Determining appropriateness of sampling integrated circuit test data in the presence of manufacturing variations

• US 9,287,185 B1
• Filed: 06/29/2015
• Issued: 03/15/2016
• Est. Priority Date: 06/29/2015
• Status: Expired due to Fees

• Current Assignee: GlobalFoundries, Inc.
• Original Assignee: GlobalFoundries, Inc.
• Inventors: Bonilla, Griselda, Li, Baozhen, Linder, Barry P., Stathis, James H., Wu, Ernest Y., Zhao, Kai
• Primary Examiner(s): Menz, Laura
• Application Number:
  US14/753,771
• Time in Patent Office: 260 Days
• US Class Current: 1/1
• CPC Class Codes: G01R 31/2894 Aspects of quality control [QC] G01R31/31718 takes precedence; program control for QC G05B19/41875
• CPC Class Codes: H01L 22/14 for electrical parameters, e.g. resistance, deep-levels, CV, diffusions by electrical means
• CPC Class Codes: H01L 22/20 Sequence of activities consisting of a plurality of measurements, corrections, marking or sorting steps

Abstract

Methods and systems determine an original statistical variance of an original failure distribution of a component (that is common to all chips tested) that occurs during manufacturing of wafers containing such chips. These methods and systems determine a first statistical variance of a reconstructed failure distribution, relative to sample size; and determine a second statistical variance of a mean time to failure of the component, relative to sample size. The first and second statistical variances are combined into a total reconstruction variance. Methods and systems determine whether the original statistical variance is less than the total reconstruction variance to identify whether the process of creating the reconstructed failure distribution can be used. Therefore, these methods and systems prohibit testing of the additional wafers manufactured using the specific wafer design and manufacturing process when on the original statistical variance is less than the total reconstruction variance.

Claims

1. A method comprising:

testing a wafer comprising integrated circuit chips to produce test data;

determining an original statistical variance of an original failure distribution using the test data;

performing a process of creating a reconstructed failure distribution using sampling sizes of the test data that are less than all of the test data;

determining a first statistical variance of the reconstructed failure distribution relative to sample size;

determining a second statistical variance of a mean time to failure relative to sample size using the test data;

combining the first statistical variance and the second statistical variance into a total reconstruction variance;

determining whether the original statistical variance is less than the total reconstruction variance to identify whether additional wafers can be subjected to sample-based testing using the process of creating the reconstructed failure distribution with the sampling sizes;

prohibiting the sample-based testing of the additional wafers based on the original statistical variance being less than the total reconstruction variance; and

permitting the sample-based testing of the additional wafers based on the original statistical variance not being less than the total reconstruction variance.

2. Method according to claim 1, each of the integrated circuit chips containing identically designed devices, the method further comprising:

selecting a number of integrated circuit chips per wafer to sample in the sampling sizes that are less than all of the integrated circuit chips; and

selecting a number of devices per integrated circuit chip to sample in the sampling sizes that are less than all of the identically designed devices in the integrated circuit chips.

3. Method according to claim 1, all of the test data being a multiple of the sampling sizes.

4. Method according to claim 1, further comprising altering the sample sizes to include more samples based on identifying that the process of creating the reconstructed failure distribution cannot be used.

5. Method according to claim 1, further comprising determining the original failure distribution based on all of the test data of the test data.

6. Method according to claim 1, each of the integrated circuit chips containing identically designed devices, all of the test data and the sample sizes comprising the test data from a limited number of integrated circuit chips per wafer that are less than all of the integrated circuit chips on the wafer and a limited number of devices per integrated circuit chip that are less than all components on the integrated circuit chips.

7. Method according to claim 1, the original statistical variance of the original failure distribution comprising at least one of:

a chip-to-chip variance; and

a device-to-device variance.

