JP2003004654A - Optical scanner and defect detector - Google Patents

Abstract

PROBLEM TO BE SOLVED: To realize a high-resolution optical scanner by which not only the existence of a defect but also the irregular shape on the surface of a sample can be detected. SOLUTION: The optical scanner is provided with a beam splitter 2 by which a light beam advancing to the sample 4 from a light source 1 is separated from a reflected beam advancing to a photodetector 6 from the sample, and a means by which the sample and a light spot are moved relatively. The surface of the sample is scanned by the light spot, and information on the surface region of the sample is detected on the basis of reflected light from the surface of the sample. A light blocking plate is arranged in an optical path between the beam splitter and the photodetector, and half the optical path on one side in a direction corresponding to the scanning direction of the light spot on the surface of the sample is blocked. When the light blocking plate is arranged in the optical path, the reflected light from the light spot is displaced from its optical axis due to the law of reflection when a very small slope causing the defect exists on the surface of the sample, and a quantity of light blocked by the light blocking plate is changed. As a result, a convex defect and a concave defect can be discriminated on the basis of an output signal from the photodetector.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an optical scanning device and a defect detecting device using such an optical scanning device.

[0002]

2. Description of the Related Art With the miniaturization of LSIs, semiconductor wafers,
There is a strong demand for the development of a defect detection apparatus capable of accurately detecting fine defects existing on the surface of patterned wafers, mask blanks, photomasks, and the like. In particular, if there is a defect in the semiconductor wafer before the semiconductor element control process is performed, the manufacturing yield is significantly reduced.
There is a demand for the development of a defect detection device capable of accurately detecting a microscopic defect of about m.

In order to optically detect minute defects, it is necessary to use an optical system having high resolution. A confocal optical system is known as an optical system capable of achieving high resolution. FIG. 1 is a diagram for explaining the principle of a conventional confocal optical system. The light beam generated from the point light source or the laser light source 1 enters the objective lens 3 via the half mirror 2 that functions as a beam splitter. The objective lens 3 focuses the incident light beam in the form of a minute spot and projects it on the sample 4 arranged at the focal position of the objective lens. The reflected beam from the sample surface is reflected by the half mirror 2 and enters the photodetector 6 through the spatial filter 5 having the pinhole 5a. In this confocal optical system, the light reflected by the sample surface placed at the focal position of the objective lens 3 is pinhole 5
However, the reflected light from the sample surface deviated from the focus position cannot pass through the pinhole, and the output signal intensity of the photodetector is lowered. Therefore, the unevenness information of the sample surface is detected from the output signal intensity from the photodetector.

[0004]

Although the conventional confocal optical system can detect the height information of the sample surface, when the sample surface is located at a higher level or a lower level than the focal position of the objective lens. Both of them were detected as a decrease in the light amount of the light beam incident on the photodetector, and it was not possible to determine whether the light beam had a convex shape or a concave shape. Therefore, when it is mounted on a defect detection device, it is only possible to detect the presence of the defect, and it is not possible to detect the shape or property of the defect. In this case, the unevenness of the sample can be detected by sweeping the objective lens or the sample in the optical axis direction at the position where the defect is detected, but it is impossible to determine the unevenness in real time. Moreover, the resolution that can be detected is at most about several 100 nm, which does not meet the requirement for detecting a fine defect of about 50 nm.

Furthermore, in order to improve the resolution in the height direction, a method of incorporating a two-beam interference optical system can be considered.
Adjustment of the optical system is not easy because it is easily affected by air fluctuations and vibrations. There is also a method of introducing a shearing interference optical system using a Nomarski prism, but since the amount of shearing is small, the size in the lateral direction becomes large, making it difficult to observe when the sample surface is gently inclined. I will end up. Further, since the Nomarski prism is expensive, the cost of the optical system is also high.

Therefore, it is an object of the present invention to eliminate the above-mentioned defects and to realize an optical scanning device capable of accurately detecting not only the presence of defects but also the uneven shape of the sample surface.

Further, another object of the present invention is to provide a defect detecting apparatus capable of accurately detecting a minute defect of about 50 nm or less.

