Method of monitoring a production process using a linear combination of measured variables with selected weights

• US 7,479,395 B2
• Filed: 07/25/2002
• Issued: 01/20/2009
• Est. Priority Date: 01/25/2000
• Status: Expired due to Fees

• Current Assignee: Infineon Technologies AG
• Original Assignee: Infineon Technologies AG
• Inventors: Voigtländer, Knut, Bell, Ferdinand, Knobloch, Dirk, Zimpel, Jan
• Primary Examiner(s): CHEN, KIN CHAN
• Application Number:
  US10/205,080
• Publication Number:
  US 20030008507A1
• Time in Patent Office: 2,371 Days
• Field of Search: 438/9, 438/7, 438/706, 438/710, 438/712, 216/60
• US Class Current: 438/9
• CPC Class Codes: G01N 21/25 Colour; Spectral properties, i.e. comparison of effect of material on the light at two or more different wavelengths or wavelength bands
• CPC Class Codes: H01L 22/12 for structural parameters, e.g. thickness, line width, refractive index, temperature, warp, bond strength, defects, optical inspection, electrical measurement of structural dimensions, metallurgic measurement of diffusions electrical measurement of diffusions H01L22/14
• 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

The invention relates to a method for monitoring a manufacturing process, which by using a linear combination of measured variables with judiciously chosen weighting, produces a suitable (optimal) signal for determining each parameter. The extraction of the parameters from the measured variables is thus greatly simplified and in many cases becomes actually possible for the first time. According to the invention, large amounts of data may now be prepared, such that the crucial information (parameters) can be obtained from the data, almost in real time, or in real time.

Claims

1. A method of monitoring a production process, which comprises:

determining intensities of various wavelengths in an emission spectrum at predefined times;

forming at least one linear combination from the intensities of various wavelengths, and selecting weights of the linear combination such that a predefined curve shape is approximated; and

determining at least one characteristic variable from the linear combination.

2. The method according to claim 1, wherein:

the intensities of the various wavelengths are in a spectral range between 200 nm and 950 nm.

3. The method according to claim 1, wherein:

the emission spectrum is an emission spectrum of a plasma process.

4. The method according to claim 3, wherein:

the characteristic variable is an end point of the plasma process.

5. The method according to claim 1, which comprises:

using a step-like function as a predefined curve shape.

6. The method according to claim 5, wherein:

the step-like function has at least one free parameter to match to the characteristic variable that will be determined.

7. The method according to claim 1, which comprises:

determining the weights of the linear combination by configuring the intensities of various wavelengths in a matrix and performing a principal component analysis based on the matrix.

8. The method according to claim 7, which comprises:

using a singular value decomposition to perform the principal component analysis.

9. The method according to claim 1, which comprises:

determining the weights of the linear combination by minimizing a function based on the predefined curve shape and the linear combination.

10. The method according to claim 9, which comprises:

as a result of the minimizing of the function, determining important principal components of a measured variable matrix to approximate the predefined curve shape.

11. The method according to claim 9, which comprises:

as a result of the minimizing of the function, determining a free parameter of a step-like function.

12. The method according to claim 1, which comprises:

determining the weights of the linear combination by maximizing a function based on the predefined curve shape and the linear combination.

13. The method according to claim 1, which comprises:

performing the determining of the characteristic variable by smoothing the linear combination.

14. The method according to claim 13, which comprises:

using a zero-phase filter to perform the smoothing of the linear combination.

15. The method according to claim 1, which comprises:

performing the determining of the characteristic variable by using qualitative criteria.

16. The method according to claim 1, which comprises:

obtaining the intensities of various wavelengths from a plurality of runs selected from a group consisting of trial runs and production runs; and

determining the weights of the linear combination by using the intensities of various wavelengths.

17. The method as claimed in claim 16, which comprises:

determining the weights of the linear combination by configuring intensities of various wavelengths from a plurality of runs, selected from a group consisting of trial runs and production runs, in a matrix and performing a principal component analysis based on the matrix.

18. The method according to claim 1, wherein:

the emission spectrum of a plasma etching process is used.

19. The method according to claim 1, wherein:

the various wavelengths include a large number of wavelengths.

20. A method of monitoring a production process, which comprises:

determining intensities of various wavelengths in an emission spectrum of a plasma process at predefined times;

forming at least one linear combination from the intensities of various wavelengths, and selecting weights of the linear combination such that a predefined curve shape is approximated; and

determining an end point of the plasma process from the linear combination.