8. A method comprising:

manufacturing a wafer comprising integrated circuit chips using a wafer design and a manufacturing process that uses wafer manufacturing machines;

testing the wafer to produce test data;

determining an original statistical variance of an original failure distribution of a component common to all of the integrated circuit chips using the test data;

performing a process of creating a reconstructed failure distribution of the component using sampling sizes of the test data that are less than all of the test data;

determining a first statistical variance of the reconstructed failure distribution relative to sample size;

determining a second statistical variance of a mean time to failure of the component relative to sample size using all of the test data of the test data;

combining the first statistical variance and the second statistical variance into a total reconstruction variance;

determining whether the original statistical variance is less than the total reconstruction variance to identify whether additional wafers manufactured according to the wafer design and the manufacturing process can be subjected to sample-based testing using the process of creating the reconstructed failure distribution with the sampling sizes;

prohibiting the sample-based testing of the additional wafers based on the original statistical variance being less than the total reconstruction variance; and

permitting the sample-based testing of the additional wafers based on the original statistical variance not being less than the total reconstruction variance.

9. Method according to claim 8, each of the integrated circuit chips containing identically designed devices, the method further comprising:

selecting a number of integrated circuit chips per wafer to sample in the sampling sizes that are less than all of the integrated circuit chips in the wafer design; and

selecting a number of devices per integrated circuit chip to sample in the sampling sizes that are less than all of the identically designed devices in the integrated circuit chips in the wafer design.

10. Method according to claim 8, all of the test data being a multiple of the sampling sizes.

11. Method according to claim 8, further comprising altering the sample sizes to include more samples based on identifying that the process of creating the reconstructed failure distribution of the component cannot be used.

12. Method according to claim 8, further comprising determining the original failure distribution of the component of the wafer based on all of the test data of the test data.

13. Method according to claim 8, each of the integrated circuit chips containing identically designed devices, all of the test data and the sample sizes comprising the test data from a limited number of integrated circuit chips per wafer that are less than all of the integrated circuit chips on the wafer and a limited number of devices per integrated circuit chip that are less than all of the components on the integrated circuit chips.

14. Method according to claim 8, the original statistical variance of the original failure distribution comprising at least one of:

a chip-to-chip variance; and

a device-to-device variance.

15. A computer program product for testing wafers, the computer program product comprising a computer readable storage medium having program instructions embodied therewith, the program instructions executable by a computer to cause the computer to perform a method comprising:

testing a wafer comprising integrated circuit chips to produce test data;

determining an original statistical variance of an original failure distribution using the test data;

performing a process of creating a reconstructed failure distribution using sampling sizes of the test data that are less than all of the test data;

determining a first statistical variance of the reconstructed failure distribution relative to sample size;

determining a second statistical variance of a mean time to failure relative to sample size using the test data;

combining the first statistical variance and the second statistical variance into a total reconstruction variance;

determining whether the original statistical variance is less than the total reconstruction variance to identify whether additional wafers can be subjected to sample-based testing using the process of creating the reconstructed failure distribution with the sampling sizes;

prohibiting the sample-based testing of the additional wafers based on the original statistical variance being less than the total reconstruction variance; and

permitting the sample-based testing of the additional wafers based on the original statistical variance not being less than the total reconstruction variance.

16. The computer program product according to claim 15, each of the integrated circuit chips containing identically designed devices, the method further comprising:

selecting a number of integrated circuit chips per wafer to sample in the sampling sizes that are less than all of the integrated circuit chips; and

selecting a number of devices per integrated circuit chip to sample in the sampling sizes that are less than all of the identically designed devices in the integrated circuit chips.

17. The computer program product according to claim 15, all of the test data being a multiple of the sampling sizes.

18. The computer program product according to claim 15, further comprising altering the sample sizes to include more samples based on identifying that the process of creating the reconstructed failure distribution cannot be used.

19. The computer program product according to claim 15, further comprising determining the original failure distribution based on all of the test data of the test data.

20. The computer program product according to claim 15, each of the integrated circuit chips containing identically designed devices, all of the test data and the sample sizes comprising the test data from a limited number of integrated circuit chips per wafer that are less than all of the integrated circuit chips on the wafer and a limited number of devices per integrated circuit chip that are less than all components on the integrated circuit chips.