Pattern defect inspection method and apparatus

• US 20050117796A1
• Filed: 11/24/2004
• Published: 06/02/2005
• Est. Priority Date: 11/28/2003
• Status: Active Grant

• Current Assignee: Hitachi High-Technologies Corporation (Hitachi, Ltd.)
• Original Assignee: Hitachi High-Technologies Corporation (Hitachi, Ltd.)
• Inventors: Suzuki, Katsuya, Matsui, Shigeru
• Granted Patent: US 7,457,455 B2
• Time in Patent Office: Days
• US Class Current: 382/145
• CPC Class Codes: G01N 21/95607 using a comparative method
• CPC Class Codes: G03F 7/7065 Defects, e.g. optical inspection of patterned layer for defects
• CPC Class Codes: G06T 2207/30148 Semiconductor; IC; Wafer
• CPC Class Codes: G06T 7/001 using an image reference approach

Abstract

The pattern defect inspection apparatus is operable to detect defects by comparing a detection image, which is obtained through scanning by an image sensor those patterns that have the identical shape and are continuously disposed on the object under tested at equal intervals in row and column directions, with a reference image obtained by scanning neighboring identical shape patterns in the row and column directions. This apparatus has a unit for generating an average reference image by statistical computation processing from the images of identical shape patterns lying next to the detection image including at least eight nearest chips on the up-and-down and right-and-left sides and at diagonal positions with the detection image being intermediately situated. The apparatus also includes a unit that detects a defect by comparing the detection image to the average reference image thus generated.

Claims

1. A pattern defect inspection apparatus for detecting defects by comparing a detection image obtainable through scanning, by an image sensor, patterns having an identical shape and being continuously laid out on an object to be tested at equal intervals in row and column directions with a reference image obtained by scanning patterns of the identical shape neighboring in the row and column directions thereof, said apparatus comprising:

means for generating an average reference image by statistical computation processing from images of identical shape patterns lying next to the detection image including at least four nearest chips on up-and-down and right-and-left sides with the detection image being intermediately situated; and

means for detecting a defect by comparing the detection image to the average reference image thus generated.

4. The pattern defect inspection apparatus according to claim 1, wherein the statistical computation processing for generating the average reference image is a processing for calculating an adaptable quadratic curved plane from the nearest identical shape patterns lying next to the detection image on the up-and-down and right-and-left sides and at the diagonally neighboring positions.

5. The pattern defect inspection apparatus according to claim 4, wherein a calculation scheme for obtaining the adaptable quadratic curved plane is a least-squares method, and wherein said apparatus further comprises means for calculating a mean square error simultaneously during this calculation, and means for using the resultant mean square error to determine a threshold value upon judgment of defect detection, and means for judging whether a defect is present or absent based on the threshold value.

2. A pattern defect inspection apparatus for detecting defects by comparing a detection image obtainable through scanning by an image sensor patterns having an identical shape and being continuously laid out on an object to be tested at equal intervals in row and column directions with a reference image obtained by scanning patterns of the identical shape neighboring in the row and column directions thereof, said apparatus comprising:

means for generating an average reference image by statistical computation processing from images of identical shape patterns lying next to the detection image including at least eight nearest chips on up-and-down and right-and-left sides and at diagonally neighboring positions with the detection image being intermediately situated; and

means for detecting a defect by comparing the detection image to the average reference image thus generated.

3. The pattern defect inspection apparatus according to claim 2, wherein said statistical computation processing for generating the average reference image comprises calculating a quadratic curved plane relating to a signal intensity on each corresponding pixel of the image of said eight nearest chips, obtaining an average signal intensity of the quadratic curved plane with respect to a pixel corresponding to a pixel at a g(x, y) point of said detection image, and comparing the average signal intensity to the signal intensity at said g(x, y) point.

6. A pattern defect inspection apparatus for detecting defects by comparing a detection image obtainable through scanning by an image sensor patterns having an identical shape and being continuously laid out on an object to be tested at equal intervals in row and column directions with a reference image obtained by scanning patterns of the identical shape neighboring in the row and column directions thereof, said apparatus comprising:

means for generating an average reference image by statistical computation processing from both the detection image and the identical shape patterns of at least eight nearest chips lying next to the detection image on up-down and right-left sides and at diagonally neighboring positions with at least the detection image being intermediately situated; and

means for detecting a defect by comparing the detection image to the average reference image thus generated.

7. A pattern defect inspection method for detecting defects by comparing a detection image obtainable through scanning by an image sensor patterns having an identical shape and being continuously laid out on an object to be tested at equal intervals in row and column directions with a reference image obtained by scanning patterns of the identical shape neighboring in the row and column directions thereof, said method comprising the steps of:

generating an average reference image by statistical computation processing from images of identical shape patterns lying next to the detection image including at least four nearest chips on up-and-down and right-and-left sides with the detection image being intermediately situated; and

detecting a defect by comparing the detection image to the average reference image thus generated.

8. A pattern defect inspection method for detecting defects by comparing a detection image obtainable through scanning by an image sensor patterns having an identical shape and being continuously laid out on an object to be tested at equal intervals in row and column directions with a reference image obtained by scanning patterns of the identical shape neighboring in the row and column directions thereof, said method comprising:

generating an average reference image by statistical computation processing from images of identical shape patterns lying next to the detection image including at least eight nearest chips on up-and-down and right-and-left sides and at diagonally neighboring positions with the detection image being intermediately situated; and

detecting a defect by comparing the detection image to the average reference image thus generated.