JPH0690140B2 - Optical surface inspection method - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION [Industrial field of application] The present invention scans the surface of a sample having a mirror-like surface such as a substrate for semiconductors and a substrate for magnetic disks with a light beam to cause unevenness of the sample surface and adhesion of foreign matter. The present invention relates to a method for optically inspecting the presence or absence of these irregularities, foreign matter, etc. by receiving scattered light due to.

[Conventional technology]

FIG. 5 shows a general configuration of an apparatus for carrying out this type of optical inspection method. The light source 1 is a light source that projects a thinly focused light beam like a laser light source. The projected light beam is a beam expander 2 and a rotating polygon mirror 3.
The sample 5 is scanned through the f.theta. Lens 4 to form a light beam 6 which is focused on the sample 5. Beam expander 2
Adjusts the incident thin light beam to a light beam with a width of about 5 mm,
The rotating polygon mirror 3 rotates its beam about its axis,
The f · θ lens 4 is a lens in which the rotational angular velocity of the incident light beam and the scanning velocity are in a proportional relationship, and the rotating light beam is linearly formed on the surface of the sample 5 at a constant speed while always keeping the focused state on the surface of the sample 5. To scan. If there is an irregular portion 7 such as unevenness or foreign matter on the sample 5, scattered light 8 is generated there. The scattered light 8 is detected by the light detection means 9 surrounded by the alternate long and short dash line. In the light rendering means 9, the scattered light 8 that is incident by the incident light condensing system 10 having a cylindrical lens on the light incident surface becomes almost parallel light and is incident on the optical fiber bundle 11, and is emitted and condensed by the light condensing system 12 and the band bus. After passing through the filter 13, it is detected by a photodetector 14 such as a photomultiplier tube and given as a pulse component of an output signal 16 from an amplifier 15 connected to the photodetector 14. The bandpass filter 13 is for blocking light having a wavelength other than the wavelength of the light projected from the light source 1 to improve the S / N ratio of the incident light. This pulse component 16 is subjected to counting or peak value measurement in the measuring circuit 17,
Analysis of the number and size of foreign matter is performed. The sample 5 is a sample transport mechanism that constitutes a relative moving means 18 surrounded by a two-dot chain line.
It is supported by a conveyor belt 20 provided in 19 and moves relative to the light projecting means in a direction perpendicular to the scanning direction of the light beam 6, that is, a direction perpendicular to the paper surface. This allows the light beam 6 to scan the entire surface of the sample 5 for inspection.

Since the surface of sample 5 is mirror-finished, sample 5
The light beam that illuminates
It is projected at an angle not incident on 10. Therefore, only the weak scattered light 8 scattered at the abnormal portion 7 on the surface of the sample 5 enters the photodetector 14. The particle size of fine particles to be inspected by this type of inspection apparatus is about 0.1 μm to 10 μm. For example, a silicon substrate for semiconductor manufacturing has a diameter of 150 mm.
Several to several tens of fine particles adhere to the substrate of No. 1, and most of them have a particle diameter of 1 μm or less. The scattered light from such fine particles is extremely weak, but the photomultiplier tube as the photodetector 14 has sufficient sensitivity to detect these scattered lights.

[Problems to be solved by the invention]

FIG. 6 is a plan view showing the sample 5 and one scanning line 23 of the light beam 6 for scanning the sample 5 as seen from above.

Output signal when particles 7a, 7b, 7c adhere to scanning line 23
16 produces pulse components corresponding to these.

The inspection device totalizes and outputs the pulse components detected as described above over the entire surface of the sample 5. At this time, the peripheral portion 24 partitioned by the broken line on the sample 5 is an area which is not used as a substrate because the peripheral portion 24 is often contaminated by contact with these cases when the sample 5 is stored or transported. Therefore, the pulse components obtained from the fine particles 7b and 7c in the peripheral portion 24 are excluded from the counting, and the inspection region of the communication unit is excluded.
For 25, only the pulse component due to the fine particles 7a must be output.

