Method for controlling semiconductor wafer processing

• US 5,661,669 A
• Filed: 06/07/1995
• Issued: 08/26/1997
• Est. Priority Date: 12/17/1993
• Status: Expired due to Term

• Current Assignee: Texas Instruments, Inc.
• Original Assignee: Texas Instruments, Inc.
• Inventors: Barna, Gabe G., Mozumder, Purnendu Kanti
• Primary Examiner(s): Ramirez, Ellis B.
• Assistant Examiner(s): PIPALA, EDWARD J
• Application Number:
  US08/482,599
• Time in Patent Office: 811 Days
• Field of Search: 364/468, 364/550, 364/552, 364/488, 364/489, 364/491, 364/148-151, 364/578, 364/554, 364/553, 364/556, 364/557, 364/558, 364/468.28, 204/192.13, 427/8, 427/10
• US Class Current: 702/84
• CPC Class Codes: G05B 13/042 in which a parameter or coefficient is automatically adjusted to optimise the performance
• CPC Class Codes: G05B 17/02 electric

Abstract

A system (10) for run-to-run control of semiconductor wafer processing is provided. An input/output device (12) receives a desired quality characteristic for a particular semiconductor fabrication process. A generating circuit (22) uses a model to generate appropriate process parameters for a processing unit (20) and an expected quality characteristic. An adjusting circuit (16) functions to adjust process parameter inputs of the processing unit (20). In-situ sensor (18) functions to measure a quality characteristic of the process in the processing unit (20) on a real-time basis. A comparing circuit (24) functions to compare the measured quality characteristic with the expected quality characteristic. A model adjusting circuit (26) may adjust the model of the generating circuit (22) if the measured quality characteristic varies from the expected quality characteristic by more than a predetermined statistical amount.

Claims

1. A method for run-to-run control of a processing unit, comprising:

generating process parameters using a model which provides a transfer function from said process parameters to an expected quality characteristic provided by a user;

adjusting selected inputs of the processing unit to process a device under manufacture using said generated process parameters;

measuring a quality characteristic of said process in real-time during said process using an in-situ sensor comprising an optical device operable on a single wavelength to generate a signal proportional to the concentration of a predetermined chemical species present during processing of the device;

comparing said measured quality characteristic with said quality characteristic; and

adjusting said model if said measured quality characteristic varies from said expected quality characteristic by more than a predetermined statistic amount.

2. The method of claim 1 wherein said processing unit is a semiconductor processing unit and said device under manufacture is a semiconductor wafer.

3. The system of claim 2, wherein said adjusting step comprises operating a plurality of actuating devices operable to adjust selected process parameter inputs of the processing unit.

4. The system of claim 2, wherein said comparing step subtracts said expected quality characteristic from said measured quality characteristic to generate a difference value and averages said difference value with between two to eight difference values from previous devices.

5. The system of claim 2, wherein said sensor comprises a monochromator and further comprising the steps of:

storing a value of said signal generated by said monochromator at predetermined intervals during the processing to generate data representative of an optical emission trace; and

converting a slope at an end-point of said optical emission trace to a quality characteristic for the process.

6. The system of claim 2, wherein said adjusting step comprises adjusting a constant term of said model if said measured quality characteristic varies from said expected quality characteristic by more than a predetermined statistical amount.

7. The system of claim 2, wherein said adjusting step weights a difference between said measured quality characteristic and said expected quality characteristic for the current device with said difference for up to twenty-five previous devices to adjust a constant term of said model.

8. A method for run-to-run control of a processing unit, comprising:

generating process parameters using a plurality of models each of which provides a transfer function from said process parameters to at least one of a plurality of quality characteristics at least one of said characteristics being unmeasurable after said process is complete;

adjusting selected inputs of the processing unit to process a device under manufacture using said generated process parameters;

measuring a quality characteristic of said process in real-time during said process with an in-situ sensor;

comparing the measured quality characteristics with expected quality characteristics; and

adjusting the model if the measured quality characteristics vary from the expected quality characteristics by more than a predetermined statistical amount.

9. The method of claim 8 wherein said processing unit is a semiconductor processing unit and said device under manufacture is a semiconductor wafer.

10. The system of claim 9, wherein said adjusting step comprises operating a plurality of actuating devices operable to adjust selected process parameter inputs of the processing unit.

11. The system of claim 9, wherein said comparing step subtracts said expected quality characteristic from said measured quality characteristic to generate a difference value and averages said difference value with between two to eight difference values from previous devices.

12. The system of claim 9, wherein said sensor comprises a monochromator and further comprising the steps of:

storing a value of said signal generated by said monochromator at predetermined intervals during the processing to generate data representative of an optical emission trace; and

converting a slope at an end-point of said optical emission trace to a quality characteristic for the process.

13. The system of claim 9, wherein said adjusting step comprises adjusting a constant term of said model if said measured quality characteristic varies from said expected quality characteristic by more than a predetermined statistical amount.

