JP2701872B2 - Surface inspection system - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION [Industrial Application Field] The present invention relates to a photomask used in a photoetching process such as an IC lead frame, and a photomask and a product used in an IC manufacturing process. The present invention relates to an apparatus for inspecting defects such as scratches, pinholes, black spots, and dust of an industrial product having a broken structure.

[Conventional technology]

Conventionally, the defect inspection of industrial products as described above is naked eye or
It is customary to use a microscope to perform visual observations, but it requires a lot of labor to inspect a large number of products, and it lacks inspection accuracy and reliability because it is a sensory test. There was a problem.

In order to solve such a problem, for example, with respect to defect inspection of an imposed pattern in which unit patterns are arranged at equal pitches, such as an IC photomask, two unit patterns of the unit patterns are arranged. A signal obtained by imaging each of the corresponding portions with an imaging unit such as a line sensor is compared with a predetermined threshold and binarized, and the signals are compared with each other to recognize a different portion as a defect, An apparatus for performing a defect inspection is used.

[Problems to be solved by the invention]

In such a conventional defect inspection method for imposed patterns in which binarized pixels are directly compared with each other, the pixels of the imaging system must be set smaller than the size of the defect to be detected in principle. In order to perform good detection operation with
The error of the mechanical system during the operation must be made sufficiently smaller than the pixel, so that the device becomes extremely accurate and is expensive. In addition, when there is a large error in the equal pitch arrangement or when the arrangement is not equal pitch, it cannot be detected, and furthermore, defects that do not reach the binary threshold cannot be inspected.

The present invention has been made to solve the above problems, and provides an imprint pattern defect inspection apparatus capable of inspecting imprint pattern defects with high accuracy in a short time without being affected by imposition accuracy or arrangement. The purpose is to:

[Means for solving the problem]

The present invention relates to an apparatus for inspecting a defect of an imposed pattern in which a plurality of unit patterns are arranged, wherein the imaging means for imaging a part of one imposed pattern, and the one imposed pattern and the imaging means are relatively positioned. Moving means for moving to the unit pattern, measuring means for measuring the relative position of the unit pattern and the imaging means to image from the image data obtained by imaging each edge region set in the unit pattern of the one imposed pattern, Image processing means for processing image data obtained by imaging to detect a defect; and control means for controlling each of the means, wherein the control means is configured to control the one of the ones based on a measurement result of the measurement means. The moving means is driven and controlled such that the relative positions of each unit pattern for imaging the imposed pattern and the imaging means are equal to each other.

Further, the present invention is characterized in that image data obtained by imaging is image data obtained by frame integration.

Further, the present invention is characterized in that the image pickup unit also picks up an image of the area of the transmission part at each image pickup position, and based on the image pickup result, the image processing unit detects a light amount fluctuation of the light source and corrects the image data of the unit pattern. I do.

[Operation] The defect inspection apparatus for an imposed pattern of the present invention measures the relative position between the pattern to be imaged and the imaging unit by comparing the total value of the pixel data of each edge area set for the unit pattern. By capturing the image data obtained by imaging when the relative position of the pattern to be imaged and the imaging unit are made equal to each other based on the measurement result, and comparing the captured image data of the unit patterns, a defect of the imposed pattern is detected. Can be detected. In addition, the light amount fluctuation is obtained by capturing the image data of the area of the transmission part at each imaging position, the image data of the unit pattern is corrected by the light amount fluctuation, and the image data of the unit patterns is compared with each other to reduce the influence of the light amount fluctuation. Can be removed.

〔Example〕

 Hereinafter, the present invention will be described in detail based on examples.

FIG. 1 is a block diagram showing an embodiment of an imposed pattern inspection apparatus according to the present invention, FIG. 2 is a diagram showing an example of an imposed pattern, FIG. 1 (a) is a diagram showing a unit pattern, and FIG. FIG. 3B is a view showing an imposed pattern in which the unit pattern (a) is arranged at a plurality of places, FIG. 3C is a view showing an etching region set for the unit pattern, and FIG. FIG. 9 is an explanatory diagram illustrating a detection operation. In the figure, 1 is an imposition pattern, 2 is an XY stage, 3 is an illumination unit, 4 is a power supply, 5 is an imaging lens, 6 is a beam splitter, 7 is an enlarging lens, 8 and 9 are imaging elements, and 10 and 11 are An image sensor driving circuit, 12 is an image processing device for defect detection, 13 is an image processing device for position detection, 14 is an XY stage driving circuit, and 15 is a control unit.