For this reason, in the conventional inspection method, the surface of the sample 5 is divided into a number of areas, coordinates are determined, and the areas and coordinates in which the pulse components are obtained are stored, and after the entire scanning is completed, the peripheral portion is A method other than the one corresponding to 24 is adopted, or the mutual positional relationship among the light beam 6, the sample transport mechanism 19, and the sample 5 is determined in advance, and the synchronization signal generator 21 linked with the sample transport mechanism 19 is used. A method of excluding the count in the peripheral portion 24 based on the output synchronization signal is also used. Therefore, an advanced signal processing system such as a high-speed storage circuit having a large storage capacity and a high-speed arithmetic processing circuit, a power supply circuit for the signal processing system, and the like are required, and the device becomes large and expensive.

Further, since the conventional inspection apparatus is made on the premise that the sample carried by the human is measured, the sample 5 mounted with the uncertainty of the installation position by the human is placed between the sample 5 and the light beam 6 relative to each other. Sample transport mechanism with precisely defined position
An intermediate mechanism for reattaching to 19 is built into the inspection device. In these intermediate mechanisms, for example, a cassette holding a sample at a predetermined position by hand is used to successively transport the sample transport mechanism 19
It is a complicated mechanism such as a cassette elevator that transfers the sample to the sample transfer mechanism 19 and accurately positions it on the sample transfer mechanism 19. Further, as the sample transfer mechanism 19, there are an XY stage, a rotary stage for arranging a plurality of samples 5 on the circumference around an axis for rotational movement, and a parallel belt type transfer device for moving the sample 5 in one direction. Both of these are also large and expensive, since they are provided with means for precisely defining the zones and coordinates on the sample 5 to be transported, as previously described.

From the above, the conventional inspection device is large and expensive in both the signal processing system and the transfer mechanism, and is not suitable for being directly incorporated in the manufacturing line. Therefore, the sample must be manually carried to the inspection apparatus or installed in the apparatus, and at this stage, there is a problem that contamination due to adhesion of fine particles to the sample cannot be avoided.

The present invention provides an inexpensive, compact and simple inspection apparatus capable of counting abnormal points in a predetermined region of a sample by a simple signal processing mechanism and a simple moving mechanism, except for the above problems. With the goal.

[Means for solving problems]

The present invention is such that the intensities of scattered light from the boundary of the circular sample, the region outside the circular sample, and the surface of the circular sample are different from each other, and the relative number of scanning of the light beam relative to the light beam of the sample in the direction intersecting the scanning direction. That is, the scanning time of the light beam is proportional to the movement distance of the light beam on the scanning line, that is, the position where the light beam starts to enter the sample, that is, the edge of the sample. Is detected in two directions, the scanning direction and the direction intersecting the scanning direction, and the start and end of the inspection area in the relative movement direction of the sample are given in advance as the number of scans after the edge of the sample is detected. Every other, the inspection area is set by previously giving the start and end of the inspection area in the scanning direction as the elapsed time after detecting the edge of the sample,
The abnormal portion is counted within the inspection area.

That is, the first determination level and the second determination level are set in advance above and below the level of the DC component corresponding to the scattered light from the region other than the sample in the output signal of the light detecting means that receives the scattered light, When a pulse component exceeding the first judgment level is given for the first time after the output signal of the light detecting means gives a DC component exceeding the second judgment level in a certain scanning, the above-mentioned DC component exceeds the second judgment level. And the time difference T between the time point at which the pulse component exceeds the first determination level is obtained, and the first gate signal is issued at the time point when the number of scans after that time point reaches a predetermined first value. , And from the time when the DC component of the output signal first exceeds the second determination level for each scan after the time when the first gate signal is issued, to the time when the DC component of the output signal falls below the second determination level. Determination level is a time required from a time point falls below to the light beam moves the peripheral portion of the sample, a predetermined time given as a function of the diameter and the number of scans of the sample with the second measuring the time difference t ₁ A _second gate signal is issued when t ₂ has elapsed, and counting of the pulse component of the output signal in the scan is started according to the AND condition between the first gate signal and the second gate signal, while the time difference T And t ₁ and a predetermined time t ₂ from 2 [T−
The value of (t ₁ + t ₂ )] is derived, and the second gate signal is cut off at the time when the above-mentioned 2 [T− (t ₁ + t ₂ )] has elapsed from the time when the second gate signal was issued. The counting of the pulse component in the scanning is ended, and when the number of times of scanning reaches a predetermined second value, the first gate signal is cut off to end the counting in one sample.

