JP3202089B2 - Surface defect inspection method for periodic patterns - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method for inspecting a surface defect of a sample having a periodic opening such as a shadow mask.

[0002]

2. Description of the Related Art Conventionally, various methods have been proposed for defect inspection of industrial products in which unit patterns are periodically arranged, and among defects detected from a direction perpendicular to a sample surface, a video signal photographed by a camera is used. A method of detecting a pattern by examining a pattern, a method of detecting a defect by means of comparing a signal obtained by similarly photographing a pattern without a defect, or a method of detecting a periodic pattern light when irradiating coherent light. Although a method utilizing the diffraction phenomenon is known, oblique light inspection is conventionally performed by human eyes.

[0003]

In a sample having a periodic aperture, a surface defect due to a lack of a hole is a defect which is relatively easy to be found by observing oblique light by visual observation, but is an overlook caused by a human inspection. There is a problem that occurs. For defects visible from the perpendicular direction to the sample,
Although many inspection methods have already been proposed, no effective method has been proposed for oblique light inspection, and this is one of the obstacles to full automation of inspection.

SUMMARY OF THE INVENTION The present invention has been made to solve the above problems, and has as its object to provide a method of inspecting a surface defect of a periodic pattern, which can improve the inspection accuracy and enhance the possibility of fully automatic inspection. .

[0005]

SUMMARY OF THE INVENTION The present invention illuminates a sample with a light source, detects transmitted light from an oblique direction by means of a linear area photographing means, and photographs a linear area to eliminate defects in the periodic pattern. The detection method is characterized in that an elevation angle exceeding the slice level first when the predetermined slice level of the average value change amount of the sensor incident light amount with respect to the elevation angle change for a predetermined number of times is set as the optimum elevation angle.

Further, the present invention relates to a method for preparing an arbitrary sample 1 in advance.
When the optimum elevation angle of the entire surface of the sample measured as a reference is used as the reference data, and the optimum elevation angle of another sample is obtained, the optimum elevation angle is measured only at one arbitrary position, and the measurement result and the optimum elevation angle of the same position in the reference data are obtained. The data obtained by adding / subtracting the difference from the reference data to the entire reference data is used as the optimum elevation angle of the sample.

Further, according to the present invention, the maximum value of the input signal or its differential signal is obtained at any of a plurality of measurement points while scanning the sample at the set optimum elevation angle, and the maximum value or the maximum value is determined. The value obtained by adding / subtracting the above offset is used as the inspection slice level around each measurement point.

Further, the present invention is characterized in that the boundary of the perforated portion is identified from the shape of the primary differential waveform of the sensor input signal.

[0009]

According to the present invention, a linear sensor is used to control the elevation angle of a sensor or a sensor in the direction of an axis while moving the sample or the sensor in one axis direction so that the elevation angle at which the sensor looks at the sample is always an optimum angle. By capturing obliquely transmitted light from the sample, performing image processing such as differentiation on the captured image signal to detect surface defects, and rotating the sample or sensor to inspect from each direction, periodic aperture It is possible to automate the oblique light inspection of the sample having the above, improve the inspection accuracy, and improve the feasibility of fully automatic inspection.

[0010]

FIG. 1 is a view showing an embodiment of a surface defect inspection apparatus used in the present invention, and FIG. 2 is a view showing a parallel light detecting optical system. In the figure, 1 is a CCD line sensor camera, 2 is a linear light source, 3 is a lens, 5 is a rotary stage, 6 is a uniaxial moving stage, 7 is a sample, 8 is a holding column, 9 is a camera controller and a signal processing device, 10 Is a controller, 11 is a control computer, 13 is a camera slide device, 14 is an elevation angle variable device, and 15 is a glass plate.

