System and method for finding defective tools in a semiconductor fabrication facility

• US 6,701,204 B1
• Filed: 05/22/2000
• Issued: 03/02/2004
• Est. Priority Date: 05/22/2000
• Status: Expired due to Fees

• Current Assignee: Qimonda AG (Infineon Technologies AG), Mosel Vitelic Incorporated, ProMOS Technologies, Inc.
• Original Assignee: Infineon Technologies AG, Mosel Vitelic Incorporated, ProMOS Technologies, Inc.
• Inventors: Nicholson, Mark
• Primary Examiner(s): Von Buhr, Maria N.
• Application Number:
  US09/577,695
• Time in Patent Office: 1,380 Days
• Field of Search: 700/95, 700/108-111, 700/117, 700/119-121, 700/174-177, 716/2-4, 438/10, 438/11, 438/14-18, 702/35, 702/182-185
• US Class Current: 700/121
• CPC Class Codes: H01L 22/20 Sequence of activities consisting of a plurality of measurements, corrections, marking or sorting steps

Abstract

A system and method for finding a defective tool in a semiconductor fabrication facility is disclosed. When the tools process the wafers, data representing the time period during which each wafer passes through each tool is sent to a database. The wafers are tested for defects, and lots having wafers with common failure signatures are determined. A lot list for each tool is generated, a positive weight value is assigned to each bad lot, and a negative weight value is assigned to each good lot. A cumulative value is calculated for each tool by sequentially adding the weight values of each lot in the lot list and keeping the cumulative value above or equal to zero. The tool with the largest maximum cumulative value is the tool that is most likely to be defective.

Claims

1. A method for finding a defective tool in a semiconductor fabrication facility having a plurality of tools for processing a plurality of semiconductor wafers, the method comprising the steps of:

performing a series of wafer processing steps on the plurality of wafers with the plurality of tools;

generating a database having data representative of the time during which each wafer passed through each of said plurality of tools;

testing the plurality of wafers to determine whether a wafer is associated with a failure signature;

generating a lot list for each tool, the sequential order of the lots in the lot list representing the process sequence of the lots;

assigning a weight value to each lot in the lot list, said weight value being a predetermined positive value for lots having wafers with said failure signature, and said weight value being negative for lots having wafers without said failure signature;

generating a cumulative value for each tool by sequentially summing the weight values of each lot in the corresponding lot list, and assigning a peak cumulative value for each tool as the maximum cumulative value during the summation process; and

assigning the tool with the greatest peak cumulative value as the tool most likely to cause said fail signature.

2. The method of claim 1, wherein said the absolute value of said positive weight value is larger than the absolute value of said negative weight value.

3. The method of claim 1, wherein said step of testing the plurality of wafers includes testing performed by an in-line parameter tester before said series of wafer processing steps are completed.

4. The method of claim 1, wherein said step of testing the plurality of wafers includes testing performed by a defect tester after said series of wafer processing steps are completed.

5. The method of claim 1, wherein circuitry is gradually formed on the plurality of wafers during said series of wafer processing steps, and said step of testing the plurality of wafers includes testing the electrical parameters of the circuitry gradually formed on the plurality of wafers.

6. The method of claim 1, wherein said step of testing the plurality of wafers measures the physical properties of the plurality of wafers.

7. The method of claim 5, wherein a wafer is associated with said failure signature when the electrical parameters of a wafer resembles a predetermined failure characteristic pattern.

8. The method of claim 1, wherein after the step of assigning the tool most likely to cause said fail signature further includes a step of displaying a graphical representation of the cumulative values of each tool.

9. The method of claim 1, wherein the step of generating a cumulative value sequentially sums the weight value of each lot in the corresponding lot list, keeping the cumulative value not less than zero during the summation process.