A part of the imposed pattern 1 is illuminated by the illumination unit 3 which is turned on by the DC stabilizing power supply 4, and an optical image by the imaging lens 5 is captured by the imaging device 8 by driving the imaging device driving circuit 10. The image processing device 12 converts the video signal from the image sensor 8 into an A /
A defect is detected based on the D-converted image data. 13 is an image processing apparatus for measuring the relative position between the imaging pattern 13 and the imaging unit. The optical image obtained by the imaging lens 5 is separated by a beam splitter 6, and this image is enlarged by a lens 7. The video signal obtained by imaging with the image sensor 9 is A / D converted and output to the control unit 15 for obtaining the relative position from the image data.
The control unit 15 includes an XY stage drive circuit 14, image processing devices 12, 13,
The entire system is controlled and the inspection operation is performed.

Next, an operation for detecting a defect in the imposition pattern according to this embodiment will be described. First, the imaging control unit 15 regions A ₁ indicated by broken lines in Figure 2 by driving the XY stage 2 (b) by an XY stage driving circuit 14 is set to be an imaging region of the imaging element 8, the image processing The image data of the unit pattern is taken into the image memory of the device 12. In addition, the edge areas W _X1 , W _X2 , W _Y1 and W _Y2 are set, and the portion hidden by the pattern in the edge area is expressed in gradations, for example, as 8-bit data for each pixel, and the sum of the numerical values is obtained for each edge area. Then by driving the XY stage on the basis of the design values of the array of unit patterns, to drive the Menzuke pattern A ₂ to the next unit pattern is the imaging region. However, when an area to be imaged is moved from one unit pattern to another unit pattern using a mechanism such as an XY stage based on the design value of the pattern, an error of the moving mechanism, an imposition position error of the pattern, pattern not sufficient pattern relative position repeatability of the imaging system in the area a ₁ and a ₂ because there is such error when mounting the stage or the like, which may produce a large error. As described above, when comparing image data captured in a state where the relative positions are different from each other, since a position error becomes a defect signal, a minute defect cannot be detected.

For correcting an error of such a relative position, in the area A _2, likewise the edge region and did in region A ₁ W _X1, W _X2,
The sum of 8-bit pixel data is _{calculated for} each of W _Y1 and W _Y2 . By determining the total value of the pixel data for each edge area in this manner, the influence of jagged edges can be reduced. Furthermore W _X1 -W _X2, W _Y1 Calculating the -W _Y2 pattern and the relative position information of the imaging system is obtained, the deviation direction and approximate magnitude of the comparison when the relative position and the value of the area A ₁ is obtained . Since this process is a pixel data operation of the obtained area, a result can be obtained in several tens of ms. Then, based on this result, the position is corrected by feeding back to the XY stage, and the relative position error can be reduced to the order of the position detection accuracy. Although the position detection accuracy of this method is affected by the pixel size and the video signal S / N, the pixel size is reduced by increasing the imaging magnification of the position detection imaging element 9 by the lens 7 as shown in FIG. In addition, the influence of noise included in the video signal is obtained by calculating the sum of a large number of pixel data. And the detection accuracy can be improved. For example, if the number of pixels in the vertical direction of W _X1 is 100, the noise component is In the case of one frame of image data, even if there is a 5% change in the data for each pixel, the average by adding 100 pixels is about 0.5%. When the pixel size is about several%, that is, when the pixel size is 50 μm, the position can be detected with an accuracy of 1 to 2 μm. If the target image data of this position detection processing is frame-integrated data, the noise component can be further reduced, and the same applies even if the number of pixels is increased by setting the number of edge areas of W _X and W _Y to be large. The effect is obtained. Also, since this method calculates the difference between the pixel data sums of the edge areas set at both ends of the unit pattern, there is a difference in the pattern between the two unit patterns, and there is a slight variation in the illumination intensity during each imaging. Is also canceled out, and there is an advantage that no detection error occurs.

After the above relative position correcting operation captures the image data in the area A ₁ and likewise region A _2, to detect a defect from the difference between the two data. 3A and 3B illustrate a method of detecting a defect. FIG. 3A shows a portion including defects P and Q, and FIG. 3B shows a portion corresponding to FIG. The image data of one frame of the scanning line passing over the defect in (a) is V _a1 , and the corresponding data on (b) is V _b1 . Usually, the data for one frame contains several percent of random noise in time, so the difference data between the two data is obtained.
From S1, _only a few% of the pixel area, that is, only a defect of several hundred μm ² or more when the pixel is 100 μm can be detected. However, if the image data in the regions A ₁ and A ₂ is data obtained by frame integration over a plurality of frames, the random noise is expressed as the number N of integrated frames. And the S / N is improved, and a defect signal buried in noise such as V _an and V _bn appears. By taking the difference between the two data, even a fine defect like S ₂ shown in (h) It can be detected.

