, wherein said at least one model structure may be user-defined as either least squares adaptive, gradient adaptive or non-adaptive.

14. The run-to-run controller of claim 10, wherein process noise and model integrity metrics for said at least one model structure are displayed by said user interface.

15. The run-to-run controller of claim 10, wherein said model parameters for said at least one model structure are user-adjustable between runs.

16. A method for controlling a manufacturing process, comprising the steps of:

(a) downloading variables form a plurality of variable parameter tables to a plurality of processing tools prior to operation of each tool;

(b) operating said processing tools according to said downloaded variables;

(c) receiving, at a supervisory station, metrology data from a plurality of metrology tools monitoring operation of said processing tools;

(d) applying said metrology data to at least one model structure relating said metrology data to said variables including a target set-point for a next processing tool, and generating an output thereby; and

(e) updating one or more of said variable parameter tables in response to said output and controlling operation of the next processing tool using another model structure associated with the next processing tool in a feed-back control loop.

17. The method of claim 16, wherein steps (a) through (e) are repeated to effectuate run-to-run control of the manufacturing process.

18. The method of claim 16, wherein step (d) comprises the step of applying said metrology data to a feedback model.

19. The method of claim 16, wherein step (d) comprises the step of applying said metrology data to a feed-forward model.

20. The method of claim 16, wherein step (d) comprises the step of applying said metrology data to both a feedback model and a feed-forward model.

21. The method of claim 16, wherein step (c) comprises the step of automatically transferring said metrology data from said metrology tools to said supervisory station.

22. The method of claim 16, further comprising the step of setting a target set-point for said at least one model structure.

23. The method of claim 16, wherein each of said variable parameter tables is associated with exactly one of said processing tools.

24. The method of claim 16, wherein each of said metrology tools is associated with exactly one of said processing tools.

25. The method of claim 16, further comprising the step of interactively selecting said at least one model structure via a user interface from a plurality of predefined model formats.

26. The method of claim 25, wherein said predefined model formats comprise a linear model format, a quadratic model format and a cubic model format.

27. The method of claim 22, wherein step (e) comprises the step of adjusting said one or more of said variable parameter tables in response to a comparison between said output and said target set-point.

28. The method of claim 25, wherein said at least one model structure may be user-defined as either least squares adaptive, gradient adaptive or non-adaptive.

29. The method of claim 25, further comprising the step of displaying process noise and model integrity metrics at said user interface.

30. The method of claim 25, wherein said at least one model structure comprises a multi-input, multi-output model.

31. The method of claim 26, further comprising the step of determining model parameters for said at least one model structure from experimental data using a fading-memory least-squares algorithm.

32. The method of claim 26, further comprising the step of adjusting said model parameters for said at least one model structure between runs.

33. A run-to-run controller for controlling a manufacturing process, comprising:

a first processing tool;

a first metrology tool for obtaining metrology data from said first processing tool;

a second processing tool;

a second metrology tool for obtaining metrology data from said second processing tool; and

a supervisory station, said supervisory station comprising

an interface for receiving said metrology data from said first metrology tool; and

a first model structure relating said metrology data from said first metrology tool to a target set-point for said second processing tool;

a second model structure used in controlling operation of said second processing tool in a feed-back control loop; and

a first variable parameter table for said first processing tool and a second variable parameter table for said second processing tool;

wherein all or part of said first variable parameter table is downloaded to said first processing tool prior to operation of said first processing tool, wherein all or part of said second variable parameter table is downloaded to said second processing tool prior to operation of said second processing tool, and wherein one or more variables in said second variable parameter table are modified in response to application of said first model structure to said received metrology data.

34. The run-to-run controller of claim 33, wherein said supervisory station adjusts table parameters in said second variable parameter table in response to the metrology data from said second metrology tool in order to maintain operation of said second processing tool at a desired target point.

35. The run-to-run controller of claim 33, further comprising an interface for receiving said first model structure or said second model structure, or both, from an external plug-in unit.

36. The run-to-run controller of claim 33, further comprising a user interface for selecting one of a plurality of predefined model formats from which said first model structure is generated.

