JPH0115001B2 - - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to an inspection method using a laser beam diffraction pattern spatial frequency filtering method for detecting defects in a regular pattern using the diffraction phenomenon of light, particularly when a plurality of patterns are inspected. The present invention relates to a method for detecting common defects in a plurality of patterns to be inspected, each of which has unit apertures regularly arranged from bright spots appearing on an inverse Fourier transform surface.

In recent years, a laser beam pattern spatial frequency filtering method (hereinafter simply referred to as the spatial filtering method) using laser light has been developed as a method for detecting defects in regular patterns such as those of various metal filters and IC masks. It is starting to be used. This spatial filtering method consists of 2
The conventional method for processing dimensional images, in which a two-dimensional image is scanned with a small light beam and the resulting output signal is processed by an electronic computer, has problems such as complexity of processing, long processing time, and high cost. It has the advantage of being problem-free, allowing for simple spatial parallel processing of two-dimensional images, and allowing high-speed inspection with an inexpensive optical system. For this reason, it is mainly used as a defect inspection device for industrial products having regular two-dimensional patterns, such as metal meshes and IC masks.

The basic optical system configuration of this type of defect inspection equipment is
As shown in the figure. That is, laser oscillator 1
The laser beam 2 emitted from the laser beam 2 is expanded by a collimator 3, becomes a parallel beam 4, and hits an object to be inspected 5. The inspected object 5 is placed on the front focal plane of the Fourier transform lens 7, and the Fourier transform spectrum of the inspected object 5 appears on the back focal plane due to the light diffracted when passing through the inspected object. . Further, a spatial filter 8 (for example, a negative photographic film recording the intensity distribution of the Fourier transform spectrum, a directional high-cut spatial filter, etc.) is disposed on the back focal plane of the Fourier transform lens 7. Of the Fourier transformed spectra, only the spectra corresponding to normal patterns are absorbed by the spatial filter 8, and the spectra corresponding to defective patterns are transmitted.

Here, since the spatial filter 8 is arranged on the front focal plane of the inverse Fourier transform lens 10, the light 9 transmitted through the spatial filter 8 is inversely Fourier transformed by the inverse Fourier transform lens 10, and the light 9 is inversely Fourier transformed by the inverse Fourier transform lens 10. 10 appears as a spatially filtered inverse Fourier transform image, that is, an image of only the defective portion of the object 5 to be inspected. This is detected by the photodetector 11. A photodetector 11 is placed on the back focal plane of the inverse Fourier transform lens.
If a screen is placed instead, the defective area will appear as a bright spot on the screen, making it possible to visually recognize the defect.

The pattern to be inspected in this defect inspection apparatus is, for example, as shown in FIG. 2, in which unit rectangular openings P are regularly arranged, and the defects in this pattern to be inspected are as shown in FIGS. There is something like that.

Therefore, it is necessary to reliably and quickly detect these defects in the pattern to be inspected.

Hitherto, several proposals have been made by the applicant of this patent as this type of defect inspection apparatus (for example, Japanese Patent Application Laid-open Nos. 117106-1982 and 133830-1982).

However, since these defect inspection devices inspect each pattern one by one, they detect essential defects in the manufacturing process, such as defects in the original mask, defects in the exposure equipment, defects in the etching equipment, and other defects that result in the same type of defective product occurring one after another. cannot be detected.

The present invention has been made in view of the above-mentioned points. First, the first object to be inspected is moved in the X direction and the Y direction within a plane perpendicular to the optical axis to detect the address and detect defects on the entire surface of the object to be inspected. Scanning is performed, and then defect detection scanning is performed in the same manner for the second and subsequent inspection objects, and 3× centered on the defect coordinates of the first pattern is
This provides a method for detecting common defects by comparing the coordinate range of No. 3 with the patterns of the second and subsequent sheets.

An embodiment of the present invention will be described below with reference to FIGS. 4 to 11.

FIG. 4 shows an example of the configuration of an apparatus for carrying out the present invention, and this apparatus is used in combination with the optical system shown in FIG. 1.

In this figure, 12 is a half mirror, which extracts a part of the output light of the laser oscillator 1 and supplies it to the control signal generation unit 13 to detect fluctuations in laser intensity, and performs inverse Fourier transformation using a correction signal corresponding to this fluctuation. The output of the defect detection unit 15 optically coupled to the surface photodetector 11 is corrected.

