Image analysis – Applications – Manufacturing or product inspection
Reexamination Certificate
2000-02-03
2003-06-03
Boudreau, Leo (Department: 2621)
Image analysis
Applications
Manufacturing or product inspection
C250S370100, C356S622000, C356S237400
Reexamination Certificate
active
06574359
ABSTRACT:
BACKGROUND OF THE INVENTION
1. Field of the Invention
This invention relates generally to semiconductor fabrication, and more particularly to methods and apparatus for inspecting features on a wafer surface.
2. Description of the Related Art
Wafer inspections are routinely performed at numerous stages during the processing of semiconductor integrated circuits. Two common inspection points include after-develop-inspection (“ADI”) and post-chemical-mechanical-polishing (“CMP”) inspection. ADI is performed to ensure that the steps leading up to and including resist develop have been performed correctly and to within the specified tolerances for the particular process. The goals of the ADI procedure are to verify that the correct mask has been used, that the resist film is free from contamination, scratches, bubbles, striations, etc., and that the defect types and populations are cataloged to enable subsequent yields and defect occurrences to be correlated.
The goals of post CMP inspection procedure are to ensure that the polished film exhibits an acceptable level of defects in the form of surface particles, pits or craters and scratches. It is desirable for the post CMP inspection procedure to not only identify the locations of defects on the polished film but also to characterize the topography of the defects, that is, whether the defects are pits or surface particles. The need to distinguish pits and craters from surface particles arises because such classes of defects are the product of quite different mechanisms during CMP processing, and thus entail different remedial measures. Surface particles on the polished film may indicate the presence of particulate contaminants originating from the CMP tool polish pad, the slurry or from another source. Pits or craters on the other hand are frequently the result of rip-outs that occur during CMP. Rip-outs are usually caused by poor process control during CMP. Typical causes for such poor CMP process control include, for example, undesirably large abrasive particles in the CMP slurry, undesirable variations in pad pressure applied to the polished surface, and irregular dispersion of CMP slurry during CMP to name just a few. Some types of films are more susceptible to rip-out than others. For example, tetra-ethyl-ortho-silicate (“TEOS”) films often exhibit higher rates of CMP rip-out than other types of insulator materials. This propensity for CMP rip-out for TEOS is due largely to the kinetics of the chemical vapor deposition processes used to form such films.
Conventional techniques for detecting defects in resist and CMP films include, for example, optical microscopy and laser scanning. In optical microscopy inspection, a human operator visually inspects the surface of the resist or polished film to identify and catalog defects. Given the immense quantities of defects typically present on a resist or polished film, the operator only observes a small quantity of defects in selected areas of the film. Furthermore, the operator will typically inspect only a small percentage of the wafers in a given lot. During the inspection, the operator catalogs the defects observed and characterizes them based on topography, that is, as a pit or crater or a surface particle. From the very limited sampling, predictions are then made about the size and composition of the defect population for not only the wafer inspected but also for the overall lot of wafers. Although the human operator is in many cases able to distinguish surface particles from pits in the inspected film, the process is time consuming and may yield inaccurate predictions about the size and composition of the defect population for a given wafer or lot of wafers since inspection time constraints only allow for a relatively small percentage of a film surface to be visually inspected.
Laser scanning defect inspection has eliminated some of the time constraints that are associated with human operator optical inspection. In many conventional laser inspection systems, a laser is focused down to a relatively small spot which is then scanned across the surface of the resist or polished film. Light reflected off the film is detected by two or more photodetectors that are positioned on either side of the scan axis. The detectors are designed to read the intensity of the scattered radiation. In many cases, current laser defect inspection systems do an acceptable job of identifying and distinguishing flat plane defects, such as pattern defects. However, current laser scanning systems do a relatively poor job of identifying the topography of a particular defect, that is, whether the defect is a pit or a surface particle. The difficulty stems from the fact that the average intensity of scattered radiation from a given size pit or surface particle is about the same. Thus, while conventional laser scanning techniques may be used to more quickly identify the total population of defects on a given film, the exact composition of the population of defects may not be determined.
The present invention is directed to overcoming or reducing the effects of one or more of the foregoing disadvantages.
SUMMARY OF THE INVENTION
In accordance with one aspect of the present invention, a method of inspecting a feature on a film surface is provided that includes illuminating the feature with laser radiation and detecting radiation scattered from the feature with a plurality of detectors. Each of the plurality of detectors has a known position. A position of the feature observed by each of the plurality detectors is computed based upon the radiation scattered from the feature. An A first value for each of the plurality of detectors is computed that is the scalar product of the known position of a given detector with the difference of the position of the feature observed by that given detector and the average of the positions observed by each of the plurality of detectors. A second value indicative of the topography of the feature is computed by summing first values for the plurality of detectors.
In accordance with another aspect of the present invention, a method of inspecting a feature on a film surface is provided that includes illuminating the feature with laser radiation and detecting radiation scattered from the feature with D detectors. Each of the D detectors has a known position {right arrow over (u)}
d
where d designates a given detector. A position {right arrow over (p)}
d
of the feature observed by each of the D detectors is computed based upon the radiation scattered from the feature. An average position {right arrow over (P)} of the positions {right arrow over (p)}
d
observed by each of the D detectors is computed and a scatter height descriptor S for the feature is computed according to the equation:
S
=
∑
d
(
p
⇀
d
-
P
⇀
)
·
u
⇀
d
D
.
In accordance with another aspect of the present invention, a method of inspecting a film surface is provided that includes subdividing the film surface into a plurality of pixel locations and sequentially illuminating each of the pixel locations with laser radiation. Radiation scattered from each of the pixel locations is detected with D detectors, where each of the D detectors has a known position {right arrow over (u)}
d
where d designates a given detector. A position {right arrow over (p)}
d
of each of a plurality of defects on the film surface observed by each of the D detectors is computed according to the equation:
p
⇀
d
=
∑
ij
p
⇀
ij
I
ijd
∑
ij
I
ijd
,
where I
ijd
is the intensity of scattered laser radiation detected by a given detector d for a given pixel location {right arrow over (P)}
ij
. An average position {right arrow over (P)} of the positions {right arrow over (p)}
d
observed by each of the D detectors is computed according to by the equation:
P
⇀
=
∑
ijd
p
⇀
ij
I
ijd
∑
ijd
I
ijd
.
A scatter height descriptor S for each of the plurality of defects is computed according to the equation:
S
=
∑
Advanced Micro Devices , Inc.
Boudreau Leo
Honeycutt Timothy M.
Patel Shefali
LandOfFree
Method and apparatus for inspecting wafer defects does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Method and apparatus for inspecting wafer defects, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method and apparatus for inspecting wafer defects will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3121773