Method for detecting defects on wafers, system for detecting defects on wafers
A method for detecting defects on a wafer including the steps of obtaining a reference image of a chip pattern formed on a reference wafer, using a computer algorithm to analyze the reference image to produce a division map for the chip pattern; setting respective thresholds for divisions of the division map, obtaining a comparison data between a test image of the chip pattern formed on a test wafer and the reference image, using the division map and the thresholds to examine the comparison data to identify a defect in the test image.
1 . A method for detecting defects on a wafer, comprising:
obtaining a grayscale reference image of a chip pattern formed on a reference wafer;
using a computer algorithm to analyze the grayscale reference image to produce a division map for the chip pattern, wherein the division map comprises a plurality of divisions;
grouping the divisions according to the grayscale levels of the divisions, wherein the divisions are grouped into 5 to 10 groups;
setting respective grayscale difference thresholds for the groups of the divisions, wherein a grayscale difference threshold of a group of darker divisions is smaller than a grayscale difference threshold of a group of whiter divisions;
obtaining a grayscale difference data between a grayscale test image of the chip pattern formed on a test wafer and the grayscale reference image; and
using the division map and the grayscale difference thresholds of the divisions to examine the comparison data to identify a defect in the grayscale test image.
2 . The method for detecting defects on a wafer according to claim 1 , wherein the comparison data comprises differences in grayscale value between the grayscale test image and the grayscale reference image.
3 . The method for detecting defects on a wafer according to claim 1 , wherein the grayscale difference thresholds of the divisions of the same group are the same.
4 . The method for detecting defects on a wafer according to claim 1 , further comprising:
recognizing a boundary between two of the divisions having different grayscale difference thresholds; and
shifting the boundary toward one of the two divisions that has a smaller grayscale difference threshold.
5 . The method for detecting defects on a wafer according to claim 4 , wherein the boundary is shifted by 3 um.
6 . The method for detecting defects on a wafer according to claim 1 , wherein the grayscale reference image and the grayscale test image are obtained by a same inspection tool.
7 . The method for detecting defects on a wafer according to claim 6 , wherein the inspection tool is a SEM or an optical imaging tool.
8 . A system for detecting defects on a wafer, comprising:
an inspection tool, configured to obtain a grayscale reference image and a grayscale test image of a chip pattern; and
a computer, configured to:
analyze the grayscale reference image to produce a division map for the chip pattern, wherein the division map comprises a plurality of divisions;
group the divisions according to the grayscale levels of the divisions, wherein the divisions are grouped into 5 to 10 groups;
set respective grayscale difference thresholds for the groups of the divisions, wherein a grayscale difference threshold of a group of darker divisions is smaller than a grayscale difference threshold of a group of whiter divisions;
calculate differences between the grayscale reference image and the grayscale test image to provide a grayscale difference data; and
examine the grayscale difference data based on the division map and the grayscale difference thresholds of the divisions to identify a defect in the grayscale test image.
9 . The system for detecting defects on a wafer according to claim 8 , wherein the inspection tool is a SEM or an optical imaging tool.
10 . The system for detecting defects on a wafer according to claim 8 , further comprising a memory configured to store the grayscale reference image, the grayscale test image, and the division map.
11 . The system for detecting defects on a wafer according to claim 8 , wherein the computer is further configured to:
recognize a boundary between two of the divisions having different grayscale difference thresholds; and
shift the boundary toward one of the two divisions that has a smaller grayscale difference threshold.