14. The system of claim 9, wherein said adjusting step weights a difference between said measured quality characteristic and said expected quality characteristic for the current device with said difference for up to twenty-five previous devices to adjust a constant term of said model.

15. A method for run-to-run control of a processing unit, comprising:

generating process parameters and a plurality of expected quality characteristics using a plurality of models and desired quality characteristics;

adjusting selected inputs of the processing unit to process the device under manufacture using said generated process parameters;

measuring a quality characteristic of said process utilizing a real-time, in-situ sensor comprising a monochromator operable to generate a signal proportional to the concentration of a predetermined chemical species present during processing of the device;

storing a value of a signal generated by said monochromator at predetermined intervals during the processing to generate data representative of an optical emission trace; and

converting a slope at an end-point of said optical emission trace to a quality characteristic for the process;

comparing the measured quality characteristics with the expected quality characteristics;

adjusting the model if the measured quality characteristics varies from the expected quality characteristics by more than a predetermined statistical amount.

16. The method of claim 15 wherein said processing unit is a semiconductor processing unit and said device under manufacture is a semiconductor wafer.

17. The system of claim 16, wherein said adjusting step comprises operating a plurality of actuating devices operable to adjust selected process parameter inputs of the processing unit.

18. The system of claim 16, wherein said comparing step subtracts an expected quality characteristic from said measured quality characteristic to generate a difference value and averages said difference value with between two to eight difference value from previous devices.

19. The system of claim 16, wherein said adjusting step comprises adjusting a constant term of said model if said measured quality characteristic varies from said expected quality characteristic by more than a predetermined statistical amount.

20. The system of claim 16, wherein said adjusting step weights a difference between said measured quality characteristic and said expected quality characteristic for the current device with said difference for up to twenty-five previous devices to adjust a constant term of said model.

21. A method for run-to-run control of a processing unit, comprising:

generating process parameters and an expected quality characteristic using a model and a desired quality characteristic provided by a user;

adjusting selected inputs of the processing unit to process a device under manufacture using said generated process parameters;

measuring a signal correlated to a quality characteristic of said process which is not measurable after the process is complete utilizing a real-time, in-situ sensor;

comparing said measured quality characteristic with said expected quality characteristic;

adjusting said model if said measured quality characteristic varies from said expected quality characteristic by more than a predetermined statistical amount.

22. The method of claim 21 wherein said processing unit is a semiconductor processing unit and said device under manufacture is a semiconductor wafer.

23. The system of claim 22, wherein said adjusting step comprises operating a plurality of actuating devices operable to adjust selected process parameter inputs of the processing unit.

24. The system of claim 22, wherein said comparing step subtracts said expected quality characteristic from said measured quality characteristic to generate a difference value and averages said difference value with between two to eight difference values from previous devices.

25. The system of claim 22, wherein said sensor comprises a monochromator and further comprising the steps of:

storing a value of said signal generated by said monochromator at predetermined intervals during the processing to generate data representative of an optical emission trace; and

converting a slope at an end-point of said optical emission trace to a quality characteristic for the process.

26. The system of claim 22, wherein said adjusting step comprises adjusting a constant term of said model if said measured quality characteristic varies from said expected quality characteristic by more than a predetermined statistical amount.

27. The system of claim 22, wherein said adjusting step weights a difference between said measured quality characteristic and said expected quality characteristic for the current device with said difference for up to twenty-five previous devices to adjust a constant term of said model.

28. A method for run-to-run control of a processing unit, comprising:

generating process parameters and an expected quality characteristic using a model and a desired quality characteristic provided by a user;

adjusting selected inputs of the processing unit to process a device under manufacture using said generated process parameters;

measuring a quality characteristic of said process utilizing a real-time, in-situ monochromator for measuring a signal correlated to a non-directly observable quality characteristic of said process and generating a signal proportional to the concentration of a predetermined chemical species present during processing of the device;

storing a value of signal at predetermined intervals during the processing to generate data representative of an optical emission trace; and

converting a slope at an end-point of said optical emission trace to a quality characteristic for the process;

comparing said measured quality characteristic with said expected quality characteristic;

adjusting said model if said measured quality characteristic varies from said expected quality characteristic by more than a predetermined statistical amount.

29. The method of claim 28 wherein said processing unit is a semiconductor processing unit and said device under manufacture is a semiconductor wafer.

30. The system of claim 29, wherein said adjusting step comprises operating a plurality of actuating devices operable to adjust selected process parameter inputs of the processing unit.

31. The system of claim 29, wherein said comparing step subtracts said expected quality characteristic from said measured quality characteristic to generate a difference value and averages said difference value with between two to eight difference values from previous devices.

32. The system of claim 29 wherein said adjusting step comprises adjusting a constant term of said model if said measured quality characteristic varies from said expected quality characteristic by more than a predetermined statistical amount.

33. The system of claim 29, wherein said adjusting step weights a difference between said measured quality characteristic and said expected quality characteristic for the current device with said difference for up to twenty-five previous devices to adjust a constant term of said model.