JPH06167458A - Foreign matter inspecting apparatus - Google Patents

Abstract

PURPOSE:To surely and stably inspect foreign matter in a simple structure. CONSTITUTION:Light sources 2, 3 radiate on a reticle 6 in the diagonal direction and a diffracted light occurring on a circuit pattern is condensed by an objective lens 41. Then the diffracted light of an angle of 0 deg. from the pattern is shaded by a shade section of a linear type space filter 44 so that diffracted light of an angle of 45 deg. or 90 deg. from the pattern is neither incident on the objective lens 41 nor detected by a sensor 51. As a scattered light from foreign matter has not the directivity, it is spread in a whole face of a Fourier-transform face. A part of the scattered light passes through a permeation section of the space filter 44 and focuses on the sensor 51 so that the foreign matter is differentiated with the circuit pattern to be detected.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a foreign matter detecting device for detecting foreign matter adhering to a circuit pattern of a reticle or a photomask (hereinafter referred to as "reticle"), and more particularly to a submicron on a reticle having a phase shift film. Which are suitable for inspecting foreign substances.

[0002]

2. Description of the Related Art In the exposure process of a reticle used for manufacturing an LSI or a printed board, if a minute foreign substance is present on a circuit pattern, the shadow of the foreign substance is also transferred, resulting in the total number of chips. There is a problem that becomes defective. This problem has become more serious with the recent high integration of LSIs, and the presence of, for example, smaller foreign matter in submicron units has become unacceptable. Therefore, in terms of management of the reticle and the like, a foreign matter inspection device has been conventionally proposed, and generally, the foreign matter inspection of a circuit pattern of the reticle or the like is performed diagonally above the inspection surface with a light source having good directivity such as a laser beam. It is said that a method of irradiating the particles with a foreign substance to generate scattered light on the foreign substance and detecting the scattered light from the foreign substance is advantageous in terms of inspection speed and sensitivity. However, in the above inspection method,
When the circuit pattern is irradiated, not only scattered light is generated from the foreign matter, but also diffracted light is generated from the edge portion of the circuit pattern, so it is necessary to distinguish and detect only the foreign matter from this diffracted light. Various technologies have been proposed and put to practical use.

As prior art No. 1 for that purpose, a foreign matter inspection apparatus having a linearly polarized laser, means for obliquely irradiating the laser light at a specific incident angle, and an oblique imaging optical system using a polarizing plate and a lens. (For example, JP-A-54-101
390), and when linearly polarized light is radiated, only the foreign matter is detected by shining it by utilizing the fact that the diffracted light from the circuit pattern and the scattered light from the foreign matter have different polarization directions. There is something I did. As the conventional technique No. 2, focusing on the fact that the diffracted light generated at the edge portion of the circuit pattern has directivity, but the scattered light due to foreign matter has no directivity,
There is a configuration (for example, Japanese Patent Laid-Open No. 57-80546) that discriminates foreign matter by taking a logical product of the detection outputs of a plurality of detectors installed obliquely. As the conventional technique No. 3, by utilizing the phenomenon that the diffracted light from the edge portion of the circuit pattern is concentrated only in a certain specific direction, the scattered light from the foreign matter is scattered in all directions, There is one that arranges a plurality of detectors to discriminate foreign substances (for example, Japanese Patent Laid-Open Nos. 60-154634 and 60-154635). As the fourth conventional technique, a means for irradiating the sample to be inspected with a laser beam obliquely and scanning is provided above the sample to be inspected so that the irradiation point of the laser beam and the converging point surface are substantially coincident with each other. A first lens that collects scattered light and a linear light-shielding plate (a linear light-shielding plate that is provided on the Fourier transform surface of the first lens and that shields regularly diffracted light from the circuit pattern of the sample to be inspected (linear shape) Spatial filter), a second lens for performing an inverse Fourier transform on scattered light from a foreign substance obtained through the spatial filter, and a laser beam irradiation point on the sample to be inspected provided at the image forming point of the second lens. A foreign matter inspection device including a slit that shields scattered light from other than the above and a light receiver that receives scattered light from the foreign matter that has passed through the slit (for example, Japanese Patent Laid-Open Nos. 59-65428 and 1- No. 117024, 1- 53,943 JP)
There is. This device pays attention to the fact that the circuit pattern is generally composed of the same direction or a combination of two or three directions within the field of view, and the diffracted light by the circuit pattern in that direction is installed on the Fourier transform plane. By removing with a filter, only scattered light from the foreign matter is emphasized and detected. As the prior art No. 5, the circuit pattern of the reticle / photomask such as LSI is repetitive (periodic).
By focusing on the fact that the diffracted light from the periodic circuit pattern is focused on the Fourier transform plane at a unique position of the shape of each circuit pattern, each inspected sample As a configuration that can be rewritten or exchanged with a spatial filter corresponding to, the spatial filter is created by adapting it to the Fourier transform image of various circuit patterns, and diffracted light from the circuit pattern can be effectively changed by exchanging each inspected sample. Aiming at improving the detection ability by shielding the light from the light (for example, JP-A-56-89714 and JP-A-1-158308).
No. 2-114386, No. 2-174424
No. 3, gazette 3-130647 gazette, 4-10915.
No. 2). As prior art No. 6, when an array-like detector such as a one-dimensional solid-state image sensor is used and a foreign matter is detected across pixels forming the array and the pixels, the output from the foreign matter is a plurality of pixels. Are dispersed and detected. As a result, the output from the detector becomes smaller by the dispersed amount, and there is a risk that foreign matter may be missed.To avoid this, the array of detectors is tilted with respect to the scanning direction of the sample stage. (For example, Japanese Unexamined Patent Publication No. 61-104242) or those using an array-shaped detector having a special shape or arrangement (for example, Japanese Unexamined Patent Publication No. 61-1).
No. 04244). As the prior art No. 7, since the unevenness of illumination or fluctuation affects the reproducibility and accuracy of detection, automatic calibration is performed using a standard sample in which the intensity of scattered light is measured in advance (for example, JP-A-60- 038827
Issue gazette). As prior art 8, a large amount of scattered light generated from a large foreign substance is prevented from being mistaken for scattered light from a large number of small foreign substances (for example, Japanese Patent Laid-Open No. 56-1
No. 32549). As related to the inspection of other minute foreign matters, there are techniques related to the Schlieren method, a phase contrast microscope, a diffraction image of a light source with a finite size, and the like, for example, Hiroshi Kubota, "Applied Optics (Iwanami Zensho)", p. It is described on page 136. In addition, the technology related to materials that can be used for rewritable spatial filters has been developed by Toshiharu Tanaka.
Others "Optical Measurement Handbook (Asakura Shoten)" No. 106
Pp. 109-109.