The output of the defect unit 15 is sent to the defect recognition system 1.
6 is given. The defect recognition system 16 is also provided with an address signal from the XY coordinate signal generating unit 14. This address signal is given every time the object to be inspected 5 moves a predetermined distance in accordance with the movement in the X direction and the Y direction within a plane perpendicular to the optical axis.

As a result, the defect recognition system 16 associates the defect detection signal from the defect detection unit 15 with the address signal from the XY coordinate signal generation unit 14 to determine the defect position coordinates, and, together with the input information from the keyboard 17, sends the defect detection signal to the printer 18, for example. Print output can be sent via .

FIG. 5 shows the signal processing system having the configuration shown in FIG. 4 in more detail. First, the control signal generation unit 13 for correcting laser intensity fluctuation is composed of a laser intensity fluctuation correction signal generation circuit 13a and a control signal generation section 13b, and when the intensity of the laser beam changes, it generates a laser intensity fluctuation correction signal. The output of the generation circuit 13a changes, and based on this, the control signal generation section 13b forms a control signal and supplies it to the defect discrimination circuit 15b of the defect detection unit.

The defect determination circuit 15b provides the data interface 103 of the defect recognition system 16 with defect data obtained by determining the level of the signal from the photoelectric converter 15a optically coupled to the photodetector 11.

In the defect recognition system 16, an address signal from the XY coordinate signal generation unit 14 is applied to an address interface 101, and an
The XY stage interface 102 exchanges signals for scanning the object 5 to be inspected via the Y stage controller 19 and the XY stage driver 20 . Signals from the address interface 101, the XY stage interface 102, and the data interface 103 are sent to the CPU 105 via the bus line 104.
Provided to RAM 107, programs in ROM 106 and console interface 108
Calculations are performed based on inputs from the switch 109, keyboard 110, etc., and the output is outputted via the console interface 108 and printer 111.

FIG. 6 shows an example of scanning of the object 5 to be inspected using the apparatus of the present invention. This scanning is ultimately understood as a relative movement between the object to be inspected 5 and the photodetector 11, and the photodetector 11 is configured as an array in which a predetermined number of unit detectors are linearly arranged. Defect detection scanning is performed by moving the inspection object 5 relative to the inspection object 5 in a direction perpendicular to its longitudinal direction.

If the length of the array is L, then the photodetector 11 is
After scanning once in the direction, the photodetector 11 is moved in the Y direction by L to perform the next X direction scan, and the photodetector 11 is further moved in the Y direction by L to perform the X direction scan. By repeating this process, the entire surface of the inspection object 5 can be scanned. In this case, it is reasonable to perform scanning in the X direction by alternately repeating scanning in one direction and the other direction as shown in the figure. The amount of movement of the photodetector 11 in the X direction and the Y direction is detected by a linear encoder or the like, and this detection signal is used as an address signal.

FIG. 7 shows an example of the configuration of a position detecting section for detecting and scanning only a predetermined surface of the object to be inspected when performing the scanning shown in FIG. 6. in this case,
A marker M is provided at a predetermined portion of the object to be inspected 5, and this marker M is connected to sensors S ₁ , S ₂ and S ₃ , S fixed in a predetermined positional relationship with respect to the photodetector 11. ₄ , the object to be inspected 5 is fed in the X direction and the Y direction. sensor
The range between S ₁ and S ₂ is the drive system acceleration range, and S ₃ ,
The same is true for _S4 . The reason why there are two acceleration ranges is because the object to be inspected 5 is fed in two directions in the X direction.

As shown in FIG. 6, feeding in the X direction and Y direction is repeated, and during feeding in the X direction, the surface to be inspected of the object to be inspected 5 is scanned for defect detection by the photodetector 5. When this scanning is performed a predetermined number of times, that is, the number of times covering the entire surface of the object 5 to be inspected, it returns to the setting position.

These operations are performed by the defect recognition system 1 in Fig. 5.
6, the CPU 105 at address interface 1 from the XY coordinate signal generation unit 14
Address signal given through 01, ROM1
06 and the signals formed based on the memory contents of the RAM 107 are sent to the XY stage interface 102.
via the X-Y stage controller 19,
This is done by applying it to the Y stage driver 20.