Further, another object of the present invention is to provide a defect detecting apparatus which can determine an appropriate processing process to be performed on a sample based on the detected defect information.

[0009]

An optical scanning device according to the present invention comprises a sample stage for supporting a sample, a light source for generating a light beam, and an objective lens for focusing the light beam to form a light spot on the sample surface. , A photodetector that receives the reflected light from the sample surface, and is arranged in the optical path between the light source and the objective lens, and separates the light beam from the light source toward the sample and the reflected beam from the sample toward the photodetector. Beam splitter, and means for relatively moving the sample and the light spot, the sample surface is scanned by the light spot,
In an optical scanning device for detecting information on the surface area of a sample by reflected light from the sample surface, a light shielding plate is arranged in the optical path between the beam splitter and the photodetector, and this light shielding plate allows It is characterized in that one half of the optical path in the direction corresponding to the scanning direction of the light spot is shielded.

In samples such as semiconductor wafers and mask blanks, many defects have a convex or concave inclined surface. When the sample surface is inclined, the reflected light from the light spot is displaced in the direction away from the optical axis according to the law of reflection. Therefore, by disposing a light shielding plate that shields the optical path of one half of the direction corresponding to the scanning direction of the light spot on the sample surface in the optical path of the imaging optical system,
The amount of light incident on the light receiving element of the photodetector changes significantly depending on whether the inclined surface is concave or convex. This change in the amount of light can significantly improve the resolution of defect detection sensitivity.
Further, in the case of a convex defect, a positive peak and a negative peak are alternately generated in the output signal from the light receiving element, and in the case of a concave defect, a negative peak and a positive peak appear alternately. Therefore, the shape of the defect can be determined based on the output signal from the light receiving element.

A process embodiment of the optical scanning device according to the present invention is such that, when a sample stage supporting a sample, a light source for generating a radiation beam, and m and n are natural numbers of 2 or more,
A two-dimensional diffraction grating that converts the radiation beam into a beam array of light beams arranged in a matrix of m rows and n columns, a beam deflecting device that deflects the beam array in a first scanning direction, and a beam array of the beam array. An objective lens that focuses a light beam in a spot shape to form an array of m × n light spots on a sample, and a light beam that is arranged in an optical path between the light source and the objective lens, and travels from the light source toward the sample. And a beam splitter for separating a reflected beam from the sample toward the photodetector, and a light receiving element arranged in a matrix of m ′ rows and n ′ columns when m ′ and n ′ are natural numbers of 2 or more, Each light receiving element is arranged at a pupil position in the optical path between the photodetector for receiving the reflected light from the light spot formed on the sample and the beam splitter and the photodetector, and on the sample surface. Comprising a light shielding plate for shielding the optical path in the direction of one half corresponding to the scanning direction of the spot,
The reflected beam from the surface of the sample is made incident on the light receiving element of the photodetector through the beam deflector, and the light spot array formed on the sample is moved in the direction orthogonal to the first direction. It is characterized in that the light spots adjacent to each other are formed so as to have an equal interval when projected onto. In this embodiment, since the sample surface is scanned by the m × n light spot array, the sample surface can be scanned at an extremely high speed and an optical scanning device having higher resolution can be realized.

[0012]

FIG. 2 is a diagram for explaining the principle of the optical scanning device according to the present invention. The same members as those used in FIG. 1 will be described with the same reference numerals. Sample 4
2A shows a state in which the light spot scans an inclined surface that rises upward, and FIG. 2B shows an inclined surface that gradually decreases. Indicates the state to do. In FIGS. 2A and 2B, the light spot scans the sample surface along the arrow direction, and the solid line represents virtual reflected light from the flat sample surface located at the focal plane of the objective lens. The broken line shows the reflected light from the inclined surface of the defect. In the present invention, the beam splitter 2 and the photodetector 6 are
A light-shielding plate 7 is arranged in the optical path between and, and the light-shielding plate 7 allows a light spot on the optical path from the sample 4 to the photodetector 6 to scan one surface of the front side or the rear side in the scanning direction for scanning the sample surface. The light is shielded (in FIG. 2, the front side of the optical path is shielded).