[Action]

The first and the second DC components, respectively, corresponding to scattered light from a region other than the sample in the electric signal output from the light detection means are provided above and below, respectively.
If the determination level and the second determination level are set, the DC component of the output signal exceeds the low second determination level at the time when the light beam starts projecting on an area other than the sample.
This makes it possible to know when the projection of the light beam has started. Further, when the sample moves relative to the light beam and the scanning line comes into contact with the edge of the sample, strong scattered light is generated in a pulse shape, so that the electric signal is a pulse exceeding the high first judgment level. Give a component. Therefore, in a certain scan, the pulse signal exceeding the first judgment level is not given until the output signal gives the DC component exceeding the second judgment level, as described above. It corresponds to touching the boundary of the sample. Therefore, when the time difference T between the time when the DC component of the output signal exceeds the second determination level and the time when the pulse component exceeds the first determination level for the first time, the time difference T is obtained, and the scanning speed of the light beam is constant. In the sample, the distance between the projection start point of the light beam and the central axis of the sample can be obtained.

Further, since the relative movement amount of the sample with respect to the light beam is constant in the direction intersecting with the scanning direction between one scanning and the next scanning, the relative movement amount of the sample in that direction can be obtained by the number of scannings after the above point. . Until the number of times of scanning reaches the number of times of scanning of a predetermined first value, the peripheral portion which is not the object of inspection is scanned. When the number of times of scanning reaches the first value, the light beam reaches the inspection area for the first time in the direction intersecting with the scanning direction, and at this time point, the first gate signal is issued to count the abnormal portion of the sample. Will be possible.

On the other hand, the surface of the sample is mirror-finished, and the scattered light from the surface is extremely weak and greatly falls below the level of the scattered light from outside the sample. Therefore, when the time difference t ₁ from the time when the DC component of the output signal exceeds the second judgment level to the time when it falls below the second judgment level is measured, the scanning speed of the light beam is constant and t ₁ is The position of the sample becomes clear corresponding to the distance in the scanning direction from the projection start point of the beam to the boundary of the sample. Since the predetermined time t ₂ given separately is the time for the light beam to pass through the peripheral portion in the scanning direction, the light beam scans when the time t ₂ elapses from the time when the DC component of the output signal falls below the second determination level. Since the inspection area is reached in the direction, the second gate signal is issued at this point, and the counting of the abnormal portion in the scanning can be started.

Therefore, the AND of the first gate signal and the second gate signal
The condition is satisfied only when the light beam passes through the inspection area, and the pulse component of the output signal can be counted under the condition of this AND condition. At this time, only the abnormal portion in the inspection area is counted. Equivalent to.

Further, the value derived as 2 [T- (t ₁ + t ₂ )] is the time when the light beam passes through the inspection area in the scanning direction.
When the second gate signal is turned off after the time [T- (t ₁ + t ₂ )] has elapsed, the above AND condition is not satisfied, and the counting of pulse components in the scan is completed. The number of scans of the second value is the number of scans until the light beam reaches the end of the inspection region in the direction intersecting the scanning direction. Therefore, if the first gate signal is cut off when the number of scans reaches the second value, the AND condition is not satisfied in the subsequent scans, and the counting for the sample is completed.

In this way, it is possible to count abnormal points only in the inspection area without previously defining areas or coordinates on the sample.