Now, when the switch 109 of the defect recognition system 16 is operated, a start signal is generated and the stage on which the inspected object 5 is mounted is moved from the setting position in the X direction from the system 16 via the XY stage controller 19 and the XY stage driver 20. Start moving in one direction. When the marker M is detected by the sensor _S1 , this detection signal is given to the system 16 via, for example, the XY coordinate signal generation unit 14, and based on this, the system 16 outputs the stage driver 2
0 to accelerate the stage. Next, at the same time as the marker M is detected by the sensor _S2 , an address counter reset pulse is given to the system 16 from the XY coordinate signal generation unit 14, and after the count contents of the address counter (not shown) are cleared, the address is reset. Signal counting begins. From now on, the inspection object 5 is scanned for defect detection until the marker M is detected by the sensor _S3 , and the resulting defect data is sent to the system 16, which uses the data from the XY coordinate generation unit 14 through interrupt processing. Defect position data is calculated and written to the RAM 107.

When the marker M is detected by the sensor _S3 , the defect detection scan ends, and the object 5 to be inspected is then sent in the X direction until the marker M passes the sensor _S4 . When the marker M passes the sensor _S4 , the system 16 uses the XY stage controller 19 and the XY stage driver 20 to move the stage by one pitch (L) in the Y direction, and then moves the stage in the opposite direction to the X direction. move it. In this case, marker M is first detected by sensor S ₄ and then sequentially by sensors S ₃ , S ₂ and S ₁ . Then, the area between sensors S ₄ and S ₃ becomes an acceleration area, and the area between S ₃ and S ₂ becomes the inspection target movement, and when it passes the position of S ₄ , it moves in the Y direction again, and then
A forward feed in the direction is performed.

After such stage feeding, when the entire surface of the inspection object 5 is scanned, a scan stop signal is given from the system 16 to the XY stage controller 19 and the XY stage driver 20, and the defect detection scan is completed. At the same time, the stage is returned to its setting position.

The coordinates written together with the defect data by this defect detection scan can be expressed as follows. In other words, a) Number of apertures of the photodetector; N B) Number of apertures of the photodetector; n (1, 2, 3...
N) C) Length in the Y direction of one aperture of the photodetector; l D) Number of scans; s (1, 2, 3...) When the defect coordinates (x, y) are [x, l (sN- n)], and the Y-direction scanning amount, that is, the Y-direction feed amount L, is represented as L=N·l.

8 to 9 are flowcharts showing the steps of the method of the present invention carried out using the above-mentioned apparatus, and the following description will be made according to this flowchart.

First, the apparatus is started and a defect signal of the first pattern is detected (S ₁ ). When detecting this defective signal, address signals from the X and Y stages are detected (S ₂ ). The defect signal and address signal are then stored in the RAM 107 (FIG. 5) in association with each other (S ₃ ).

Next, a second pattern is set in the device and a defect signal of the second pattern is detected (S ₄ ). When this defective signal is detected, address signals from the X and Y stages are detected (S ₅ ). Then, the defect signal and address signal are related to each other in the RAM 107 (Fig. 5).
(S ₆ ).

The defect data of the first pattern and the defect data of the second pattern stored in the RAM 107 (FIG. 5) are compared by the CPU 105 (FIG. 5) to determine whether there is a common defect. This determination is made as to whether or not the defect coordinates of the second pattern exist within a 3×3 coordinate range centered on the defect coordinates of the first pattern (S ₇ ).

This increases the tolerance for minute variations in defect positions between objects to be inspected and the accuracy of the mechanical detection position of the device, making it possible to prevent common defects from being overlooked. If the defect coordinates corresponding to the second pattern do not exist, the defect coordinates are canceled (S ₈ ), and if they do exist, the common defect coordinates are output (S ₉ ). This output is performed, for example, by the printer 111 (FIG. 5).

FIG. 9 shows the steps S ₁ to _{S 3} (S ₄ to _{S 6} ) of detecting and storing defect signals and address signals in FIG.
In particular, step S ₃ (S ₆ ) is shown in detail and will be explained.

That is, after detecting a defective signal (S ₁ or S ₄ ) and detecting an address signal (S ₂ or S ₅ ), it is checked whether the defective addresses are consecutive (S ₁₁ ).
do. First, it is determined whether or not it is continuous (S ₁₂ )
If the address is not consecutive, the defective address is stored ( _S13 ). When storing, it is determined whether the memory is over (G ₁₄ ), and if the memory is over, the message is transferred (S ₁₅ ); otherwise, the process moves to the next step in FIG.

On the other hand, if it is continuous, calculate the area specification line (S ₁₆ )
After dividing the pattern portion of the object to be inspected into regions, it is determined whether a defect is within the region designation line (S ₁₇ ). If it is within the area, the defective address is stored (S ₁₈ ).