When scanning a convex inclined surface in which the light spot gradually rises, as shown in FIG. 2 (a), the reflected beam reflected by the sample surface is caused by the law of reflection so that the sample surface is the objective lens. It is displaced rearward in the scanning direction as compared with the case of scanning a flat reference surface located on the focal plane. On the other hand, when scanning a concave inclined surface where the light spot gradually lowers, as shown in FIG. 2B, the reflected beam reflected on the surface of the sample is compared with the case where the reference surface is scanned according to the law of reflection. It is displaced to the front side in the scanning direction. Therefore, according to the present invention, when the light shield plate 7 shields the front half of the optical path in the scanning direction, when the convex inclined surface is scanned, and when the reference plane is scanned, the light shield plate 7 shields the light. The light amount of the beam is reduced, and more reflected light is incident on the photodetector 6. On the other hand, when scanning a concave inclined surface in which the light spot gradually lowers, compared with the case of scanning the reference surface, the light blocking plate 7
As a result, the light amount of the reflected beam that is shielded increases, and a smaller amount of reflected light enters the photodetector 6.

FIG. 3 shows the output signal of the photodetector when the light spot scans the portion where the convex defect exists on the sample surface and the output signal of the photodetector when the portion where the light spot scans the concave defect. Show. A curve a shows an output signal intensity when scanning a convex defect, a curve b shows an output signal intensity when scanning a concave defect, and a curve c shows a light spot having a convex defect when the light shielding plate 7 is not present. Or, it shows the output signal intensity when the concave defect is scanned. When the light spot scans a portion of the sample surface where a convex defect exists, a convex peak is generated first, and then a concave peak is generated continuously in time. On the other hand, when the light spot scans a portion where a concave defect is present, a concave peak is generated first, and a convex peak is generated continuously in time. When the light-shielding plate is not used, as shown by the curve c, only a gradually changing concave peak is generated. Therefore, the shape of the defect can be determined from the output signal of the photodetector, and when the convex peak and the concave peak are alternately generated, it can be determined that the convex defect is present, and When peaks and convex peaks alternate, it can be determined that a concave defect is present. In addition, when the arrangement position of the light shielding plate is reversed, that is, when the light shielding plate is arranged on the rear side in the scanning direction of the optical path, only the generation order of the convex peak and the concave peak is opposite, and the convex defect is When present, a concave peak is generated first, and then a convex peak is generated.

FIG. 4 is a diagram showing the construction of an example of the optical scanning device according to the present invention. The laser beam generated from the laser light source 10 enters the one-dimensional diffraction grating 11. The diffraction grating 11 converts the incident laser beam into n sub beams. It is assumed that these sub-beams are aligned in the plane of the paper. The n sub beams emitted from the diffraction grating 11 pass through the first and second relay lenses 12 and 13,
The light enters the objective lens 15 via the beam splitter 14. The objective lens focuses the incident n light beams into minute spots to form n light spots on the sample to be inspected. The sample stage 17 supporting the sample 16 is xy
The xy stage has a drive mechanism. Sample stage 1
By moving 7 in a direction orthogonal to the paper surface, the sample surface is scanned by n light spots.

The reflected beam from the light spot formed on the sample 16 again passes through the objective lens 15 and is reflected by the beam splitter 14, and passes through the relay lens 18 and the photodetector 1
It is incident on 9. The photodetector 19 is a linear image sensor having a plurality of light receiving elements arranged in a line.
Then, the reflected beams from the respective light spots on the sample respectively enter the corresponding light receiving elements. The light receiving elements of the linear image sensor are separated from each other by a light blocking member, and the light blocking member or the frame thereof functions as a pinhole for restricting light incident on the light receiving element. Therefore, the linear image sensor itself has a spatial filter having pinholes. In this example, the light shielding plate 20 is arranged in the optical path between the relay lens 18 and the photodetector 19. For the sake of convenience of the drawings, the light shield plate 20 is shown so as to shield one half of the optical path in the direction orthogonal to the scanning direction, but in reality, the one half on the front side or the rear side in the scanning direction (the direction orthogonal to the paper surface). Shall be shaded.

