Real time control of plasma etch utilizing multivariate statistical analysis

• US 5,479,340 A
• Filed: 09/20/1993
• Issued: 12/26/1995
• Est. Priority Date: 09/20/1993
• Status: Expired due to Fees

• Current Assignee: Intel Corporation, Sematech, Inc.
• Original Assignee: Intel Corporation, Sematech, Inc.
• Inventors: Fox, Edward P., Kappuswamy, Chandru
• Primary Examiner(s): Voeltz, Emanuel T.
• Assistant Examiner(s): MILLER, CRAIG S
• Application Number:
  US08/124,146
• Time in Patent Office: 827 Days
• Field of Search: 364/152, 364/153, 364/148, 364/130, 364/164, 364/165, 364/468, 364/477, 364/554, 364/552, 364/550, 437/7, 437/8, 437/225, 437/228, 437/939
• US Class Current: 700/33
• CPC Class Codes: B24B 37/013 Devices or means for detecting lapping completion
• CPC Class Codes: G05B 21/02 electric
• CPC Class Codes: H01J 2237/334 Etching
• CPC Class Codes: H01J 37/32018 Glow discharge
• CPC Class Codes: H01J 37/32082 Radio frequency generated discharge H01J37/32357, H01J37/32366, H01J37/32394 and H01J37/32403 take precedence
• 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
• CPC Class Codes: H01L 22/26 Acting in response to an ongoing measurement without interruption of processing, e.g. endpoint detection, in-situ thickness measurement endpoint detection arrangements in CMP apparatus B24B37/013, in discharge apparatus H01J37/32

Abstract

Hotelling'"'"'s T² statistical analysis and control is used to provide multivariate analysis of components of an RF spectra for real time, in-situ control of an ongoing semiconductor process. An algorithm calculates the T² value which is then used to generate a feedback signal, if the T² value is out of range, to indicate an out-of-tolerance condition.

Claims

1. A method for providing real time control of a semiconductor manufacturing process within a reactor by utilizing a multivariate statistical analysis to provide process control, comprising the steps of:

a) determining reactor operating parameters which are associated with characterizing an on-going performance of said semiconductor manufacturing process, said parameters having correlated characteristics such that a change in one of said parameters causes a change in at least another one of said parameters, wherein said performance of said semiconductor manufacturing process is also changed;

b) measuring said parameters by utilizing a monitoring device coupled to said reactor when said reactor is in operation to perform said semiconductor manufacturing process;

c) providing measured reactor operating parameters as input variables to be operated on by performing said multivariate statistical analysis;

d) performing said multivariate statistical analysis by use of a processor while said reactor is performing said semiconductor manufacturing process to obtain a probability value which corresponds to said performance of said semiconductor manufacturing process, wherein if said probability value is at a predesignated mean value, it corresponds to a desired performance of said semiconductor manufacturing process, but a variance from said mean value corresponds to a variation from said desired performance of said semiconductor manufacturing process, said multivariate statistical analysis evaluating overall performance of said semiconductor manufacturing process based on all of said parameters and not basing performance separately on each individual parameter;

e) determining if said probability value resides within a predesignated acceptable range from said mean value for said semiconductor manufacturing process, wherein said probability value beyond said acceptable variation range indicates an unacceptable operating condition for said semiconductor manufacturing process;

f) utilizing said probability value to monitor operation of said reactor while said semiconductor manufacturing process is on-going in said reactor;

g) performing steps b) through f) until said semiconductor manufacturing process is completed or until said unacceptable operating condition is detected;

h) correcting said unacceptable operating condition once detected by adjusting said reactor parameters until said probability value is back within said acceptable variation range;

wherein real time in-situ control of said semiconductor manufacturing process is achieved by utilizing said multivariate statistical analysis.

2. The method of claim 1 wherein said multivariate statistical analysis is determined by utilizing Hotelling'"'"'s T² distribution to calculate said probability value.

3. The method of claim 2 wherein said semiconductor manufacturing process is terminated in step h) when said unacceptable operating condition is noted in step g).

4. The method of claim 1 wherein said semiconductor manufacturing process is terminated in step h) when said unacceptable operating condition is noted in step g).

5. A method for providing real time control of a plasma etch process within a plasma reactor by utilizing a multivariate statistical analysis to provide process control, comprising the steps of:

a) determining components of a radio frequency (RF) spectra which are associated with characterizing an on-going performance of said plasma etch process, said RF components having correlated characteristics such that a change in one of said RF components causes a change in at least another one of said RF components, wherein said performance of said plasma etch process is also changed;

b) measuring said RF components by utilizing a monitoring device coupled to said reactor when said reactor is in operation to perform said plasma etch process;

c) providing measured RF components as input variables to be operated on by performing said multivariate statistical analysis;

d) performing said multivariate statistical analysis by use of a processor while said reactor is performing said plasma etch process to obtain a probability value which corresponds to said performance of said plasma etch process, wherein if said probability value is at a predesignated mean value, it corresponds to a desired performance of said plasma etch process, but a variance from said mean value corresponds to a variation from said desired performance of said plasma etch process, said multivariate statistical analysis evaluating overall performance of said plasma etch process based on all of said components and not basing performance separately on each individual component;

e) determining if said probability value resides within a predesignated acceptable range from said mean value for said plasma etch process, wherein said probability value beyond said acceptable variation range indicates an unacceptable operating condition for said plasma etch process;

f) utilizing said probability value to monitor operation of said reactor while said plasma etch process is on-going in said reactor;

g) performing steps b) through f) until said plasma etch process is completed or until said unacceptable operating condition is detected;

h) correcting said unacceptable operating condition once detected by adjusting said RF components until said probability value is back within said acceptable variation range;

wherein real time in-situ control of said plasma etch process is achieved by utilizing said multivariate statistical analysis.

6. The method of claim 5 wherein said multivariate statistical analysis is determined by utilizing Hotelling'"'"'s T² distribution to calculate said probability value.

7. The method of claim 6 wherein said RF components are derived from filtering a RF signal used to generate plasma in said plasma reactor.

8. The method of claim 7 further including the step of storing values of said measured RF components in a memory prior to performing said multivariate statistical analysis.

9. The method of claim 8 wherein said RF components are comprised of fundamental RF voltage and current, harmonic values of RF voltage and current, and a phase difference between said fundamental voltage and current.

10. The method of claim 9 wherein said plasma etch process is terminated in step h) when said unacceptable operating condition is noted in step g).