An example in which a defect inspection of an imposed pattern is performed according to the above embodiment is described. An image of a 50 m / m square area is taken at 100 μm per pixel, and a diameter of 15 μm is obtained from image data obtained by integrating 30 frames.
m defects could be detected.

By the way, in the above embodiment, if there is a change in the brightness of the light source or the like, the image data of the scanning line passing over the defect P with respect to the unit pattern as shown in FIG. FIG. 5A shows V ₁ , and the image data obtained by scanning the defect-free position shown in FIG. 4B becomes V ₂ (FIG. 5B). .
The difference Vd (FIG. 5 (c)) may be detected as a defect due to fluctuations in the brightness of the light source in addition to the defect.

Therefore, the position detection image processing device 13 shown in FIG. 1 sets a window W indicated by a broken line in FIG. 4, for example, to set image data of the transparent portion (white portion), that is, in the area of FIG. detecting the average I _B of the pixel values in the window W in the area of the average I _a of the pixel values in the window W, FIG. 4 (b), α = I _a
/ A I _B calculated as a correction value, it performs the following calculation equation in the control unit 15.

V ₂ ′ = αV ₂ (FIG. 6 (b)) Vd ′ = V ₁ −V ₂ ′ (FIG. 6 (c)) By this processing, the brightness of the light source is reduced as shown in FIG. 6 (c). Differences in image data due to fluctuations and the like can be prevented from appearing in the difference image, and as a result, even minute defects can be detected.

〔The invention's effect〕

As described above, according to the present invention, the imposition position accuracy of the minute defect of the industrial product having the imposition pattern, the arrangement method,
Detection can be performed automatically without being affected by mechanical accuracy and the like, and effects such as improved detection accuracy, reliability, and efficiency can be obtained. Further, when comparing the image data obtained by capturing each unit pattern, after correcting the change in the image data of the illumination unit, it is possible to detect even a minute defect by comparing the difference between the image data with a threshold value.

[Brief description of the drawings]

FIG. 1 is a block diagram showing an embodiment of an imposed pattern defect inspection apparatus according to the present invention, FIG. 2 is a diagram showing an example of an imposed pattern, and FIG. 1 (a) is a diagram showing a unit pattern. FIG. 3B is a diagram showing an imposition pattern in which unit patterns are arranged at a plurality of positions, FIG. 3C is a diagram showing an edge region set for the unit pattern, and FIG. 3 is a defect detection operation of the present invention. FIG. 4 is a diagram showing a unit pattern at a position where a defect exists and a position without a defect. FIG. 5 is a diagram showing image data when there is a light amount fluctuation. FIG. FIG. 7 is a diagram showing image data in the case. DESCRIPTION OF SYMBOLS 1 ... Surface pattern, 2 ... XY stage, 3 ... Lighting part, 4 ... Power supply, 5 ... Imaging lens, 6 ... Beam splitter, 7 ... Magnifying lens, 8, 9 ... Imaging element, Ten,
11 image sensor driving circuit, 12 image processing device for defect detection, 13 image processing device for position detection, 14 XY stage driving circuit, 15 control unit.

Claims (3)

(57) [Claims] In an apparatus for inspecting a defect of an imposed pattern in which a plurality of unit patterns are arranged, an image pickup means for picking up an image of a part of one imposed pattern, and the one imposed pattern and the image pickup means are relatively positioned. Moving means for moving the unit pattern, and measuring means for measuring a relative position between the unit pattern and the image pickup means, which are imaged from image data obtained by imaging each edge area set in the unit pattern of the one imposed pattern. Image processing means for processing image data obtained by imaging to detect a defect, and control means for controlling each of the means, wherein the control means is configured to execute the first processing based on the measurement result of the measurement means. A defect inspection apparatus for an imposed pattern, wherein the moving means is driven and controlled such that the relative positions of each unit pattern for imaging two imposed patterns and the imaging means are equal to each other. 2. An apparatus according to claim 1, wherein the image data obtained by imaging is frame-integrated image data. 3. The image pickup means also picks up an image of a transmissive area at each image pickup position, and based on the image pickup result, the image processing means detects a change in the light amount of the light source and corrects the image data of the unit pattern. Or a defect inspection device for an imposed pattern according to 2.