Apparatus and method for automatically detecting defects on silicon dies on silicon wafers

• US 6,504,948 B1
• Filed: 04/15/1999
• Issued: 01/07/2003
• Est. Priority Date: 09/04/1997
• Status: Expired due to Fees

• Current Assignee: Texas Instruments, Inc.
• Original Assignee: Texas Instruments, Inc.
• Inventors: Thorne, Richard, Schemmel, Floyd
• Primary Examiner(s): WERNER, BRIAN P
• Application Number:
  US09/292,263
• Time in Patent Office: 1,363 Days
• Field of Search: 382/145, 382/148, 382/149, 382/151, 382/217-219, 382/228, 382/274, 382/275, 348/87, 348/120, 356/237.4, 356/237.5, 438/16, 250/559.39, 250/559.45, 702/40, 702/82, 702/172, 702/179, 702/194
• US Class Current: 382/149
• CPC Class Codes: G06T 2207/30148 Semiconductor; IC; Wafer
• CPC Class Codes: G06T 7/001 using an image reference approach

Abstract

An apparatus and method for automatically detecting defects on silicon dies on silicon wafers comprising a silicon wafer acquisition system (30) and a computer (32) connected to said silicon wafer image acquisition system (10), wherein said computer (32) automatically aligns a silicon wafer (16), calibrates the image acquisition system (30), analyzes die images by determining a statistical die model from a plurality of dies, and compares the statistical die model to silicon die images to determine if the silicon dies have surface defects, is disclosed.

Claims

1. A method for automatically detecting detecting on silicon dies on a silicon wafer comprising the steps of:

aligning a silicon wafer;

calibrating the focal plane of an image acquisition system;

adjusting the lighting conditions of said image acquisition system;

identifying a sample of said silicon dies using an image acquisition system;

calculating a statistical die model from said sample;

comparing the statistical die model to said silicon dies at a pixel neighborhood level and a die level;

determining if said silicon dies have surface defects at said pixel neighborhood level and said die level;

determining if the percentage of defective dies exceeds a maximum threshold trip point;

adjusting the trip point of the statistical model die level comparison if said percentage of defective dies exceeds the maximum threshold; and

displaying the results of the comparison.

2. The method as recited in claim 1 further comprising the step of compiling a list of said silicon dies that were found to be defective.

3. The method as recited in claim 2 wherein said list of silicon dies is displayed on a display unit.

4. The method as recited in claim 2 further comprising the step of printing said list of silicon dies.

5. The method as recited in claim 2 wherein said results of the comparison are displayed graphically as an image representing said silicon wafer on a display unit to summarize the location of defective silicon dies.

6. A method for automatically detecting defects on silicon dies on a silicon wafer comprising the steps of:

aligning a silicon wafer;

calibrating the focal plane of an image acquisition system;

adjusting the lighting conditions of said image acquisition system;

identifying a sample of said silicon dies using an image acquisition system;

calculating a statistical die model from said sample;

comparing the statistical die model to said silicon dies at a pixel neighborhood level and a die level;

determining if said silicon dies have surface defects at said pixel neighborhood level and said die level;

determining if the percentage of defective dies exceeds a maximum threshold trip point;

adjusting the trip point of the statistical model die level comparison if said percentage of defective dies exceeds the maximum threshold; and

displaying a graphical image of said silicon wafer and indicating on said graphical image of said silicon wafer the locations and identities of said defective dies.

7. A method for automatically detecting defects on silicon dies on a silicon wafer comprising the steps of:

aligning a silicon wafer;

calibrating the focal plane of an image acquisition system;

adjusting the lighting conditions of said image acquisition system;

identifying a sample of said silicon dies using an image acquisition system;

calculating a statistical die model from said sample;

comparing the statistical die model to said silicon dies at a pixel neighborhood level and a die level;

determining if said silicon dies have surface defects at said pixel neighborhood level and said die level;

determining if the percentage of defective dies exceeds a maximum threshold trip point;

adjusting the trip point of the statistical model die level comparison if said percentage of defective dies exceeds the maximum threshold;

storing the average gray scale value for each of said pixel neighborhoods for each individual one of said silicon dies within a silicon chip matrix; and

displaying said silicon chip matrix.