〔Example〕

FIG. 1 is a block diagram showing an embodiment for carrying out the method of the present invention. The configuration of the optical system from the light source 1 to the light detecting means 9 is the same as that in FIG. Regarding the relative moving means 31, the sample transport mechanism 32 has an extremely simple structure in which the transport belt 20 is the main component, and the sample 5 moves relative to the scanning direction of the light beam 6 in the vertical direction. .

Output signal output from the amplifier 15 included in the photodetection means 9
The 16 is divided into two, one of which is guided to the pulse shaping circuit 33 and the other of which is guided to the inspection area setting circuit 34 surrounded by a broken line. In the inspection area setting circuit 34, the inspection area is set based on the waveform of the output signal 16 by the x-direction inspection area setting circuit 35 in the scanning direction and the y-direction inspection area setting circuit 36 in the sample moving direction. The gate circuit 3 depends on the AND condition of the second and first gate signals 37 and 38 output when it is determined that the beam is within these set inspection areas.
9 opens, pulse signal 40 shaped by pulse shaping circuit 33
To the counting circuit 41. Inspection area setting circuit 34 and counting circuit
An input / output circuit 42 is connected to 41 to input a constant for setting an inspection area and display a counting result of the counting circuit 41. The display of the counting result is performed after the first gate signal 38 of the y-direction inspection area setting circuit 36 is cut off, and the total value for each sample is shown.

2 and 3 show the output signal 16 in the embodiment of the present invention.
8 is a source diagram showing that the inspection area 25 can be set on the sample 5 from FIG.

FIG. 2 shows the sample 5 and the scanning lines 23a, 23b and 23c at different positions relative to the sample 5 and the output signal 1 for the scanning lines.
The waveforms of 6a, 16b, 16c and the second gate signal 37 are shown respectively.

In the present invention, the x coordinate and the y coordinate are hypothesized on the sample 5 to define the inspection area 25.

Point O at which the light beam 6 begins to be projected from the f / θ lens 4.
Is the same in every scan, this point O is taken as the reference of the x coordinate, and the scanning speed of the light beam 6 is constant, so the distance from the point O depends on the elapsed time after the start of projection of the light beam 6. Establish.

Also the point P ₂ across the edge P ₂ of the light beam 6 is first sample 5 as a reference point of the y-coordinate, the distance in the y direction from the point P ₂ is determined by the number of scans after the detection of the point P _2.

The scanning line 23a is the case where the sample is advancing toward the scanning line 23a, and the output signal 16a for the scanning line 23a will be described. The output signal 16a before the light beam 6 is projected to the point O
Is at the base level 43 which is the noise level of the photodetector 14 and the amplifier 15. When the light beam begins to be projected from the point O, the output signal 16a rises to the level 44 corresponding to the amount of scattered light outside the sample corresponding to the scattered light 20 from the surface of the sample transport mechanism 32 shown in FIG. The scanning line 23a is at the points P ₃ and P ₃ ′ at the conveyor belt
When it crosses 20, the transport belt 20 has a better reflectance than the surface of the sample transport mechanism 32, and therefore specular reflection light that does not enter the incident light collection system 10 is generated, and at that time, the output signal 16a
Is instantaneously reduced to produce a negative going pulse signal 45. When the light beam is no longer projected at the point O ', the output signal drops again to the base level 43.

A first determination level 46 and a second determination level 47 are set above and below the level 44 corresponding to scattered light outside the sample. By doing so, it can be known that the light beam has begun to be projected from the point O because the DC component of the output signal 16a first shifts from the level lower than the second determination level to the level higher than the second determination level. This point is the same for every scan line.