If it is outside the inspection target area, it is first determined whether the number of consecutive addresses in the Y direction is greater than or equal to a predetermined number ED (Y) (S ₁₉ ), and if it is, the defective address is canceled (S ₂₀ ). Next, the same procedure is adopted for the X direction. As a result, a defect address is stored only when there are abnormally consecutive defects even outside the inspection target area (S ₁₈ ).

FIG. 10 shows in more detail the procedure for checking the continuity of defective addresses in FIG. 9. First, it is determined whether the defect is at the edge of the defect detection scanning range (S ₁₁₁ ), and if it is not, the continuity in the Y direction is checked (S ₁₁₄ ), and then the continuity in the X direction is checked (S 114). Check (S ₁₁₅ ). On the other hand, at the end of the defect detection scanning range, consideration must be given to the possibility that defects are continuous between the adjacent defect detection scanning range. Therefore,
The end defect address is memorized ( _S112 ), and the end of the next defect detection scan is checked ( _S113 ) to confirm its continuity.

By detecting this defect continuity, the inspection range can be expanded to the vicinity of the periphery of the pattern of the object to be inspected.

FIGS. 11a and 11b show examples of the object 5 to be inspected by the method of the present invention, where a is an IC mask and FIG. 11b is a shadow mask. In the case of an IC mask, the inspection object pattern A is circular, and a square area designation line B indicated by a broken line is set inside this circle. In the case of a shadow mask, the pattern portion A has the shape of a CRT image plane, and a rectangular area designation line B is set inside the pattern portion A.

As described above, the present invention allows a first object to be inspected to be inspected in the X direction within a plane perpendicular to the optical axis in an optical system using a laser beam diffraction pattern spatial frequency filtering method.
While moving in the Y direction and detecting the address, defect detection scanning is performed on the entire surface of the object to be inspected, and then defect detection scanning is performed in the same way on the second and subsequent objects to be inspected, and the defect coordinates of the first pattern are centered. Since common defects are detected by comparing the 3×3 coordinate range with the second and subsequent patterns, it is possible to limit the detection target and perform common defect inspection efficiently with fewer oversights. Common defects are highly likely to develop into serious problems such as rotout, and being able to reliably detect common defects is extremely advantageous in terms of manufacturing technology.

[Brief explanation of drawings]

Fig. 1 is a diagram showing the optical system configuration of a defect inspection device using the spatial filtering spatial frequency method used in the present invention, and Fig. 2 is a diagram showing a pattern in which unit rectangular openings are regularly arranged as an example of the inspection target. ,
3A to 3D are diagrams showing examples of defects in the pattern of FIG. 2, FIG. 4 is a diagram illustrating a defect signal and scanning signal detection system of the apparatus according to the present invention, and FIG. Figures 6 and 7 showing one embodiment
The figure is an explanatory diagram of the scanning operation for detecting defects in the above embodiment, FIG. 8 is a flowchart showing the procedure of the method of the present invention, and FIGS. A flowchart showing possible procedures, and FIGS. 11a and 11b are explanatory drawings showing examples of objects to be inspected by the method of the present invention. 1... Laser oscillator, 3... Collimator, 5...
...Object to be inspected, 8...Spatial frequency filter, 11...
...Optical inspection device, 15... Defect detection unit, 16...
...defect recognition system. S...Sensor, M...Marker.

Claims (1)

[Scope of Claims] 1. A pattern in which unit apertures are regularly arranged is placed in an optical system using a laser beam diffraction pattern spatial frequency filtering method to obtain an inverse Fourier transform image, and this transform image is used to determine the shape of the pattern. In the method of detecting defects, the first
The object to be inspected is moved in the X and Y directions within a plane perpendicular to the optical axis of the optical system, and the entire surface of the object to be inspected is scanned for defects while detecting the address. The second inspected object is scanned for defect detection in the same manner as the first inspected object, and the second 1. A method for detecting common defects in a regular pattern, characterized in that common defects are detected by comparing the inspection results of an object to be inspected. 2. In the method described in claim 1,
The predetermined coordinate range is a regular pattern common defect detection method, wherein the predetermined coordinate range is a 3×3 coordinate range. 3. In the method according to claim 1 or 2, the defect detection scan of the first or second object to be inspected is performed to detect pattern surrounding information and defect information by detecting continuity of defects. A common defect detection method for regular patterns that enables the identification of