37. The run-to-run controller of claim 36, wherein said plurality of predefined model formats include a linear model format, a quadratic model format and a cubic model format.

38. A method of controlling a manufacturing process, comprising the steps of:

(a) obtaining metrology data from a first metrology tool with respect to a first processing tool;

(b) applying said metrology data to a first model structure relating said metrology data to a target set-point for a second processing tool and using a second model structure to control operation of the second processing tool in a feed-back control loop;

(c) modifying one or more variables in a variable parameter table for said second processing tool in response to application of the first model structure to the received metrology data;

(d) downloading said one or more modified variables to said second processing tool;

(e) operating said second processing tool in accordance with said downloaded variables.

39. The method of claim 38, wherein steps (a) through (e) are repeated to effectuate run-to-run control of the manufacturing process.

40. The method of claim 39, further comprising the steps of:

obtaining metrology data from a second metrology tool with respect to said second processing tool;

applying said metrology data from said second metrology tool to a second model structure relating said metrology data from said second metrology tool to said target set-point for said second processing tool; and

modifying one or more variables in said variable parameter table for said second processing tool.

41. The method of claim 40, wherein said step of applying said metrology data from said second metrology tool to said second model structure comprises the step of applying said metrology data from said second metrology tool to a feedback model.

42. The method of claim 40, further comprising the step of selecting, via a user interface, a model formats from among a plurality of model formats from which said second model structure is generated.

43. The method of claim 42, wherein said plurality of model formats comprise a linear model format, a quadratic model format and a cubic model format.

44. A supervisory station for managing run-to-run control of a manufacturing process, comprising:

an interface for receiving metrology data from a plurality of said metrology tools;

a control model relating said metrology data from one metrology tool to a target set-point for a next processing tool;

another control model used in controlling operation of said next processing tool in a feed-back control loop; and

a memory, said memory storing a plurality of output control variables for controlling the operation of a plurality of processing tools; and

at least one processor for executing one or more processes for controlling said metrology tools by adjusting said output control variables based upon the received metrology data and one or more control models, said control models selectable from a plurality of control model formats stored in said memory wherein one or more control variables for controlling the operation of the next processing tool are modified in response to application of the control model to the received metrology data.

45. The supervisory station of claim 44, wherein a control model is selected individually for each processing tool, each control model relating some or all of the received metrology data to output control variables for a particular processing tool according to the control model selected for the particular processing tool.

46. The supervisory station of claim 44, wherein said control model formats include a feedback control model.

47. The supervisory station of claim 44, wherein said control model formats include a combined feedback control model and a feed-forward control model.

48. The supervisory station of claim 44, further comprising a user interface for selecting the control model for each processing tool from the available control model formats.

49. The supervisory station of claim 44, wherein said control model formats include a linear model format, a quadratic model format and a cubic model format.

50. The supervisory station of claim 44, further comprising an interface for receiving one or more of the control model formats from an external plug-in unit.

51. The supervisory station of claim 44, wherein all or part of each variable parameter table is downloaded to its corresponding processing tool prior to operation of the processing tool, and wherein said variables modified in response to said received metrology data are automatically downloaded to said processing tools without user intervention being required.

52. The supervisory station of claim 45, further comprising plurality of variable parameter tables, one for each of said processing tools, stored in said memory, said variable parameter tables collectively associated with a manufacturing process recipe.

53. The supervisory station of claim 46, wherein said control model formats include a feed-forward control model.

54. The supervisory station of claim 47, further comprising a target set-point associated with said feedback control model, said supervisory station adjusting said target set-point based upon an output of said feed-forward control model.

55. The supervisory station of claim 52, wherein each control model relates some or all of the received metrology data to one or more variables of one of the variable parameter table for the control model'"'"'s corresponding processing tool, whereby said variables are modified in response to the received metrology data according to the control model.

56. The supervisory station of claim 55, wherein all or part of each variable parameter table is downloaded to its corresponding processing tool prior to operation of the processing tool, and wherein said variables modified in response to said received metrology data are automatically downloaded to said processing tools without user intervention being required.