[0004]

By the way, in the above-mentioned prior art, if the foreign matter to be detected becomes extremely small, for example, in the submicron unit, there are the following problems. That is, in the related art 1, since the difference between the polarization direction of scattered light from a minute foreign substance and the polarization direction of diffracted light from the edge portion of the circuit pattern becomes small, it is difficult to distinguish only the foreign substance. There is. In the prior art Nos. 2 and 3, it is difficult to adopt an optical system having a sufficient light-collecting ability due to the device configuration, and it is difficult to detect weak scattered light generated from minute foreign matter. Is. In the prior art No. 4, it is assumed that the regular diffracted light from the circuit pattern is blocked by the linear spatial filter, but the diffracted light from the intersection of the circuit patterns has a specific position like the diffracted light from the straight portion. Therefore, it is not possible to completely block the diffracted light from the intersection of the circuit patterns by the spatial filter, and moreover, a circuit having a microstructure pattern of the micron order accompanying the recent high integration of LSIs. In the pattern, the diffracted light generated from this has a behavior similar to that of the scattered light from the foreign matter, so that there is a problem that it becomes more difficult. Recently, in order to improve the transfer resolution of the circuit pattern on the reticle, a transparent film called a phase shift film or a phase shifter between the circuit patterns on the reticle, or a semitransparent film (about 1 / wavelength of the exposure light source is used). A reticle provided with a circuit pattern having a film thickness of an odd multiple of 2) has been developed. Since this film has a structure which is formed by a metal thin film such as chromium and has a size several times as large as the circuit pattern having a thickness of about 0.1 μm, the diffracted light from the edge portion of the film is Compared with the conventional diffracted light from the edge portion of the circuit pattern, it becomes several times to several tens of times larger, and the detection sensitivity of foreign matter is significantly reduced. For this reason, it is practically difficult to detect foreign matter by separating it from the circuit pattern with a linear spatial filter having a simple structure. In the prior art No. 5, as described above, the principle that the diffracted light from the periodic circuit pattern is focused on the Fourier transform plane at a position peculiar to the shape of each circuit pattern, and most of these are used. In the prior art, the sample to be inspected is illuminated with light that is transmitted through or incident on the sample (bright field), and therefore a spatial filter that blocks specularly transmitted light or specularly reflected light (zero-order diffracted light) is arranged. . However, since the 0th-order diffracted light is significantly brighter than the 1st-order and diffracted light, a spatial filter that shields the 0th-order diffracted light requires a precise position, and the ratio of the spatial filter in the aperture (NA) of the detection system is high. Becomes large, the effective NA decreases, and 16M
It is unsuitable for detecting foreign matter in submicron units required for 0.5 micron LSI or later represented by DRAM. In particular, for the phase shift reticle, the shape of the phase shift film is more unstable than that of the circuit pattern made of the metal thin film, so that it becomes difficult to position the spatial filter. Further, in all the prior arts in Part 5, in order to make the range of the 0th-order diffracted light which is remarkably bright as compared with the diffracted light of the 1st or higher order, the illumination by the parallel beam is performed
Since the illuminance of the illumination is lower than that of the illumination by the condensed beam, there is a problem that some measure is required for detecting submicron order with low S / N. Furthermore, since the repetitiveness (periodicity) of the circuit pattern is premised, the scattered light from the circuit pattern cannot be removed by the spatial filter in the non-repeatable portion partially present in the LSI chip, Some kind of countermeasure is necessary. Japanese Unexamined Patent Publication No. 4-109152 discloses an example of this measure. This is because the non-repeatable part is extracted in advance from design data and the corresponding part is masked during automatic inspection. I try to exclude it from the range. Then, it is a method of estimating the adhesion state of foreign matter on all the inspected sample areas including the masked area from the detected state. This is an effective method when it is not necessary to detect all the foreign matters on the entire surface. However, since the reticle serves as an original plate of an LSI and all foreign matters on the entire surface of the sample on which the circuit pattern is formed must be detected, the method of masking the inspection area as described in the above publication cannot be applied. There is. The prior art No. 6 has a problem that not only the special detector which is necessary for the construction and the special detector has to be specially manufactured, but also the special optical system has to be constructed, which increases the cost. Prior art No. 7 has a problem in that it corresponds to an array-shaped detector suitable for high-speed detection and in terms of configuration accuracy for minute foreign matter. In the prior art No. 8,
Since only one point of a large foreign matter is represented, there is a problem that the shape of a slender foreign matter cannot be accurately recognized.