FIG. 5 is a diagram showing the structure of an example of the defect detecting apparatus according to the present invention. A laser beam generated from the laser light source 30 is incident on the diffraction grating 31 and converted into a plurality of light beams aligned in the first direction. For the sake of clarity of the drawing, it is assumed that the diffracted light is diffracted within the plane of the paper. These light beams are transmitted to the first and second relay lenses 32.
And 33 to enter the polarization beam splitter 34, and the light passes through the polarization beam splitter 34 and enters the carbano mirror 35. The galvanometer mirror 35 periodically deflects the incident light beam at a predetermined frequency in a second direction orthogonal to the first direction (scanning direction orthogonal to the paper surface). The plurality of light beams reflected by the galvanometer mirror enter the objective lens 39 via the third and fourth relay lenses 36 and 37 and the quarter-wave plate 38. This objective lens focuses a plurality of incident light beams in the form of minute spots and projects them onto the sample 40 to be inspected for defects. Therefore, a plurality of minute light spots are formed in a line on the sample 40 along the direction corresponding to the first direction. Since these light spots are deflected in the second direction orthogonal to the direction of the light spot row, the sample 40 is scanned by the plurality of light beams, and thus the sample 40 is two-dimensionally scanned by the plurality of light beams. .

The sample 40 is placed on a stage 41 having an xy drive mechanism. The stage 41 moves at a predetermined speed in a direction (first direction) orthogonal to the deflection direction of the galvanometer mirror. Therefore, the sample is two-dimensionally scanned by the plurality of light spots. The reflected light from each light spot on the sample is condensed by the objective lens 39 and is incident on the galvanometer mirror 35 via the quarter-wave plate 38, the relay lenses 37 and 36. Then, it is descanned by the galvanometer mirror and enters the polarization beam splitter 34. Since the incident reflected beam is transmitted through the quarter-wave plate 38 twice, its polarization plane is rotated by 90 °. As a result,
The reflected beam from the sample reflects off the plane of polarization of the polarizing beam splitter and is separated from the illumination beam from the light source. The reflected beam emitted from the polarization beam splitter passes through the relay lens 42 and enters the linear image sensor 43. The linear image sensor 43 has a plurality of light receiving elements arranged in a line along a direction corresponding to the light spot on the sample.

An optical system between the sample and the linear image sensor constitutes an image forming optical system, and a light shielding plate 44 is arranged in the optical path between the polarization beam splitter 34 and the photodetector 43.
In the drawing, the light shielding plate 44 is shown so as to shield the optical path in the direction orthogonal to the scanning direction of the galvano mirror, but in reality, it shields one half of the optical path in the scanning direction of the galvano mirror which is the second direction. Arrange to do. The light blocking member or frame that separates each light receiving element of the linear image sensor functions as a spatial filter having a pinhole that restricts the light beam incident on the light receiving element.

Since the reflected beam from each light spot on the sample 40 is descanned by the galvanometer mirror 35, the reflected beam from each light spot on the sample is incident on each corresponding light receiving element of the linear image sensor 43. And stay stationary. The output signal from each light receiving element of the photodetector is amplified by the amplifier 45 and supplied to the defect detection circuit 46.

FIG. 6 is a circuit diagram showing the configuration of an example of the defect detection circuit. This example comprises a first comparator 50 for comparing the output signal from each light receiving element of the photodetector 43 with a negative limit value and a second comparator 51 for comparing with a positive limit value. The outputs of the first and second comparators are connected to the OR circuit 53. As described above, when a positive peak and a negative peak are alternately generated from the light receiving element, it is determined as a convex defect, and when a negative peak and a positive peak are alternately generated, it is determined as a convex defect. To do. If only a negative peak occurs, it is determined that the defect is a steep slope. Further, when only a positive peak is generated, it can be determined that the substance having high reflectance is locally attached. Therefore, by using the present invention, not only the existence of a defect but also the shape and property of the defect can be determined, and the processing process after the inspection can be accurately determined.

Since the video signal can be generated by reading the charge accumulated in each light receiving element of the photodetector at a predetermined reading frequency, the optical system shown in FIG. 5 can be used as a microscope.

