Browse Prior Art Database

Increased collection of defect information in inline defect inspection

IP.com Disclosure Number: IPCOM000188724D
Original Publication Date: 2009-Oct-20
Included in the Prior Art Database: 2009-Oct-20
Document File: 2 page(s) / 108K

Publishing Venue

IBM

Abstract

All the optical defect images are stored on optical PLY tool for each inspection, however majority of these images typically are not used; they were removed after a period of time due to storage limitation on the tool. More useful information should be extracted and used from these optical images. This invention utilizes vast amount of images previously wasted and condenses them into manageable information by fail pattern, killer defect images etc. Condensed information can find many uses such as fail pattern prediction, limited yield calculation for defect types, prioritize defect SEM review, etc

This text was extracted from a PDF file.
At least one non-text object (such as an image or picture) has been suppressed.
This is the abbreviated version, containing approximately 92% of the total text.

Page 1 of 2

Increased collection of defect information in inline defect inspection

In this invention, we analyze defect image as soon as it is created; mapping defect location and shape to layout information to predict circuit failure patterns. If a real circuit failure is predicted, the optical image, predicted fail pattern, corresponding layout image are saved together with the basic defect information (coordinates, size, etc). If the defect does not cause real circuit failure, discard its optical image. This will help store relevant killer defect images while discarding other information not so useful. Saved information has many possible applications: classify these killer defects based on defect type, and calculate yield impact of each defect type to prioritize defect elimination efforts.

Here is the flow chart of this invention:

Here is an example showing how fail pattern in SRAM can be predicted based on location and shape of the defect:

1

[This page contains 4 pictures or other non-text objects]

Page 2 of 2

In a lot of cases, precise coordinate of the defect is not needed to predict fail patterns. For example, in memory layout, defect in different locations in the array will cause the same fail pattern, if it shorts the same pair of poly lines: in this case, it is more important to match determine the relationship of the defect to the design shapes...