The scanning line 23b shows the case where the edge P _{2 of the} sample 5 contacts the scanning line 23b for the first time. The light beam 6 is scattered at the edge P ₂ of the sample 5 and produces scattered light stronger than the scattered light 22 from the surface of the sample transport mechanism 32. However, since the edge P _{2 of the} sample 5 contacts the scanning line 23b at almost one point, the scattered light is generated instantaneously. Therefore, the output signal 16b for the scanning line 23b is generated when the light beam passes through the point P _2. A pulse signal 48 exceeding the first judgment level 46 is given. The point P ₂ can be known only by obtaining such a signal after the start of scanning. This point
When P ₂ is detected, the time difference T between the time when the DC component of the output signal 16 shifts from the level below the second determination level to the level above and the point P ₂ detection time is obtained. This time point T is the distance between the point O and the center axis of the sample 5 in the y direction. Equivalent to.

Further, after detecting the point P ₂ , the sample 5 moves in the y-axis direction, and the number of scans n ₁ until the operation line 23 crosses the starting point P ₆ of the inspection region 25 in the y-axis direction is the first value. Is given in advance, and when the number of scans reaches n ₁ , the first gate signal 38 is issued to satisfy one of the counting start conditions.

On the other hand, after detecting the point P ₂ , the scanning line passes through the surface of the sample 5. The scanning line 23c shows this state. The DC component of the output signal 16c corresponding to this rises from the level below the second judgment level 47 to the level 44 corresponding to the extra-sample scattered light due to the start of the light beam projection at the point O. When the light beam starts to be projected onto the sample 5, the output signal 16c is instantaneously further increased due to the scattering of light at the edge of the sample 5, and then the amount of scattered light in the sample corresponding to the scattered light 8 from the surface of the sample 5 is equivalent. Reduces to level 49. Since the surface of the sample 5 is smooth, the level 49 corresponding to the amount of scattered light in the sample is extremely low, and therefore the output signal 16
The DC component of c decreases from the level above the second judgment level 47 to the level below. The DC component of the output signal 16c is the second
The time t _{1 at the} level 44 corresponding to the amount of scattered light outside the sample above the judgment level 47 of is the distance on the scanning line 23c. This time t ₁ is measured because it corresponds to. After this time t ₁ , the light beam 6 The passes in duration t ₂ enters the inspection area 25 in the x-direction.
Since this t ₂ can be predetermined as described later, after the time t ₂ has elapsed from the time when the output signal 16c fell below the second judgment level 47, that is, the time t ₁ + t ₂
After the time elapses, the second gate signal 37 is issued so that another one of the counting start conditions is satisfied. That is, the AND condition of the first gate signal 38 and the second gate signal 37 enables counting of the pulse components included in the output signal 16c.

The intersection point P ₄ and the point O between the scanning line 23c and the central axis of the sample 5 Time t ₃ when scanning between the t ₃ = T-become _{(t 1 + t 2) (} 1). P ₄ is a point on the central axis Is. Therefore, the time taken for the light beam 6 to scan the inspection area 25 is given by 2t ₃ , so 2t ₃ = 2 [T− (t ₁
+ T ₂ )], and after that time, the second gate signal 37
Is turned off and the counting in the scan is completed. The output signal 16c includes pulse components 50, 51, 52 due to scattered light from the fine particles 7a, 7b, 7c on the surface of the sample 5, and has a peak value exceeding the second judgment level 47,
The determination of the level fluctuation of the output signal 16 based on the determination level 47 is performed only for the DC component. That is, when the output signal 16 shifts from a level lower than the second judgment level 47 to an upper level or shifts from an upper level to a lower level, the pulse component of the pulse component is changed from the time when the level shifts. The transition detection signal is not issued unless the level after transition is continued for a predetermined time that is one digit longer than the duration. Therefore, the setting of the inspection area is not affected by the above-mentioned pulse components 50, 51 and 52. Further, as shown in FIG. 2, pulse components 50, 52 due to the fine particles 7a, 7c in the peripheral portion 24
Since the second gate signal 37 is off for
These are not counted, and only the pulse component 51 due to the fine particles 7b in the inspection area 25 is counted.

Further, the number of scans n ₂ until the scan line 23 reaches the end P ₇ of the inspection region 25 on the central axis is also set in advance as a second value, and when the number of scans reaches n ₂ , the first gate The signal 38 is turned off to complete the counting for one sample.