FIG. 7 is a diagram showing a modification of the optical scanning device according to the present invention. The radiation beam emitted from the light source 61 is reflected by the total reflection mirror 62 and enters the diffraction grating 63. This diffraction grating directs an incoming radiation beam into m rows and n
It is converted into a two-dimensional beam array arranged in a matrix of columns (m and n are natural numbers of 2 or more). These m × n
The light beams of the book form a two-dimensional beam array of light beams that are equally spaced in the row and column directions. The m × n light beams enter the polarization beam splitter 65 via the Fourier transform lens 64. The beam array passes through the polarization beam splitter 65 and enters the galvanometer mirror 66. The galvanometer mirror 66 deflects m × n light beams in a first direction (main scanning direction) at a predetermined frequency.
The light beam reflected by the galvanometer mirror is relay lens 6
The light enters the objective lens 69 through the 7 and 1/4 wavelength plates 68. The objective lens 69 focuses the light beam array of m rows and n columns into a minute spot shape to form a light spot array of m rows and n columns on the sample 70.

FIG. 8 is a diagram showing the relationship between the light spot array formed on the sample 70 and the beam deflection direction (first direction) of the galvanometer mirror, that is, the main scanning direction.
In FIG. 8, white circles indicate light spots, and black circles indicate projections of the light spots on the axis L in the direction orthogonal to the main scanning direction. For clarity of the drawing, a 4 × 4 array of light spots is shown. The interval between the light spots in each row direction is p1, and the interval between the light spots in the column direction is p2. In order to scan the sample surface without gaps and without overlapping of the light spots with the light spot array of m rows and n columns, the light spots should be projected at equal intervals on the axis line in the direction orthogonal to the main scanning direction. Need to be formed. Less than,
This condition will be described. The angle formed by the axis L _{c in} the row direction of the light spot array and the axis L in the main scanning direction is θ. The projection length of the interval p1 between the light spots in the row direction with respect to the axis L is p1 × cos θ. There must be n light spot projections in the column direction within this projection length. This condition can be expressed by p1 × cos θ = p2 × sin θ. Therefore, if the following equation, tan θ = (1 / n) × (p1 / p2) is satisfied, the projections of the respective light spots of the light spot array of m rows and n columns in the direction orthogonal to the main scanning direction are equally spaced. It is formed. When the intervals between the light spots in the row and column directions are equal (p1 = p2), the light spot array is formed so as to satisfy the following formula. tan θ = 1 / n

Since the light beam array is scanned in the main scanning direction by the galvanometer mirror, the surface of the sample is scanned by the light spot array arranged in a matrix of m rows and n columns. The stage 71 that supports the sample 70 is x
The y stage is moved at a predetermined speed in a direction orthogonal to the main scanning direction which is the first direction. With this configuration, the sample surface is two-dimensionally scanned with m × n light spots.

The reflected light from the light spot reflected on the sample surface again passes through the objective lens 69, the quarter-wave plate 68 and the relay lens 67, enters the galvanometer mirror 66, is descanned, and enters the beam splitter 65. Since this reflected beam has passed through the quarter-wave plate twice, the plane of polarization is rotated by 90 ° and is reflected by the plane of polarization of the beam splitter.
The reflected beam reflected by the beam splitter enters the photodetector 75 via the relay lens 72, the total reflection mirror 72, and the relay lenses 73 and 74.

FIG. 9 is a diagram showing the structure of an example of the photodetector. The photodetector 75 includes a two-dimensional matrix light receiving element array 76 (i,
j), and each light receiving element receives the specularly reflected light from each light spot formed on the sample. Each light receiving element is composed of a photodiode and is separated from each other by a light shielding member 77. The light incident area of each light receiving element is regulated by the light shielding member 76 so that only the specularly reflected light of the light spot formed on the sample surface is incident. Therefore, the optical system of this example constitutes a confocal optical system, and higher resolution can be obtained.

A light shielding plate 80 is arranged at the pupil position of the image forming optical system between the beam splitter 65 and the photodetector 75. The light blocking plate blocks the light path on one side of the sample in the direction corresponding to the main scanning direction. In this way, by disposing one light shielding plate at the pupil position, one side of the m × n reflected beams can be shielded.