In this embodiment, CC is used as a linear area photographing means.
As shown in FIG. 1C, which is a diagram viewed from above the stage, using a D-line sensor camera, the sample is moved in a uniaxial direction (in a direction orthogonal to the main scanning direction) in order to linearly capture the transmitted light of the sample. On the uniaxial movement stage 6 for moving in the direction (x direction in the drawing), a rotation stage 5 for placing and rotating the sample (θ ₂ direction in the drawing) so that the sample can be photographed from various directions is arranged, and further thereon. A transparent plate 15 made of a transparent acrylic plate or glass plate is arranged so that light can pass therethrough.
FIG. 1 (a) showing the entire configuration inside the uniaxial movement stage 6,
As shown in FIG. 1B as viewed from the moving direction of the uniaxial movement stage, a linear light source 2 composed of a DC-lit fluorescent lamp is arranged to illuminate the sample 7 from below. FIG.
As shown in FIG. 7, the light illuminated by the linear light source 2 and transmitted through the sample 7 is focused on the lens 1a of the CCD line sensor camera so that the parallel component forms an image on the CCD line sensor camera 1 by the lens 3. Is located in the position. The CCD line sensor camera 1, as shown in FIG. 1 (b) is supported by a holding strut 8 for supporting the camera, lens, light source, to rotate the theta ₁ direction in FIG sample surface as a fulcrum by the elevation angle varying device 14 Thus, the camera can change the elevation angle at which the camera looks at the sample. Also,
The camera is configured to be slidable in the main scanning direction (the y-axis direction in the figure) by the camera slide device 13 and controlled by the camera controller and the signal processing device 9 to process the read signals and detect defects. The rotary stage 5, the one-axis moving stage 6, and the elevation angle varying device 14 are driven and controlled by a controller 10. Overall control.

Next, a method of inspecting a periodic pattern for surface defects will be described. The obliquely transmitted light of the sample moving in the x-axis direction by the uniaxial movement stage 6 is captured line by line by the CCD line sensor camera 1, the luminance distribution on the scanning line is converted into an electric signal, and input to the signal processing device 9. in this case,
When a periodic pattern is photographed with a CCD sensor, moire generated between the pattern and the pixel pitch of the CCD becomes a false defect signal and causes a reduction in S / N. It is preferable to do so. In addition, even if the defect has the same direction and size, the angle of light incident on the camera changes depending on whether it is located at the center or the end of the field of view of the camera, so that the detection sensitivity is not uniform within the same field of view. Therefore, in the present invention, the camera lens 1a is arranged at the focal position of the lens 3 as shown in FIG. 2, so that only the parallel component of the light transmitted through the sample 7 is guided to the camera lens, and the photographing angle in the visual field Is unified.

The optimum elevation angle at which the camera 1 looks at the sample is shown in FIG.
As shown in FIG. 3A, light is blocked in a normal hole 21 as shown by an arrow in the figure, and light is transmitted without blocking in a hole 23 having a defective portion 22 as shown in FIG. 3B. It is such an angle. When the shapes of the holes 31, 32, and 33 are changed in the same sample as in the shadow mask shown in FIG. 4 and the optimum elevation angle is changed, the elevation angle in FIG. The holding column 2 together with the camera 1, the lens 3 and the light source 2 so that the elevation angle is always optimized by the variable device 14.
1 is controlled by rotating it in the θ ₁ direction about an axis in the same plane as the sample. This control is performed by obtaining an optimum elevation angle in advance for each hole, and controlling the elevation angle through the controller 10 by program control by the control computer 11 so as to achieve the obtained optimum elevation angle.

The image signal for one line sequentially taken in from the camera in this manner is A / D-converted by the signal processing device 9 shown in FIG. 1A, and the defective portion is emphasized by filtering processing such as secondary differentiation. Then, the offset level (DC level) is made constant, this signal is compared with a preset slice level, and a protruding signal is recorded as a defect signal. That is, as shown in FIG. 5A, the original image obtained by the A / D conversion has a form in which a signal is placed on a normally fluctuating offset level. 5B, the offset is made almost 0 as shown in FIG. 5B, and the signal protruding at the center by the slice level S is detected as a defect. Such processing makes it possible to accurately detect defects without being affected by noise.

The filter pattern used at this time is shown in FIG.
The pattern of the second derivative as shown in FIG. 6A is good, and the size is changed by changing the distance between the elements as shown in FIG. 6B depending on the area of the defect formed on the line sensor. A series of operations from photographing to defect detection are performed in real time while photographing.

When the photographing width of the camera is smaller than the width of the sample and the image data of the entire sample width cannot be captured at one time, the apparatus having the structure shown in FIG. After the inspection, the camera is slid in the lateral direction (y direction), and the inspection operation is repeated until the inspection of the entire surface of the sample is completed. In this case, time is required for the inspection. Therefore, when speeding up is desired, FIG.
As shown in FIG.
Inspection may be performed by arranging a plurality of cameras in a horizontal row in the y direction as shown in 2, 1-3 so that the entire surface of the sample 7 can be read at a time.