As described above, the first gate signal 38 and the second gate signal 37
Since the area where the AND condition of is satisfied corresponds to the inspection area,
If the pulse component of the output signal 16 is measured while the above AND condition is satisfied, the inspection of the abnormal portion 7 in the inspection region 25 can be easily performed without setting the area or coordinates on the sample 5 in advance. It can be carried out.

Further, even if the installation position of the sample 5 is different, t ₃ can be determined from the equation (1) by measuring t ₁ , so it is not necessary to fix the position of the sample 5 in advance to the setting of the inspection area. , The mechanism for that can be omitted.

In setting the peripheral portion 24, the peripheral portion 24 does not necessarily have to have a constant width with respect to the outer periphery of the sample 5, and the shape shown in, for example, FIG. In this case, the peripheral portion 24 has a certain distance d inward from the edge of the sample 5 in the x direction, and the end points P ₂ and P ₈ are also on the central axis in the y direction.
This is the region where the same distance d as above is taken from and.

In this case, t ₂ is given as a constant value the time during which the light beam travels the distance d in the scanning direction, and the first value n ₁ of the number of scans is given the number of times the light beam 6 scans the distance d. . Further to a second value n ₂ number of scans give the number of scans between the distance 2R-d. Here, R is the radius of the sample 5.

On the other hand, when a region having a constant width d from the outer periphery of the sample 5 is given as the peripheral portion 24, the width d ′ of the peripheral portion 24 in the x direction is a function of the radius R of the sample 5 and the vertical distance y between the point P ₂ and the scanning line. Since it can be counted, the transit time of the light beam 6 with respect to the calculated d'is further calculated to be t ₂ .

That is, as revealed by the geometry in Fig. 3 (b) That is, X ² = (y−d) (2R−y−d) (3) (X + d ′) ² = y (2R−y) (4) holds, and X is deleted from the equations (3) and (4). Then the quadratic equation for d ' Is obtained. d'is the positive one of the solution of equation (5) Is.

The relationship between time t ₂ and d ′ is represented by t ₂ = k ₁ d ′ (7) where k ₁ is a constant. The relationship between the distances y, d, 2R-d and the number of scans n, n ₁ , n ₂ for these is k ₂ as a constant. Indicated by. From equations (6), (7) and (8)
t ₂ is Can be calculated as That is, this t ₂ is calculated and given for each scan.

In the above description, the distance in the y direction to the point P _{2 at} the edge of the sample 5 is determined by the number of scans after the point P ₂ is detected. However, since the time interval between one scan and the next scan is always constant, exactly the same result can be obtained even if the time interval is set to an integral multiple of the time interval of the scan. In this case, the number of scans in equation (8)
Take t, t ₄ and t ₅ as the elapsed time corresponding to n, n ₁ and n ₂ , respectively, and let k ₃ be a constant. T ₂ Calculate by

FIG. 4 is a block diagram showing an example of the circuit configuration of the inspection area setting circuit 34.

x-direction inspection area setting circuit 35 and y-direction inspection area setting circuit 36
Are surrounded by broken lines.

The x-direction inspection area setting circuit 35 includes a clock pulse generator 61, and provides a clock pulse 62 which serves as a reference for time measurement for setting the inspection area.

The point P ₂ detection circuit 63, the point O detection circuit 64, and the point P ₁ detection circuit 65 are circuits mainly composed of a comparator, and when the output signal 16 exceeds the first judgment level, respectively, The time when the DC component exceeds the second determination level and the time when the DC component of the output signal falls below the second determination level are detected. At the time of detection, the point P ₂ detection circuit 63 and the point O detection circuit respectively output signals 66,
Outputs 67. Further, the point P ₁ detection circuit is connected with a delay circuit 68 having a delay time of t ₂ given by the equation (7), (8) or (9), and time t ₁ + t ₂ from the point O detection time.
The signal 69 is output after a delay. Since the light beam 6 has reached the inspection area when this signal 69 is output, the signal 69
Causes the x-direction gate 70 to emit a second gate signal 37.