The present invention is not limited to the above-described embodiments, and various changes and modifications can be made. For example, in the above-described embodiments, the apparatus for detecting the defects existing on the sample surface such as the semiconductor wafer and the photomask blanks has been described as an example, but the optical scanning apparatus according to the present invention uses a light beam like a laser microscope. It can also be applied to an image pickup device that scans the sample surface to pick up a sample image, and by placing a light shielding plate in the optical path of the imaging optical system from the sample to the photodetector, an image pickup device with higher resolution can be obtained. Can be realized. For example, a video signal can be easily formed by reading the charge of each light receiving element at a predetermined read frequency from the photodetector of the device shown in FIG.

Various defect detection circuits can be used as the defect detection circuit. For example, it is possible to compare the output signals of adjacent light receiving elements and detect the occurrence of a defect from the comparison result.

Further, in the above-mentioned embodiment, the example in which the sample surface is raster-scanned has been described, but as a means for relatively moving the sample and the light beam array, it is connected to the sample stage supporting the sample and the stage is connected. It is possible to use a stage drive device that includes a rotation drive device that rotates and a translation drive device that translates in a direction orthogonal to the rotation axis. In this case, since the sample is spirally scanned by the m × n light spot arrays by rotating and translating the sample stage, the sample can be scanned at higher speed and the sample surface can be scanned with high resolution. .

[Brief description of drawings]

FIG. 1 is a diagram for explaining the principle of a confocal optical system.

FIG. 2 is a diagram for explaining the principle of the optical scanning device according to the present invention.

FIG. 3 is a diagram showing an optical path of a reflected beam when a light spot scans a defective portion.

FIG. 4 is a diagram showing a configuration of an example of an optical scanning device according to the present invention.

FIG. 5 is a diagram showing a configuration of an example of a defect detection device according to the present invention.

FIG. 6 is a diagram showing a configuration of an example of a defect detection circuit.

FIG. 7 is a diagram showing a modification of the optical scanning device according to the present invention.

FIG. 8 is a diagram showing a relationship between a light spot array formed on a sample and a main scanning direction.

FIG. 9 is a plan view showing an example of a photodetector.

[Explanation of symbols]

1 light source 2 beam splitter 3 Objective lens 4 samples 5 Spatial filter 6 Photodetector

Continued front page    F term (reference) 2F065 AA49 AA61 FF44 GG04 HH04                       JJ02 JJ25 LL00 LL13 LL36                       LL37 LL42 LL62 MM03 PP12                       QQ25                 2G051 AA51 AA56 AB02 BA10 BB15                       CA03 CA07 CB01 CC07 DA07                       DA08                 2H052 AA08 AC04 AC15 AC18 AC28                       AC34 AD18 AD31 AF06

Claims (12)

