Image analysis – Applications – Manufacturing or product inspection
Reexamination Certificate
1998-04-22
2001-01-16
Mehta, Bhavesh (Department: 2721)
Image analysis
Applications
Manufacturing or product inspection
C348S125000, C356S237100
Reexamination Certificate
active
06175646
ABSTRACT:
TECHNICAL FIELD OF THE INVENTION
The present invention relates in general to the field of silicon wafer processing, and more particularly, to an apparatus and method for automating the detection of defects in individual silicon chips.
BACKGROUND OF THE INVENTION
Without limiting the scope of the invention, its background is described in connection with the processing of silicon chips on silicon wafers, such as dynamic random access memory (DRAM), as an example.
Heretofore, in this field, the detection of defects on silicon wafers within individual silicon chips has been performed either manually, or using automated visual systems having a defect detection resolution of 2-4 microns. Unfortunately, the accuracy of detection in many cases was limited by the operator's ability to view a random sample of conventional silicon dies within a specified period of time.
Conventional automated systems have achieved up to a 90% success rate in detecting silicon wafers surface anomalies. These success rates, however, were achieved only under the best conditions of light and contrast. Importantly, reliability of conventional automated systems has been found to degrade rapidly if conditions of lighting and contrast are not ideal.
Furthermore, the limitations of current methods for automatically inspecting silicon dies for defects require substantial visual inspection by an operator. For example, the field of view of present systems is limited to one die, thereby excluding from analysis up to 60% of the available 256 kilobytes of data obtained from a field of view of 125×150 mils. This field of view limits spacial accuracy to about 1:16, with a sub-pixel alignment of 0.25. The problem of the limited field of view of present systems is exacerbated by the inherent difference in the lighting and contrast of the underlying silicon wafer background. The inability of present systems to cope with differences in lighting and contrast is a major stumbling block to further automation of silicon die analysis because differences in wafer background are found to occur even between different silicon wafer lots for the same type of silicon chip.
What is needed is an automated imaging system that is customizable for each silicon wafer that is processed. Also, a need has arisen for an automatic inspection system that is able to adapt to different silicon chip patterns, and that accurately detects surface defects on silicon dies on a silicon wafer. A need has further arisen for a system that is able to meet the needs of high throughput without a loss of accuracy. Finally, a need has arisen for an automated system that is able to adapt to the high precision needs of future silicon chip designs.
SUMMARY OF THE INVENTION
The present invention disclosed herein is an apparatus and method for detecting defects on silicon dies that comprises an image acquisition system and a computer connected to the image acquisition system. The image acquisition system of the present invention captures one or more die images simultaneously. The computer analyzes a random sample of die images to create a statistical die model or standardized die image. The statistical die model is then compared to silicon dies images on a silicon wafer to determine if the silicon dies have surface defects.
In one embodiment, the image acquisition system of the present invention simultaneously captures a plurality of die images in the form of a matrix of die images. An example of a matrix of silicon dies is a two by three die image matrix. The image acquisition system of the present invention uses a charge-coupled display (CCD) camera, which may be a 512 by 512 pixel CCD camera (512K CCD camera). Alternatively, a 1K×1K or 4K×4K CCD camera may be used to increase the field of view, or to increase the resolution of the image obtained over the same area as the 512K CCD camera.
The apparatus for detecting defects on silicon dies, in one embodiment, can automatically align the silicon wafer based upon a predetermined reference chip. The apparatus physically aligns the silicon wafer by instructing an automated wafer handling system as to the relative position of the silicon wafer using a coarse and a fine image adjustment.
The image acquisition system can further include a high resolution microscope for manually inspecting silicon dies, if desired. The image acquisition system of the present invention can also include an object character recognition (OCR), a Bar code reader or other system that provides the computer with information about the silicon wafer. The computer of the present invention can be connected to a display unit that displays the die images acquired by the image acquisition system. The display unit of the present invention can be, for example, a touch screen cathode ray tube that permits input from the screen to the computer. The display unit can display an image of a silicon wafer that contains a summary of the results from comparing the statistical die model with a plurality of dies. The image acquisition system of the present invention may further comprise a wafer cassette that provides the wafer handling system with silicon wafers.
The present invention also includes a method for detecting defects on silicon dies on a silicon wafer comprising the steps of, identifying a random sample of silicon dies on the surface of the silicon wafer, generating a statistical die model from the random sample, comparing the statistical die model to silicon dies, determining if the silicon dies have surface defects and displaying the results of the comparison.
More specifically, the method of the present invention involves capturing the image a plurality of silicon chips on a silicon wafer, converting the image of the silicon chips into pixels, separating the pixels into neighborhoods of pixels, determining a neighborhood value for each neighborhood of pixels, comparing each neighborhood value with the corresponding neighborhood value of the statistical die model to determine whether silicon dies are defective, and reporting the results. In the method of the present invention the neighborhood of pixels may be further defined as being a square matrix of pixels, a rectangular matrix of pixels, a circular matrix of pixels or an overlapping an adjacent neighborhood of pixels. The neighborhood of pixels may be a nine by nine matrix of pixels.
REFERENCES:
patent: Re. 33956 (1992-06-01), Lin et al.
patent: 5428442 (1995-06-01), Lin et al.
patent: 5537325 (1996-07-01), Iwakiri et al.
patent: 5621813 (1997-04-01), Brown et al.
patent: 5854674 (1997-05-01), Lin
patent: 5943551 (1999-08-01), Schemmel et al.
Schemmel Floyd
Thorne Richard
Brady W. James
Courtney Mark E.
Mehta Bhavesh
Telecky , Jr. Frederick J.
Texas Instruments Incorporated
LandOfFree
Apparatus for detecting defective integrated circuit dies in... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Apparatus for detecting defective integrated circuit dies in..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Apparatus for detecting defective integrated circuit dies in... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2462972