In view of the above-mentioned problems of the prior art, an object of the present invention is to detect fine foreign matter in submicron units adhering to a circuit pattern such as a reticle or a photomask easily and stably with a simple structure. It is to provide a foreign matter inspection device capable of performing the above.

[0006]

According to the present invention, there is provided a foreign matter inspection device for detecting foreign matter deposited on a transparent or semitransparent sample having a periodic circuit pattern, wherein the sample is parallel to the sample surface. An inspection stage that selectively moves in a direction orthogonal to each other on the same plane and a direction perpendicular to the plane, an illumination system that obliquely irradiates a desired position on the surface of the sample, and an illumination area of the sample , Which are respectively provided on the Fourier transform surface in the vertically upper position, and locally shield the diffracted light of the first or higher order generated from the pattern of the circuit pattern at a specific angle in the direction orthogonal to the irradiation direction of the illumination system, and A detection optical system having a rewritable spatial filter whose portion transmits scattered light generated from the circuit pattern, a detector for detecting scattered light from the sample through the detection optical system, and a detector. The detection signal from the vessel binarized, and a judging means for discriminating the scattered light from the foreign matter on the sample.

[0007]

Operation: Wolf's "Prinsple of Op"
According to documents such as tics "pp.647-664,
It has been argued that when the fine particles have the same size as the wavelength of the illumination light, the scattered light from the foreign matter is not uniform and has a sharp distribution. The present inventor has noticed that the minute foreign matter is missed because the scattered light from the minute foreign matter has a distribution. In the past, not only was there no mention of the numerical aperture of the detection optical system, but it was thought that when detecting a foreign substance, even if the detection optical system cannot resolve the foreign substance, it is possible to detect it. However, as shown in the above-mentioned document, since scattered light from a minute foreign substance has irregular directivity, it may not be detected by a detection optical system having a small numerical aperture (NA). It is thought that the missed detection will occur. That is, it is clear that the detection optical system having the resolution of the prior art "may detect a minute foreign substance" but not "stablely detect". It has been found that in order to achieve the target of "detection of foreign matter", it is necessary to have a resolution enough to resolve the size of the foreign matter to be detected. The examination process is described below. The simplest problem that a single sphere floating in space is irradiated with a plane wave is 190
First analyzed by Gustav Mie in 8 years. An analysis method known as Mie's theory is a quadratic series of a mathematical function called a spherical harmonics, but it is out of the gist of the present invention and will not be described. However,
The interpretation is easy to compare. Particles such as latex spheres scatter light in an incident beam through a combination of processes of reflection, refraction, absorption and diffraction. FIG. 46 shows the intensity of scattered light from spherical foreign matters (particles). FIG. 46 shows
The theoretical value of Mie scattering is applied in the case of particles adhering to the substrate, the vertical axis represents the scattered light intensity, the horizontal axis represents the wavelength λ (for example, 550 nm) of the detection light, and d represents the diameter of the detected foreign matter. It represents the dimensionless number used. Here, a region where πd / λ is smaller than about 4 (when λ = 550 nm, d = 0.
The foreign matter smaller than 7 μm) is particularly called a Rayleigh scattering region, and the scattered light from the foreign matter sharply decreases in inverse proportion to the sixth power of the diameter. Therefore, in detecting the foreign matter in this region, it is necessary to pay sufficient attention to the sensitivity of the detector. πd / λ
In a region where is larger than about 4, the scattered light is directionally scattered according to the theory of diffraction. The situation is as shown in FIG. In order to detect foreign matter in this area, the scattered light from the foreign matter has a distribution, so the NA of the detector
Should be determined by paying attention to the distribution of foreign matter. FIG. 47 shows the direction of the diffracted light when the laser illumination 4721 is performed on the foreign material 70 on the reticle 6. The diffracted light from the foreign material 70 continues to the 0th-order diffracted light 4722, the one-dimensional diffracted light 4723, and the second-order diffracted light with the angle θ. In the figure, the 0th-order diffracted light 4722 is the specularly reflected light of the laser illumination 4721, and detecting scattered light from a foreign substance means detecting diffracted light of the first order or higher. here,
θ can be obtained from the equation of diffracted light as do · sin θ = λ. However, do is the diameter,
Various definitions such as an average value of width, length, or diameter are conceivable, but since the following discussion holds regardless of the value of do, any of the above definitions does not affect the result. Therefore, here, do = d, that is, do is assumed to be the particle diameter.
In addition, the NA required for the detection optical system is πd, which is the most severe condition.
When calculated for / λ = 4, the following equation is obtained.

