System and method for controlling manufacturing apparatuses

• US 20050194590A1
• Filed: 03/02/2005
• Published: 09/08/2005
• Est. Priority Date: 03/03/2004
• Status: Active Grant

• Current Assignee: Kabushiki Kaisha Toshiba (Toshiba Corporation)
• Original Assignee: Kabushiki Kaisha Toshiba (Toshiba Corporation)
• Inventors: Noda, Tomonobu, Matsushita, Hiroshi, Sugamoto, Junji, Ushiku, Yukihiro, Kadota, Kenichi
• Granted Patent: US 7,221,991 B2
• Time in Patent Office: Days
• US Class Current: 257/48
• CPC Class Codes: G01R 31/2894 Aspects of quality control [QC] G01R31/31718 takes precedence; program control for QC G05B19/41875
• CPC Class Codes: G05B 2219/31357 Observer based fault detection, use model
• CPC Class Codes: G05B 23/0248 Causal models, e.g. fault tree; digraphs; qualitative physics
• CPC Class Codes: H01L 22/20 Sequence of activities consisting of a plurality of measurements, corrections, marking or sorting steps
• CPC Class Codes: Y02P 90/02 Total factory control, e.g. smart factories, flexible manufacturing systems [FMS] or integrated manufacturing systems [IMS]

Abstract

A control system for a manufacturing apparatus includes manufacturing information input unit acquiring time series data of apparatus parameters controlling manufacturing apparatuses; failure pattern classification module classifying in-plane distributions of failures of each of the wafers into failure patterns; an index calculation unit configured to statistically process the time series data by algorithms to calculate indices corresponding to the respective algorithms; an index analysis unit providing first and second frequency distributions of the indices categorized with and without the target failure pattern, to implement significance test between the first and second frequency distributions; and an abnormal parameter extraction unit extracting failure cause index of failure pattern by comparing value of the significance test with test reference value.

Claims

1. A control system for manufacturing apparatuses, comprising:

a manufacturing information input unit configured to acquire time series data of apparatus parameters controlling a plurality of manufacturing apparatuses, the manufacturing apparatuses performing a sequence of manufacturing processes on wafers so as to manufacture semiconductor devices;

a failure pattern classification unit configured to classify in-plane distributions of failures in each of the wafers into failure patterns;

an index calculation unit configured to statistically process the time series data by a plurality of algorithms so as to calculate a plurality of indices corresponding to the respective algorithms;

an index analysis unit configured to provide first and second frequency distributions of the indices so as to implement a significance test between the first and second frequency distributions, the first and second frequency distributions categorized with and without the target failure pattern respectively; and

an abnormal parameter extraction unit configured to extract a failure cause index of the target failure pattern by comparing a value of the significance test with a test reference value.

2. The control system of claim 1, further comprising a plurality of real time monitor units configured to monitor operation of the manufacturing apparatuses by referring to the failure cause index.

3. The control system of claim 1, further comprising an apparatus parameter collection server configured to collect the apparatus parameters from apparatus parameter acquisition units connected to the manufacturing apparatuses.

4. The control system of claim 1, further comprising a tester server configured to generate failure maps displaying the in-plane distributions based on failures measured by electrical tests of the semiconductor devices, which are performed after a sequence of wafer processes has been completed.

5. The control system of claim 4, wherein the indices include an autocorrelation coefficient of the time series data, a concordance rate of increase and decrease in variations of the time series data, a primary fitting coefficient that is a gradient of a variation of the time series data, a correlation coefficient indicating a difference between target time series data and last time series data, and another correlation coefficient indicating a difference between the target time series data and prior time series data.

6. The control system of claim 4, wherein the failure patterns are distributions of unevenly distributed clustering failures.

7. A control system for manufacturing apparatuses, comprising:

a manufacturing information input unit configured to acquire set values of apparatus parameters controlling a plurality of manufacturing apparatuses in a plurality of process modules;

an inspection information input unit configured to acquire wafer test information and defect information from an electrical test and a defect inspection of wafers processed in the process modules;

a yield calculation unit configured to calculate process yields of the process modules based on the wafer test information and the defect information; and

an optimum parameter extraction unit configured to statistically extract a target apparatus parameter having a significant difference for the process yield, so as to determine an optimum set value of the target apparatus parameter, based on the set values and the process yield, the optimum set value providing the maximum process yield.

8. The control system of claim 7, further comprising an apparatus parameter collection server configured to collect the apparatus parameters from apparatus parameter acquisition units connected to the manufacturing apparatuses.

9. The control system of claim 7, further comprising:

a tester server configured to acquire the wafer test information from a tester executing the electrical test;

a defect information server configured to acquire the defect information from a defect inspection apparatus executing the defect inspection.

10. A method for controlling manufacturing apparatuses, comprising:

acquiring time series data of apparatus parameters controlling a plurality of manufacturing apparatuses;

classifying in-plane distributions of failures of wafers into failure patterns, the wafers processed by the manufacturing apparatuses;

statistically processing the time series data by a plurality of algorithms so as to calculate a plurality of indices corresponding to the respective algorithms;

providing first and second frequency distributions of the indices so as to implement a significance test between the first and second frequency distributions, the first and second frequency distributions categorized with and without the target failure pattern respectively; and

identifying a failure cause index of the target failure pattern by comparing a value of the significance test with a test reference value.

11. The method of claim 10, further comprising monitoring operation of the manufacturing apparatuses by referring to the failure cause index.

12. The method of claim 10, wherein the indices are calculated from waveforms of the time series data.

13. The method of claim 10, wherein the indices include an autocorrelation coefficient of the time series data, a concordance rate of increase and decrease in variations of the time series data, a primary fitting coefficient that is a gradient of a variation of the time series data, a correlation coefficient indicating a difference between target time series data and last time series data, and another correlation coefficient indicating a difference between the target time series data and prior time series data.

14. The method of claim 10, wherein the failure patterns are distributions of unevenly distributed clustering failures.

15. A method for controlling manufacturing apparatuses, comprising:

acquiring set values of apparatus parameters controlling a plurality of manufacturing apparatuses in a plurality of process modules;

acquiring wafer rest information and defect information measured by an electrical test and a defect inspection of wafers processed by the process modules;

calculating process yields of the process modules based on the wafer test information and the defect information; and

statistically extracting a target apparatus parameter having a significant difference for the process yield, so as to determine an optimum set value of the target apparatus parameter, based on the set values and the process yield, the optimum set value providing the maximum process yield.

16. The method of claim 15, wherein the target apparatus parameter is extracted by one of data mining and multivariate analysis.