[0017] Then after finishing examined from a certain direction of the sample over the entire surface, the rotary stage 5 of FIG. 1 is rotated in the theta ₂ direction in FIG. 1 (c), inspect the sample from different directions, the required direction of this operation Repeat as many times. In the case of inspecting a plurality of samples of the same type, the optimum elevation angle may be determined in advance, and the inspection may be performed according to the determined optimum elevation angle at the time of inspection.

Next, an optimum elevation angle setting method will be described with reference to FIGS. FIG. 8 shows the average value of the transmitted light taken while changing the elevation angle on the horizontal axis and the elevation angle on the vertical axis in fixed steps, and shows how the amount of light transmitted through the sample increases as the elevation angle increases. . The optimum elevation angle corresponds to the rising point b in FIG. To find the point b in automatic, the average value data a _x at a certain elevation, the average value data a _{x + d} which exists in the direction in which the elevation angle is increased with respect to a _x (d ≧
It is better to look for the rise of the difference value from 1).
When noise is present as shown in FIG.
(B), an erroneous portion may be detected as rising. Therefore, if a certain slice level s is determined as shown in FIG. 9B and the first point a at which the slice level exceeds the slice level continuously for a predetermined number of times or more is set as the optimum elevation angle, the rising point without the influence of noise is obtained. You can look for.

As shown in FIG. 10, lines L _{1 to} L _n perpendicular to the sub-scanning direction for obtaining the optimum elevation angle in the sample plane are determined, and the optimum elevation angle is obtained on each line by the above-described method. May be obtained by interpolation using linear interpolation or the like.

When inspecting a large number of samples, it takes time to measure the optimum elevation angle at a plurality of points according to the above-described method.
The required takt time may not be obtained. In the case of a sample that has the same tendency in how the optimum elevation angle changes at a low surface, the measurement time of the optimal elevation angle may be reduced as follows. For any one sample in advance, L in FIG.
Measure the optimal elevation angles b _{1 to} b _n on each line of _{1 to} L _n ,
Record as reference data. In determining any other optimal elevation measures the optimum elevation c _m of any one line L _m of L ₁ ~L _n in the plane of the sample, on the same line in the c _m and the reference data From the optimal elevation angle b _m of d = c _m −
The value obtained by adding b _m for all the reference data b ₁ + d~b _n +
If d is the optimum elevation angle, one time measurement of the optimum elevation angle is sufficient, so that the time is greatly reduced.

A method for setting a slice level for inspection will be described. Scan the sample at the set optimal elevation angle, find the maximum value of the input signal or its differential signal on the line perpendicular to the sub-scanning direction determined in the same way as in the case of the optimal elevation angle measurement, and determine the maximum value or the maximum value The value obtained by adding the offset is used as the slice level used when inspecting the area between the lines. If the line for which the maximum value is to be found just falls on a defective portion, only the maximum value of the line is larger than the adjacent line, so the maximum value of a certain line is larger than the adjacent line by a predetermined size. It may be determined that there is a defect at the time when the level is large, or the slice level may be replaced by analogy calculation or the like from the maximum value of the adjacent line.

As shown in FIG. 11, if there is a portion having a hole in the sample and a portion having no hole as shown in FIG.
When this waveform is secondarily differentiated, signals a and b at the boundary between the perforated portion and the non-perforated portion due to the difference in brightness between the perforated portion and the non-perforated portion are obtained as shown in FIG. And may be detected together with the original defect signal c. In order to recognize the boundary between the perforated portion and the non-perforated portion, the difference between the primary differential waveforms of the defect signal and the boundary signal as shown in FIG.
Specifically, the first derivative waveform is compared with a predetermined positive / negative slice level, and a signal having a positive / negative polarity is detected in the vicinity of a signal exceeding the slice level if positive (lower than the slice level if negative). If not, the location may be recognized as a boundary.

Further, a method of placing a lens in front of the camera lens as shown in FIG. 2 has been described as a means for guiding the parallel component of the transmitted light through the sample to the camera lens, but as shown in FIG. Only the parallel component of the light from the light source 2 passing through the imaging line 7 may be guided to the camera 1 by the concave mirror 4.

