Image analysis – Applications – Manufacturing or product inspection
Reexamination Certificate
1998-01-12
2003-07-01
Johnson, Timothy M. (Department: 2625)
Image analysis
Applications
Manufacturing or product inspection
C382S169000, C382S219000, C382S274000, C250S305000, C250S397000
Reexamination Certificate
active
06587581
ABSTRACT:
BACKGROUND OF THE INVENTION
1. Field of the Invention
The present invention relates to a method and apparatus which obtains two-dimensional images of an object (using an electron beam or optical arrangement), removes brightness level differences or distortions between the images via correction or equalization, and compares the resultant images so as to more accurately detect any defect contained in the object, for example, defects within a fine pattern of semiconductor wafers.
2. Description of Related Art
As a background visual inspection method for semiconductor wafers, there has been used a method for picking up, by use of optical means, two-dimensional images of different wafer areas or chips having corresponding patterns thereon (which should be essentially the same), and then to compare two detected images to determine differences therebetween as defects. More particularly, a simple differential image between the images is obtained, and any portion having a large differential value is regarded as a defect.
As a differing method not based on a simple differential image, there is a method described in, for example, “Automated Visual Inspection Of LSI Wafer Patterns Using A Derivative-Polarity Comparison Algorithm”, SPIE Vol. 1567 Applications of Digital Image Processing XIV, pp. 100-109 (1991). According to this method, two optically detected brightness images to be compared are differentiated respectively, then only the polarities (in which direction the brightness gradient faces) of the differential value are extracted, and by comparing these polarities, any difference in brightness which occurs in the normal portions of two images makes it possible to detect fine defects in shape.
Also, as another comparison method, there is described “Precision Visual Inspection of LSI Wafer Pattern Using Local Perturbation Pattern Matching Method”, i.e., in the thesis journal of Electronic Information Communications Association, Vol. J72-D-2, No. 12, pp. 2041-2050 (1989). This method is arranged to permit misregistration of not more than 1 pixel between two images optically detected, and also to permit any difference in brightness up to a certain specified level, and thus such method makes it possible to detect fine defects in shape without falsely detecting any difference in brightness occurring in a normal portion.
When defects in patterns in a fabrication process for semiconductor wafers having fine structure are inspected using an optical inspection method of the aforesaid background art, there were problems in that, for example, any non-opening defect of fine conducting holes, linear etching remainder (the width of whose short side is below the resolution of the optical system), and the like, cannot be detected because of the finely made wiring pattern formed on the semiconductor wafer. On the other hand, according to the inspection method using electron beams, it is possible to inspect defects in circuit patterns in the fabrication process of semiconductor wafers having the aforesaid fine structure.
When, however, for example, electron beams are irradiated to detect circuit patterns using secondary electrons from the object, the emission efficiency of the secondary electrons is determined by not only the material of the object, but also its film thickness, the acceleration voltage of the electron beams, the potential distribution in the vicinity of the object, and the like, and therefore, portions of the same material are not always detected at the same brightness. There are also some cases where a great difference in film thickness (which would cause no problems in the operation of the element) is erroneously determined defective as a result of a detected great brightness difference. There are also some cases where charge (so-called charge-up) accumulated in one area during detection forms a potential distribution which affects detection results of another area in its vicinity, and which results in brightness levels in the second area which are distorted.
When an image of the object is thus detected using electron beams, there was a problem that a pattern that is expected to be essentially detected at the same brightness is detected at a greatly differing brightness, and the difference in brightness is subsequently (i.e., erroneously) detected as a defect using the above brightness comparison inspection system. Also, according to the comparison system using the aforesaid differential polarity of the brightness image, a good result can be obtained for an image having little noise. However, if an image has high noise (e.g., in a system detecting images at high speed using electron beams), it is difficult to apply a differential equation as the same is prone to be affected by noise. Further, a method using a local perturbation pattern matching permits differences in brightness between images up to a certain specified level unconditionally, but if such differences in brightness violate the specified level, it becomes impossible to accurately detect only actual defects without detecting false detects.
As a further problem in the art, when electron beams are irradiated to detect circuit patterns using secondary electrons from the object, it was previously described that the electron beams are affected by the electrical conductivity of the inspection object. This not only causes the brightness of the detected image to become different, but also affects the irradiation position of the electron beams by influences of an electromagnetic field effect, and any misregistered positioning appears as a distortion in the detected image. Further, since this electrical conductivity depends upon the material of the object, the amount of misregistration due to this electrical conductivity distortion depends upon the electrical conductivity, the distribution of the material, and the thickness of the material, and further, the aforesaid amount of misregistration is not constant, but varies unsteadily.
As still a further problem, since a stage with an object specimen mounted thereon travels in vacuum in the electron beam method, and because there is a limit in how much mechanical friction can be reduced, vibration caused by friction of the stage during traveling also causes distortion in the detected image. The amount of slippage due to the above causes can be greater than a delicate fluctuation of the wiring pattern due to processing, and therefore, it is difficult to directly use the comparison system using differential polarity and the method based on the local perturbation pattern matching.
In the foregoing, description has been made of the case where electron beams are irradiated to detect circuit patterns using secondary electron from the object, and similar problems also occur in an optical inspection method because of the introduction of the fabrication process of CMP (Chemical Mechanical Polishing) and larger-sized wafer. CMP requires, in the fabrication process of semiconductors, polishing of each lamination layer to flatten the upper surface thereof, so as to prevent irregularity and waviness in the wiring structure for realizing a high level wiring structure. Since the layer on the top face has been flattened by polishing at the time of the inspection, a thickness of the layer (e.g., silicon dioxide) differs at differing surface positions. As a result, during optical inspection through the layer, an interference condition differs with position when viewed optically, diversified color is presented on an optically-observed image, resulting in varied brightness when detected on the monochromic brightness image.
As a further future difficulty, uniform formation of a wiring pattern over an entire wafer surface becomes more and more difficult as wafers continue tending toward larger-sized wafers. In the case of comparison inspection, it is conceived that the farther respective comparison patterns become separated, the greater a difference in pattern becomes. Thus, it is difficult to directly use the comparison system using differential polarity presupposing the delicate fluctuation and the meth
Azuma Junzou
Hiroi Takashi
Kuni Asahiro
Maeda Shunji
Matsuyama Yukio
Antonelli Terry Stout & Kraus LLP
Chawan Sheela
Hitachi , Ltd.
Johnson Timothy M.
LandOfFree
Visual inspection method and apparatus therefor does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Visual inspection method and apparatus therefor, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Visual inspection method and apparatus therefor will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3032018