[Claims] 1. A sample stage that supports a sample, a light source that generates a light beam, an objective lens that focuses the light beam to form a light spot on the sample surface, and light that receives reflected light from the sample surface. A detector, a beam splitter arranged in the optical path between the light source and the objective lens, for separating a light beam from the light source toward the sample and a reflected beam from the sample toward the photodetector, the sample and the light spot. In the optical scanning device for scanning the sample surface with a light spot, and detecting the information of the surface area of the sample by the reflected light from the sample surface, the beam splitter and the photodetector. A light-shielding plate is arranged in the light path between the two light-shielding plates, and the light-shielding plate shields one half of the light path in the direction corresponding to the scanning direction of the light spot on the sample surface. Manabu scanning device. 2. The optical scanning device according to claim 1, wherein the means for relatively moving the light spot and the sample are means for moving the sample stage with respect to the light spot. 3. The beam deflecting device arranged in the optical path between the beam splitter and the sample stage, as the means for relatively moving the light spot and the sample. Optical scanning device. 4. A light beam generated from the light source is converted into a one-dimensional beam array of m (m is a natural number of 2 or more) sub-beams arranged in a line in an optical path between the light source and the beam splitter. 2. The optical scanning device according to claim 1, wherein a one-dimensional diffraction grating is arranged, and the surface of the sample is scanned by m light spots formed on the sample. 5. A sub-beam in which a radiation beam generated from a light source is arranged in a matrix of m rows and n columns when m and n are natural numbers of 2 or more in an optical path between the light source and the beam splitter. A two-dimensional diffraction grating for converting into a two-dimensional beam array, the sample surface is scanned by m × n two-dimensional light spot arrays, and the light shielding plate is an optical system between the beam splitter and the photodetector. The optical scanning device according to claim 1, wherein the optical scanning device is arranged at the position of the pupil. 6. As a means for relatively moving the sample and the two-dimensional light spot array, a rotation drive device for rotating and rotating the sample stage and a translation for moving the sample stage along an axis orthogonal to the rotation axis thereof. The optical scanning device according to claim 3, wherein the sample is spirally scanned by a two-dimensional light spot array using a driving device. 7. A sample stage for supporting a sample to be detected for defects, a light source for generating a light beam, an objective lens for converging the light beam to form a light spot on the sample surface, the sample and the light spot. Means for relatively moving the light source, a photodetector for receiving the reflected light from the sample surface, a light beam disposed in the optical path between the light source and the objective lens, and a light beam directed from the light source to the sample and light detection from the sample. A beam splitter that separates a reflected beam toward a beam detector, and a defect detection circuit that detects a defect existing on a sample surface from an output signal from the photodetector, wherein the beam splitter and the photodetector And a light-shielding plate that is arranged in an optical path between the light-shielding plate and a light-shielding plate that shields one half of the optical path in the direction corresponding to the scanning direction of the light spot on the sample surface, When the output signal from the output device detects that a concave peak and a convex peak alternately occur continuously in time, a defect characterized by determining that a convex or concave defect is present Detection device. 8. The light beam generated from the light source is converted into a one-dimensional beam array of m (m is a natural number of 2 or more) sub-beams arranged in a line in an optical path between the light source and the beam splitter. 8. The defect detecting apparatus according to claim 7, wherein a one-dimensional diffraction grating is arranged, and the surface of the sample is scanned with m light spots formed on the sample. 9. A sub-beam in which a radiation beam generated from a light source is arranged in a matrix of m rows and n columns when m and n are natural numbers of 2 or more in an optical path between the light source and the beam splitter. 8. The defect detecting apparatus according to claim 7, wherein a two-dimensional diffraction grating for converting into a two-dimensional beam array is arranged, and the sample surface is scanned by a light spot array of m × n light spots. 10. A sample stage for supporting a sample, a light source for generating a radiation beam, and a light arraying the radiation beam in a matrix of m rows and n columns when m and n are natural numbers of 2 or more. A two-dimensional diffraction grating for converting the beam into a beam array; a beam deflecting device for deflecting the beam array in a first scanning direction; An objective lens that forms an array of light spots, and a beam splitter that is disposed in the optical path between the light source and the objective lens and separates a light beam from the light source toward the sample and a reflected beam from the sample toward the photodetector. And, when m'and n'are natural numbers of 2 or more, m'row n '
A photodetector having light-receiving elements arranged in a matrix of rows, each light-receiving element receiving the reflected light from the light spot formed on the sample, and between the beam splitter and the photodetector. A light shielding plate disposed at a pupil position in the optical path and shielding an optical path on one side half in a direction corresponding to the scanning direction of the light spot on the sample surface, and a reflected beam from the sample surface via the beam deflecting device. When the light spot array formed on the sample is projected into the light receiving element of the photodetector and these light spots are projected in the direction orthogonal to the first direction, the intervals between the adjacent light spots are equal. An optical scanning device characterized by being formed so that 11. An angle formed by an axis line in the row direction of an m-th row and n-column light spot array formed on the sample and an axis line in a direction orthogonal to the first direction is θ, and rows of the light spot array are set. The angle θ is set so as to satisfy the expression tan θ = (1 / n) × (p1 / p2) where p1 is the spot interval in the direction and p2 is the spot interval in the column direction. 9. The optical scanning device according to item 9. 12. The intervals p1 and p2 between the light spots in the row and column directions of the light spot array are defined as p1 = p2.
10. The optical scanning device according to claim 9, wherein the light spot array is formed so as to satisfy the expression tan θ = 1 / n.