[0008]

[Equation 1]

[0009]

[Equation 2]

This means that the maximum gap between diffracted lights is 52 °. Therefore, if a detection optical system having an aperture of 52 ° or more is used, only the first-order diffracted light can be detected. It doesn't mean that you miss a foreign object.

FIG. 48 is a diagram showing the definition of NA of the optical system. In the figure, NA = sin (θ /
In the case of 2), the NA of the detection system objective lens 41 is obtained, and NA = 1 · sin (52 ° / 2) = 0.44. Therefore, it can be understood that the scattered light from the foreign matter can be detected by the detection system with an NA larger than about 0.44 without overlooking. In this case, the larger the NA, the more room for detection,
In addition, it is convenient for detecting foreign matter in the Rayleigh region. Then, even if NA ≧ 0.44 is not satisfied, NA =
If it is about 0.4, the diffracted light has a certain width,
In practice, it is possible to detect foreign matter. Conversely, NA is 0.5
If it is made larger, the diffracted light from the circuit pattern is incident on the detection system for the reason to be described later, which impedes the requirement to detect only the scattered light from the foreign matter, and the merit of intentionally increasing the NA is reduced. Therefore, the NA of about 0.4 to 0.6 becomes a practically appropriate NA.

Next, detection of foreign matter in the Rayleigh region will be described. As described above, the detection optical system having the resolution of the prior art "may detect minute foreign matter".
However, it is not a "stable detection". In order to achieve the target of "detection of foreign matter", it is necessary to have a resolution that can resolve the size of the foreign matter to be detected.
The present invention has a detection optical system having a numerical aperture (NA) that can resolve the foreign matter to be detected. Specifically, it is calculated by the following equation.

[0013]

[Equation 3]

An optical system having a value substantially close to this NA is desirable. Note that d is the size of the foreign matter to be detected, λ is the wavelength of the illumination system, and NA is the numerical aperture. If the NA of the detection system cannot be set to satisfy the above equation, the wavelength λ of the illumination system
Must be shortened to satisfy the above formula. That is, in the detection optical system for inspecting a foreign substance, it has been conventionally considered that the resolving power for resolving the foreign substance is required, but in the present invention, the detection optical system for resolving the foreign substance as shown in the above formula is required. It is based on a new way of thinking. However, the coefficient of the above equation does not need to be as large as the value for calculating a general resolution of 0.6, and according to the experiment by the inventor of the present application, it is 0.24.
It has been found that the required foreign matter detection performance is exhibited in the range of up to 0.6. In FIG. 49, the abscissa represents the particle diameter and the ordinate represents the scattering cross section. This scattering cross section is proportional to the scattered light generated from the foreign matter, and is obtained from Mie's theory of scattering. The interpretation means that when the generated scattered light is observed, the scattered light is observed as if it were the scattered light generated from the foreign substance shown by the solid line in the figure. In the figure, the cross-sectional area is also shown geometrically with a broken line. From FIG. 49, it can be seen that when observed with scattered light, the size is larger than the actual size of the foreign matter. This is exactly the reason why foreign material inspection is performed with scattered light. And the ratio is about 3 to 6 in area ratio.
Therefore, the diameter is √3 to √6. In that case, the following formula can be used to explain the previous experimental results.

[0015]

[Equation 4]

In the foreign matter inspection on the reticle, since the foreign matter size d to be detected is about ¼ of the minimum reticle size, the minimum reticle size 2.5 μm (in the case of 5: 1 reduction transfer, on the wafer). 0.5 μm, which is 16
It is 0.6 μm in the case of MDRAM) and 0.4 μm in the case of the minimum reticle size of 1.5 μm (corresponding to 64MDRAM). Therefore, in order to detect a foreign matter having a size of 0.4 μm with the detection optical system having NA = 0.4 obtained from the previous study, it can be obtained from the following equation, which is a modification of the above equation (4).

[0017]

[Equation 5]

That is, a light source having a wavelength shorter than λ = 660 nm to 460 nm is required.