The elapsed time is measured by counter 71, preset down counter 72, and up / down counter 73 clock pulse 62
Is performed by counting.

In the counter 71, the count value corresponding to the scanning time T between the point O and the point P ₂ is obtained by the signals 66 and 67, and this count value is given to the preset down counter 72 as the preset value 74.

After the next scanning, the preset down counter 72 starts counting down with the signal 67 by the point O detection, and raises the signal 75 at the time when the count value becomes “zero”, that is, when the time T has elapsed after the point O detection. Give to the down counter 73.

On the other hand, the up-down counter 73 starts counting by up-counting the clock pulse 62 with a delay of time t ₁ + t ₂ from the point O detection by the signal 69. Thus the signal
The count at the time when 75 is given is a value corresponding to T- (t ₁ + t ₂ ) = t ₃ . The up / down counter 73 starts counting down from the time when the signal 75 is obtained, and outputs the signal 76 at the time when the count value becomes “zero”, that is, when the time 2t ₃ has elapsed after the signal 69 was given. It is applied to the directional gate 70 to turn off the second gate signal 37. In this way, after the time t ₁ + t ₂ has elapsed since the point O was detected, the second gate signal 37 is given for the time 2t ₃ during which the light beam 6 scans the inspection area 25.

The y-direction inspection area setting circuit 36 includes a scanning synchronization signal generator 77, and gives a pulse signal 78 for measuring the number of scans for setting the inspection area 25. The scanning synchronization signal generator 77 may be any one that gives a pulse synchronized with the scanning cycle, for example, a clock pulse generator synchronized with the scanning cycle, a synchronization signal generator synchronized with the rotation of the rotary polygon mirror 3, and a point O. A monostable multivibrator or the like connected to the detection circuit 64 can be used.

The pulse signal 78 is measured by preset down counters 79 and 80. The preset down counter 79 is provided with the number of scans n ₁ corresponding to the peripheral portion 24 as a preset value. When the point P ₂ is detected, the signal 66 is given to start the down counting, and the count value is “zero”. Scan line
When 23 reaches the inspection area 25, the signal 81 is applied to the y-direction gate 82 so as to emit the first gate signal 38.

The signal 81 also starts down-counting of the preset down-counter 80, which is given a preset value of n ₂ −n ₁ at this time, and the time when the count becomes “zero”, that is, the scanning line.
When 23 reaches the end of the inspection area 25, the signal 83 is applied in the y direction and the first gate signal 38 is turned off.

The gate circuit 39 has a first gate signal 38 and a second gate signal 3
7, AND gate of the pulse signal 40 based on the abnormal part 7, and the AND signal is satisfied only when the light beam 6 is in the inspection area 25 and the pulse signal 40 is emitted. To the counting circuit.

〔The invention's effect〕

In the present invention, the intensity of scattered light from the light beam scanning the sample and the region outside the sample is different outside the sample, at the edge of the sample, and inside the sample. The boundary between the sample and the crossing direction is detected, and the scanning speed of the light beam, the scanning period, and the relative movement speed of the sample with respect to the light beam are constant. The scanning area is defined by the elapsed time after the boundary detection of the sample and the moving direction of the sample intersecting the scanning direction is determined by the number of scans after the boundary detection of the sample. The gate that counts the pulse signal due to the scattered light from the abnormal area is opened only when scanning, so the abnormal area in the inspection area can be inspected by a simple procedure. Rukoto can.

Therefore, there is no need for special alignment or synchronization between the optical system and the sample moving mechanism, and the sample can be set at any position. Therefore, the structure of the moving mechanism is simple and there is no need to provide a complicated signal processing circuit, the device is inexpensive, small and simple, and there are few restrictions on incorporation into the automatic inspection line, and the structure of the automatic inspection line is small. Will be easier.