10. A defective tool detection system used in connection with a fabrication facility having a plurality of tools for processing semiconductor wafers that are grouped into a plurality of wafer lots, comprising:

an electrical parameter tester for receiving the wafer lots and identifying defects on each wafer of the wafer lots;

a failure signature analyzer for identifying a failure signature representative of defect patterns on the wafers having defects;

a memory for storing a set of data representative of the time during which each wafer passed through each tool in the fabrication facility; and

an equipment commonality analyzer for receiving said set of data stored in said memory and said failure signature from said fail signature analyzer in order to generate a peak cumulative value for each tool, wherein a relatively higher peak cumulative value indicates a relatively higher likelihood that the tool is the cause of the failure signature.

11. The defective tool detection system of claim 10, wherein said equipment commonality analyzer performs the following steps:

generating a lot list for each tool in the fabrication facility representative of the sequence in which wafer lots are processed by said each tool;

assigning a weight value to each wafer lot in the lot list;

summing said weight values of the wafer lots in each lot list to obtain a cumulative value, keeping said cumulative value not less than zero during the summation process;

finding said peak cumulative value for each tool that represents the highest value accumulated during the step of summing said weight values; and

assigning the tool having the maximum peak cumulative value as the tool most likely to have caused the failure signature.

12. The defective tool detection system of claim 11, wherein wafer lots having wafers with said failure signature are assigned a positive weight value, and wafer lots not having wafers with said failure signature are assigned a negative weight value.

13. The defective tool detection system of claim 12, wherein the absolute value of said positive weight value is larger than the absolute value of said negative weight value.

14. The defective tool detection system in claim 11, further comprising an output display for displaying a graphic plot of the cumulative values versus the lot number for each tool.

15. A method for processing a plurality of semiconductor wafers using commonality analysis to effectively reduce defects on the wafers, the method comprising the steps of:

providing a plurality of tools for processing the plurality of wafers;

processing the plurality of wafers using said plurality of tools and storing time data in a database representing the time period for which each tool processed each wafer;

testing each of the plurality of wafers for a failure signature;

generating a wafer list for each tool representing the process sequence of all wafers that each tool has processed;

assigning a weight value to each wafer in the wafer list;

generating a peak cumulative value for each tool, said peak cumulative value derived from a summation of the weight values; and

assigning the tool with the largest peak cumulative value as the tool most likely to have caused said failure signature.

16. The method of claim 15, wherein the step of assigning a weight value to each wafer assigns a positive weight value to wafers having said failure signature, and assigns a negative weight value to wafer not having said failure signature.

17. The method of claim 16, wherein the absolute value of said positive weight value is larger than the absolute value of said negative weight value.

18. The method of claim 15, wherein the step of generating a peak cumulative value for each tool further comprises the steps of:

summing the weight value of each wafer in the wafer list sequentially; and

assigning the highest value accumulated during the summation process as the peak cumulative value.

19. The method of claim 15, wherein said failure signature represents a defect pattern caused by mechanical scratches.

20. The method of claim 15, wherein said failure signature represents a center spot defect pattern in which most of the defects are clustered near the center region of the wafer.

21. A method for finding a defective tool among a plurality of tools in a semiconductor fab after a plurality of wafer lots have been processed by the plurality of tools, the method comprising the steps of:

testing each wafer in the plurality of wafer lots for a failure signature; and

applying a tool-based commonality analysis to assign peak cumulative values to each of the plurality of tools; and

generating a tool list to rank the likelihood of a tool being the cause of the failure signature according to the corresponding peak cumulative values.

22. The method of claim 21, wherein said tool-based commonality analysis performs the steps of:

generating a lot list for each of the plurality of tools representing the process sequence of the wafer lots that each of the plurality of tools have processed;

assigning a weight value to each wafer lot in the lot list;

summing the weight value of each wafer lot in the lot list sequentially to generate a cumulative value, and always keeping the cumulative value not less than zero; and

assigning the highest cumulative value during the summation process as the peak cumulative value.

23. The method of claim 21, wherein the step of assigning a weight value to each lot assigns a positive weight value to wafer lots having wafers with said failure signature, assigns a negative weight value to wafer lots not having any wafer with said failure signature, and the absolute value of said positive weight value is greater than the absolute value of said negative weight value.