Next, consideration will be given to selecting a wavelength suitable for inspecting a foreign substance on a sample such as a reticle on which a circuit pattern is formed in the wavelength range. Before that, a prerequisite foreign substance is extracted from the circuit pattern. The principle of optically separating and detecting will be described. When the circuit pattern is radiated obliquely at an incident angle θ (θ <90 °) with laser light having good directivity,
The Fourier transform image of the diffracted light from the straight line portion of the circuit pattern is condensed into a thin straight line at a specific position on the Fourier transform image plane regardless of the position of the circuit pattern in the illumination field. It is known that scattered light is not biased to a specific position on the Fourier transform image plane. Therefore, it is easy to arrange a light shield plate (spatial filter) at a specific position on the Fourier transform image plane, block the diffracted light from the straight line portion of the circuit pattern by the light shield plate, and detect only the scattered light from the foreign matter. Conceivable. However, in that case, the diffracted light from the corner portion of the circuit pattern and the fine structure portion where the corner portion is continuous cannot be shielded completely. Therefore, when detection is performed by a conventional detector, for example, a detector having a detection pixel of 10 × 20 μm ² , diffracted light from a plurality of pattern corners enters the pixel and only foreign matter is detected. Can not do it. Diffracted light generated from a periodic circuit pattern having a repeating period T by illuminating the sample to be inspected with coherent parallel light such as a laser of wavelength λ, or substantially parallel light or condensed light with NA ≦ 0.2. Is Fourier-transformed by the Fourier transform lens having the focal length f, the diffracted light is discretely condensed at a position of an integral multiple of λ · f / T. As a result, it is condensed as a bright spot at a position peculiar to the shape of each circuit pattern on the Fourier transform surface. On the other hand, it is known that foreign matter having no periodicity is widely distributed, and therefore, only the image of the foreign matter can be detected by blocking the condensing position of the image of the circuit pattern on the Fourier transform surface with a spatial filter.
However, such conventional devices have the following three points in common. That is, (1) illumination is performed with parallel beams in order to increase the degree of light collection of the Fourier transform image. (2)
After the Fourier transform, the inverse Fourier transform is performed, and when the image on the inspection sample is restored (re-imaged), epi-illumination or transmitted illumination perpendicular to the sample surface is performed as a restoration of a more complete image. The diffracted light that is generated is incident on the Fourier transform lens including the 0th-order light. (3) A reticle that is required to detect foreign matters on the entire area of the sample to be inspected because it corresponds only to a circuit pattern having a periodicity and non-periodic circuit patterns existing around the periodicity pattern are excluded from the inspection target. It is difficult to apply it to foreign matter inspection such as. In view of the above, the present invention focuses on the point that the re-imaging image after the inverse Fourier transform does not have to be perfect as long as the detection of the foreign matter, that is, the presence or absence of the foreign matter can be determined. . That is, in the present invention, as described above, an inspection stage that selectively moves a sample in a direction perpendicular to the plane and a direction perpendicular to the plane on the same plane parallel to the sample surface, and a sample stage An illumination system that obliquely illuminates a desired position on the surface, and a pattern of a specific angle in a direction orthogonal to the illumination direction of the illumination system in the circuit pattern, which are respectively provided on the Fourier transform plane at a position vertically above the illumination area of the sample. A detection optical system having a rewritable spatial filter that locally blocks diffracted light of the first or higher order generated from, and transmits scattered light generated from the circuit pattern in the other part,
Since the detector is configured to detect scattered light from the sample through the detection optical system, and the determination signal that binarizes the detection signal from the detector and discriminates the scattered light from the foreign matter on the sample is configured. , When the sample is illuminated obliquely by the illumination system and the diffracted light generated in the circuit pattern is condensed by the detection optical system, the diffracted light from the specific angle pattern is blocked by the light shielding part of the spatial filter, so Since the diffracted light from the angle pattern does not enter the detection optical system, no diffracted light is detected by the detector. On the other hand, the scattered light from the foreign matter adhering to the circuit pattern spreads over the entire Fourier transform surface because it has no directivity, but the scattered light passes through the transmission part of the spatial filter,
Since the image is formed on the detector, the foreign matter is discriminated from the circuit pattern and detected. As a result, it becomes possible to reliably detect a submicron-order foreign substance on the entire surface of the reticle. At this time, if the pixel size of the detector is set to a resolution as high as the smallest pattern size on the reticle, foreign matter can be detected more without being influenced by the diffracted light from the pattern corner. In addition, the incident angle of illumination with respect to the sample to be inspected is obliquely set at a specific angle, and is significantly brighter than the diffracted light of the first order or more (generally tens to hundreds of times or more). (Folded light) does not enter the aperture of the detection optical system, which brings about the following three advantages. First, since it is not necessary to provide a light-shielding portion that shields the 0th-order diffracted light on the spatial filter, the ratio of the light-shielding portion in the aperture (NA) of the detection optical system is small, and a substantial NA is secured. , It is possible to stably detect foreign matters on the submicron order required after 0.5 micron LSI represented by 16M DRAM. The second is
Compared with the case where a remarkably bright 0th-order diffracted light enters the detection optical system, the influence of the positional deviation of the spatial filter is reduced. Not only is this an important advantage for changing the spatial filter to adapt to the shape of the circuit pattern, but it is also an advantage on the phase-shifting reticle where the shape of the film is unstable compared to the circuit pattern with a metal thin film. Enables reticle foreign matter detection. Thirdly, it is not necessary to illuminate with a parallel beam in order to narrow the range of the 0th-order diffracted light that is extremely bright as compared with the 1st-order and higher-order diffracted light as much as possible, and it is possible to detect foreign matter in the submicron order with low S / N. Lighting with a light beam can be realized.

[0020]

Embodiments of the present invention will be described below with reference to the drawings. FIG. 1 is an overall configuration diagram showing an embodiment of a foreign matter inspection apparatus according to the present invention. In FIG. 1, reference numeral 1 is an inspection stage unit, and the inspection stage unit 1 includes a Z stage 9 for fixing a reticle 6 having a pellicle 7 on an upper surface by a fixing means 8 and moving it in the Z direction, and a Z stage 9. It has an X stage 10 that moves the reticle 6 in the X direction, and a Y stage 11 that also moves the reticle 6 in the Y direction. Each of these Z stage 9, X stage 10, and Y stage 11 is selectively driven by a stage drive system 12, and the position of the reticle 6 in the Z direction is detected by a control system 13 for focus position detection. The control system 13 for detecting the focal position may be one that uses an air micrometer, or one that detects the position by a laser interferometry method, and may be one that is configured to project a fringe pattern and detect the contrast thereof. X stage 10 and Y stage 11
Is scanned in the directions orthogonal to each other on the same plane as shown in FIG. 2, and the scanning speed can be set arbitrarily. For example, the X stage 10 is set to a constant acceleration time of about 0.2 seconds, a constant velocity motion of 4.0 seconds, and a constant deceleration time of 0.2 seconds, and a stop time of about 0.2 seconds is set to 1 The maximum movement speed is about 25 mm / sec and the amplitude movement is 105 mm, and the Y stage 11 moves the reticle 6 to 0.
When the device is configured to move in steps of 5 mm in the Y direction, if it is moved 200 times during one inspection time, approximately 9
It is possible to transfer 100 mm in 60 seconds,
An area of mm mm can be scanned in about 960 seconds.