[Brief description of drawings]

FIG. 1 is a block diagram of an embodiment of the method of the present invention, FIG. 2 is a plan view showing scanning lines at different positions with respect to a sample and waveform diagrams of output signals for these scanning lines, and FIG. 3 is a sample. FIG. 4 is a block diagram showing the configuration of an inspection region setting circuit, FIG. 5 is a configuration diagram of an optical inspection device according to the prior art, and FIG. 6 shows a sample and a scanning line. FIG. 5: sample, 6: light beam, 7: abnormal part, 7a, 7b, 7c: fine particles,
8, 22: scattered light, 9: light detecting means, 14: light detector, 23, 23a, 23
b, 23c: scanning line, 37: second gate signal, 38: first gate signal, 39: gate circuit, 40: pulse signal, 41: counting circuit, 4
6: first determination level, 47: second determination level, 48: pulse signal, 50, 51, 52: pulse component.

 ─────────────────────────────────────────────────── ─── Continuation of front page (72) Inventor Hiroshi Hoshikawa 1-1, Tanabe Nitta, Kawasaki-ku, Kawasaki-shi, Kanagawa Fuji Electric Co., Ltd. (72) Keisuke Sugimoto 1-Nitta, Tanabe, Kawasaki-ku, Kawasaki-shi, Kanagawa No. 1 in Fuji Electric Co., Ltd. (56) Reference JP-A-61-132844 (JP, A) US Patent 4376583 (US, A)

Claims (5)

[Claims] 1. A foreign material on the surface of a sample, which is obtained by linearly scanning the surface of a mirror-shaped circular sample with a light beam and moving the sample relative to the light beam in a direction intersecting with the scanning direction. In the method of inspecting the sample surface by detecting the pulsed scattered light generated by the irradiation of the above-mentioned light beam to an abnormal place such as adhesion or scratch, in the area other than the sample in the output signal of the light detecting means A first determination level and a second determination level are set in advance above and below the level of the DC component corresponding to the scattered light, and the DC component in which the output signal of the photodetector exceeds the second determination level in a certain scan. When a pulse component exceeding the first determination level is given for the first time after the application of the pulse, the time point when the DC component exceeds the second determination level and the time point when the pulse component exceeds the first determination level. Time The difference T is obtained, the first gate signal is issued when the number of scans after that time reaches a predetermined first value, and the output signal is output for each scan after the time when the first gate signal is issued. DC component of the second
The time difference t ₁ from the time when the judgment level of 1 is first exceeded to the time when it falls below the second judgment level, and
It is the time required for the light beam to move around the peripheral portion of the sample from the time when it falls below the second determination level, and the second time is obtained when a predetermined time t ₂ given as a function of the diameter of the sample and the number of scans elapses. Of the first gate signal and the second gate signal are ANDed to start counting the pulse components of the output signal in the scan, while the time differences T and t ₁ and the predetermined time t ₂ To 2 [T-
The value of (t ₁ + t ₂ )] is derived, and the second gate signal is cut off at the time when the above-mentioned 2 [T− (t ₁ + t ₂ )] has elapsed from the time when the second gate signal was issued. An optical surface inspection method, characterized in that the first gate signal is turned off at the time when the counting of the pulse component in the scanning is finished and the number of times of scanning reaches a predetermined second value. 2. The optical surface inspection method according to claim 1, wherein the number of scans of the first value is determined by previously giving a time for which the number of scans is performed. . 3. A method according to claim 1 or 2, characterized in that the second number of scans is determined by giving in advance the time during which the number of scans is performed. Optical surface inspection method. 4. The method according to any one of claims 1 to 3, wherein the output signal of the light detecting means shifts from a level lower than the second determination level to a level higher than the second determination level. An optical surface inspection method, characterized in that it is determined that the DC component of the output signal has exceeded a second determination level by continuing the level above that for a predetermined time. 5. The method according to any one of claims 1 to 4, wherein the output signal of the light detecting means shifts from a level above the second determination level to a level below it. An optical surface inspection method, characterized in that it is determined that the DC component of the output signal is below a second determination level by continuing the level below that for a predetermined time.