In the above embodiment, an example in which a line sensor camera is used as a photographing device has been described. However, if a contact type linear sensor used in a facsimile, a copying machine, or the like is used, space can be saved. In the above embodiment, the sample is moved by the single-axis moving stage and rotated by the rotary stage. However, the camera, the lens, and the light source are mounted on the robot, the sample is fixed, and the imaging system performs the single-axis operation and the rotation operation. You may do it.

[0025]

As described above, according to the present invention, oblique light inspection can be automated, inspection accuracy can be improved, and the possibility of fully automatic inspection can be improved.

[Brief description of the drawings]

FIG. 1 is a diagram showing one embodiment of a surface defect inspection apparatus of the present invention.

FIG. 2 is a diagram illustrating an optical system that detects a parallel component of transmitted light.

FIG. 3 is a diagram illustrating a method of obtaining an optimum elevation angle.

FIG. 4 is an explanatory diagram in a case where an optimum elevation angle changes within the same visual field.

FIG. 5 is a diagram illustrating a defect detection method from which noise has been removed.

FIG. 6 is a diagram illustrating a configuration example of a filter.

FIG. 7 is an explanatory diagram when a plurality of cameras are used.

FIG. 8 is a diagram illustrating an optimum elevation angle setting method.

FIG. 9 is a diagram for explaining an optimal elevation angle setting method.

FIG. 10 is a diagram showing a projection line on a sample.

FIG. 11 is a diagram illustrating a method of removing a signal that appears due to a difference in brightness between a perforated portion and a non-perforated portion.

FIG. 12 is a diagram illustrating a method of removing a signal that appears due to a difference in brightness between a perforated portion and a non-perforated portion.

FIG. 13 is a diagram illustrating an optical system that detects a parallel component of transmitted light using a concave mirror.

[Explanation of symbols]

DESCRIPTION OF SYMBOLS 1 ... CCD line sensor camera, 2 ... Linear light source, 3 ... Lens, 5 ... Rotating stage, 6 ... Single-axis moving stage, 7 ...
Sample 8, holding column 9, camera controller and signal processing device 10, controller 11, control computer 13, camera slide device 14, elevation angle variable device.

──────────────────────────────────────────────────続 き Continued on the front page (58) Fields surveyed (Int. Cl. ⁷ , DB name) G01B 11/00-11/30 102 G01M 11/00-11/08 G01N 21/84-21/958

Claims (6)

(57) [Claims] 1. A method of illuminating a sample with a light source, detecting transmitted light in an oblique direction by a linear area photographing means, and detecting a defect in a periodic pattern by photographing a linear area. A method for inspecting a surface defect of a periodic pattern, wherein an elevation angle exceeding a slice level first when a predetermined slice level of an average value change amount of a sensor incident light amount continuously exceeds a predetermined number of times is set as an optimum elevation angle. 2. A method of illuminating a sample with a light source, detecting transmitted light in an oblique direction by a linear region photographing means, and photographing a linear region to detect a defect in a periodic pattern. When the optimum elevation angle of the entire sample measured for one sample is used as the reference data, and when the optimum elevation angle of another sample is obtained, the optimum elevation angle is measured at only one arbitrary position, and the measurement result and the same position in the reference data are determined. A method for inspecting a surface defect of a periodic pattern, wherein data obtained by adding / subtracting a difference from an optimum elevation angle to all reference data is used as an optimum elevation angle of the sample. 3. A method for illuminating a sample with a light source, detecting transmitted light in an oblique direction by a linear region photographing means, and photographing a linear region to detect a defect in a periodic pattern. While scanning the sample at the elevation angle, the maximum value of the input signal or its differential signal is obtained at any of a plurality of measurement points, and the maximum value or a value obtained by adding or subtracting a predetermined offset to or from the maximum value is measured around each measurement point. A surface defect inspection method for a periodic pattern, wherein the inspection slice level is used. 4. A method for illuminating a sample with a light source, detecting transmitted light in an oblique direction by a linear area photographing means, and photographing a linear area to detect a defect in a periodic pattern. A method for inspecting a surface defect of a periodic pattern, wherein a boundary of a perforated portion is identified from a shape of a first-order differential waveform. 5. The surface defect inspection method according to claim 1, wherein the imaging of the obliquely transmitted light by the linear area imaging means is performed with the focus blurred. 6. The imaging of obliquely transmitted light by the linear area photographing means is performed by rotating the specimen or the photographing means so as to photograph the specimen from each direction. The surface defect inspection method according to claim 1.