Reference numeral 2 is a first illumination system, and 3 is a second illumination system, both of which are independent of each other and have the same components. That is, the first and second illumination systems are roughly classified into laser light sources 21 and 31, and condenser lenses 22 and 32.
And is configured. In this example, the wavelength λ of the laser light source 21 is 514.5 nm, and the wavelength λ of the laser light source 31 is 54.5 nm.
Although different from 32 nm, the same wavelength may be used. The condenser lenses 22 and 32 condense the light fluxes emitted from the laser light sources 21 and 31, respectively, and irradiate them onto the circuit pattern of the reticle 6. In this case, the incident angle θ of both with respect to the circuit pattern is to avoid the objective lens 41 of the detection optical system 4 which will be described later, and the 0th-order diffracted light from the sample to be inspected is used for the objective lens 4.
To avoid entering 1
Further, in the case where the sample to be inspected is the reticle 6 with the pellicle 7 mounted, it must be smaller than approximately 80 ° in order to avoid the pellicle 7, and therefore approximately 20 ° ≦ θ ≦ 8.
A range of 0 ° is desirable. A detailed configuration example of the first illumination system 2 will be described with reference to FIG. In FIG. 3, FIG.
The same reference numerals denote the same things, 21 is a laser light source, and 22 is a condenser lens. 223 is a concave lens, 22
Reference numeral 4 is a cylindrical lens, 225 is a collimator lens, and 226 is a condenser lens.
This constitutes the condenser lens 22. Laser light source 21
Are arranged so as to have a linear deflection having an electric field vector in the X'direction (this state is called S deflection). For S-deflection, for example, when the incident angle i is about 60 °, the reflectance on the glass substrate is about 5 times higher than that of P-deflection (for example, Hiroshi Kubota “Applied Optics (Iwanami Zensho)”). This is because even smaller foreign matters can be detected. Then, in order to increase the illuminance of the first illumination system 2 and the second illumination system 3, the numerical aperture (N
A) is set to about 0.1 and the laser beam is narrowed down to about 10 μm. In this case, due to the narrowing down, the depth of focus is shortened to about 30 μm, and it may not be possible to focus on the entire area S (500 μm) of the inspection visual field 15 shown in FIG. However, in this example, as a countermeasure against this, a cylindrical lens 224 is attached so as to be tiltable around the axis X'as shown in FIG. By tilting with, the focus can be surely focused on the entire area S. Further, the cylindrical lens 224 is mounted so as to be tiltable about the axis of Y ′ in addition to the axis of X ′ shown in FIG. Is capable of performing linear illumination with high illuminance and a uniform distribution. Accordingly, when the output of the laser light source 21 is increased and the NA of the condensing optical system is decreased to increase the depth of focus, it is possible to make the illuminance distribution of illumination uniform. In FIG. 3, the first illumination system 2
Although only shown, the second illumination system 3 has the same configuration, and therefore its description is omitted here.

A detection optical system 4 includes an objective lens 41 facing the reticle 6 and an objective lens 41.
Field lens (hereinafter referred to as a field lens) 43 provided near the image forming position of, a mirror 42 for wavelength separation of the light beam condensed by the field lens 43, and linear spatial filters 44 and 444. And variable spatial filter 4
7, 447 and imaging lenses 45, 445, the inspection field of view 15 on the reticle 6 is detected by the detectors 51,
It is configured so that an image can be formed on 551. The field lens 43 has an upper focus position 46 on the objective lens 41.
Is imaged on the linear spatial filters 44 and 444 and the variable spatial filters 47 and 447. The linear spatial filters 44 and 444 are used for the inspection visual field 15 of the reticle 6.
The circuit pattern 80 is formed by including a light-shielding portion provided at the position of Fourier transform with respect to
When light is emitted, the light-shielding portion shields the first-order and higher-order diffracted light generated from the angle pattern orthogonal to the irradiation direction in the circuit pattern, while the light-transmitting portion transmits scattered light generated from foreign matter adhering to the circuit pattern. I am trying to do it. The variable spatial filters 47 and 447 are rewritable or replaceable in accordance with the Fourier transform image of scattered light generated from a portion where no foreign matter is attached,
Details will be described later. These linear spatial filters 44,
444 and the variable spatial filters 47 and 447 are both spatial light modulators and have substantially the same optical position.
It is composed of parts that have a sufficiently short optical path (thin) optically.

Reference numeral 5 denotes a judgment processing system. The judgment processing system 5 outputs detectors 51 and 551 and outputs of the detectors 51 and 551 to two.
First and second binarization circuits 52 and 552 for binarization processing
An AND circuit 53, a microcomputer 54,
The display means 55. Detector 51,551
Is composed of, for example, a charge transfer type one-dimensional fixed image pickup device, and when the X stage 10 is scanned, the detection optical system 4
The signal from the circuit pattern on the reticle 6 is detected via. In that case, if foreign matter is present on the reticle 6,
Since the input signal level and the light intensity increase, the outputs of the detectors 51 and 551 also increase accordingly. The binarization circuits 52 and 552 have preset binarization thresholds, and when the detectors 51 and 551 input an output value equal to or higher than the reflected light intensity corresponding to a foreign object of a target size, The logic level "1" is output. The logical product circuit 53 calculates the logical product of the output signals binarized by the binarization circuits 52 and 552. When the processing signal calculated by the AND circuit 53 is input via the block processing circuit 112, the microcomputer 54
According to this, it is determined whether or not there is a foreign matter, and when it is determined that there is a foreign matter, the position of the foreign matter calculated from the position information of the X stage 10 and the Y stage 11 and the pixel positions of the detectors 51 and 551. Information and detector 51,551
Is stored as foreign matter data, and the result is output to the display means 55. Figure 1
In the above, reference numerals 113 and 123 denote shading correction circuits that correct the detection values from the detectors 51 and 551 according to illumination unevenness, and 114 and 124 denote 4-pixel addition processing units, which will be described later.

Next, the operation of the foreign matter inspection apparatus will be described with reference to FIG.
From now on, it will be described with reference to FIG. 4 is a perspective view showing the principle configuration of the detection system, FIG. 5 is a plan view for explaining the angle pattern of the circuit pattern, FIG. 6 is a view showing the distribution of scattered light and diffracted light on the Fourier transform surface, and FIG. FIG. 7A is a diagram showing a corner portion of a circuit pattern, and FIG.
FIG. 8 is a detailed view of the “A” part, FIG. 8 is an explanatory view of the relationship between the scattered light detection output value from the foreign matter and the detection output value from the circuit pattern, and FIG. 9 is a diagram showing a circuit pattern having a fine structure pattern,
FIG. 10 is a diagram showing the output value levels of the detection signals detected from the foreign matters and the corners of the circuit pattern. Figure 4
In (A), 70 is the Z stage 9 by the fixing means 8.
The foreign matter on the reticle 6 fixed on the upper surface is a straight line portion 81 of the circuit pattern 80, and the corner portion 82 of the circuit pattern 80. The reticle 6 is obliquely illuminated by one (or both) of the first illumination system 2 and the second illumination system 3,
When the diffracted light generated by the circuit pattern 80 is condensed by the objective lens 41, as shown in FIG. 5, the circuit pattern 80 on the reticle 6 and the projection image 60 on the surface of the reticle 6 of the illumination system 2 or 3 are projected. The angle θ defined by the positional relationship is 0
The angle pattern at the time of °, that is, the diffracted light of the angle pattern in the direction orthogonal to the illumination direction (hereinafter referred to as the 0 ° pattern) is shown in FIG.
As shown in FIG. Here, the types of the angle θ of the circuit pattern 80 are limited to the angle patterns of 0 °, 45 °, 90 °, and as shown in FIG. Since the diffracted lights (b) and (c) are not incident on the objective lens 41,
It is not detected by the detectors 51 and 551. On the other hand, since the scattered light from the foreign matter 70 has no directivity, the objective lens 4
On the Fourier transform plane of No. 1, it spreads over the entire surface as shown in FIG. Therefore, if linear linear filters 44 and 444 each having a band-shaped light-shielding portion and a transmission portion outside the Fourier-transformed surface are arranged, the spatial filter 44,
The light-shielding portion 444 can shield the diffracted light (a) from the 0 ° pattern as shown in FIG. 4A, and the scattered light from the foreign matter 70 passes through the light-transmitting portion, so that the imaging lens 4
Since an image is formed on the detectors 51 and 551 through 5,445,
The foreign material 70 can be detected by being discriminated from the circuit pattern 80. As a result, a high NA detection optical system can be realized, and when the NA is selected to be 0.5, for example, the aperture area can be about 20 times that of the low NA detection optical system. In that case, the diffracted light from the corner portion 82 of the circuit pattern 80 may not be sufficiently shielded by the linear spatial filters 44 and 444 as shown in FIG. 4D.
Therefore, when the detection is performed by the detection pixel of 10 × 20 μm ² as in the conventional technique, the diffracted light from a plurality of pattern corner portions enters the pixel as shown in FIG. 4B. , It is not possible to detect only foreign matter. Therefore, in this embodiment, as shown in FIG. 4C, the resolution of the pixels of the detectors 51 and 551 is increased to 2 × 2 μm ² , and the influence from the circuit pattern 80 is eliminated as much as possible.
It is possible to detect foreign matter of about 5 μm. Although the pixel of the detector is set to 2 × 2 μm ² , the reason is as described below, and it is not always necessary to set it to that value.
That is, the pixel size in this case may be smaller than the smallest pattern size on the reticle 6, and therefore 16M
1/5 reduction rate of 0.5 μm process LSI such as DRAM
In the reticle for exposure with the stepper, the minimum line width of the circuit pattern is about 0.5 μm, so 0.5 × 5 =
In the case of 2.5 μm or 64 MDRAM, the minimum line width of the circuit pattern is approximately 0.4 μm, and therefore it is sufficient to detect with a pixel smaller than 0.4 × 5 = 2 μm. Actually, if the value from the pattern corner is small enough,
It may be larger or smaller. In particular,
The minimum pattern size on the reticle to be inspected is desirable. If the size is about this minimum pattern size,
Only one corner of the detector has less than two corners, and the experimental results shown in FIG. 10 show that this value is sufficient. More specifically, the minimum size on a reticle with a reduction ratio of 1/5 is about 2 μm, which is 64 μm.
In the reticle for MDRAM, the pixel size of about 1 to 2 μm is desirable. The reason for this will be described in detail with reference to FIG.
Microscopically looking at the corner portion 82 formed at the intersection in the circuit pattern 80 shown in FIG. 7A, FIG.
As shown in FIG. 5, the corners 820 are formed at continuous angles. Therefore, the diffracted light from the corners 82 also tends to spread on the Fourier transform plane, similar to the case of scattered light from foreign matter. Therefore, the linear spatial filters 44 and 444 cannot completely shield the light, and as a result, it becomes as shown in FIG. 6 (B). Therefore, when the diffracted light from the plurality of corners 82 enters one detector 51 or 551, the output V of the detector increases, so that the detection with the foreign matter 70 cannot be performed. FIG. 8 shows the state, in which the detection output value 822 from the plurality of corner portions 82 is higher than the detection output value 821 from the single corner portion 82, and the binary value is at the level of the dotted line 90. The output value 70 detected from the foreign matter 70
This means that 1 cannot be detected separately. Therefore, when diffracted light from a plurality of pattern corners enters one pixel, it cannot be detected depending on the value of level 90. As a countermeasure against this problem, in the present embodiment, the inspection field of view 15 on the reticle 6 is set to the objective lens 41,
Detectors 51, 551 via the imaging lenses 45, 445, etc.
And the imaging magnification is selected for the detectors 51 and 551 so that the diffracted light from the plurality of corners 82 causes the detection field of view 15 on the surface of the reticle 6 to be detected. The dimensions (for example, 2 × 2 μm ^{2 as} described above) are set so that they are not simultaneously incident on 551, so that even the simple detection optical system 4 can discriminate the scattered light from the foreign matter.

However, even if it is possible to detect a foreign substance having a size as in the conventional case, in the detection of a foreign substance on the order of submicron, depending on the shape of the circuit pattern 80, it may not be detected separately from a part of the corner portion 82. However, due to the high integration of the LSI, a circuit pattern having a dimension 84 of a micron order finer than the gap 83 of the normal structure portion of the circuit pattern 80 is formed as shown in FIG. In this case, the behavior of the diffracted light generated from the circuit pattern is similar to that of the scattered light from the foreign matter 70, so that it becomes more difficult to detect the foreign matter 70 separately from the circuit pattern. In the present embodiment, the measures described below are also applied to the circuit pattern having the dimension 84 of the micron order as shown in FIG. 9 to detect the foreign matter. In FIG. 10, reference numerals 701 and 702 denote scattered light detection output values from the minute foreign matter 70 on the order of submicrons. 864,8
74,865,875,866,876,867,87
Reference numeral 7 is the detection output value of the diffracted light from all the corner portions 82 formed in the angle patterns of 0 °, 45 ° and 90 °.
Reference numerals 861, 871, 862, 872, 863 and 873 respectively indicate detection output values of diffracted light from a fine structure circuit pattern having a dimension 84 of the micron order. Of these, 7
01,861,862,863,864,865,86
Reference numeral 6,867 denotes the detection output value of the first illumination system 2, and 702,871,872,873,874,875,8.
Reference numerals 76 and 877 respectively indicate detection output values by the second illumination system 3. For example, 861 ← → 871 are detection output values by illumination system at the same position of the circuit pattern, and 861 is a value by the first illumination system 2. 871 indicates the values of the second illumination system 3, respectively. Further, as can be seen from the figure, the foreign matter 70 has a smaller variation in the detection output value of scattered light depending on the irradiation direction than the circuit pattern. The dotted line 91 in the figure indicates the threshold value of the detection output value. From FIG. 10, it was found that even with the same circuit pattern, the output of the diffracted light increased depending on the irradiation direction, and furthermore, the surface of the reticle 6 was illuminated from two oblique directions opposite to each other with a 180 ° shift. In this case, it can be seen that the output value of either one of the diffracted light beams is always smaller than the output value from the foreign matter on the submicron order, as indicated by the ● mark in the figure. In the embodiment, the reticle 6
The above output values from the same position above are detected by the detectors 51 and 55.
1 is detected separately, the smaller detection output value indicated by the mark ● is adopted, and after binarized by the binarization circuits 52 and 552, a logical product is taken by the logical product circuit 53. , Submicron order foreign material 70 only circuit pattern 80
Separated from and detected. In that case, as shown in FIG.
When the threshold value 91 is set in the binarization circuits 52 and 552, the values above the threshold value 91 are detected output values 701, 7 of the foreign matter 70.
02 and detection output values 861, 863, 8 of the circuit pattern
74,875, but 2 from these circuit patterns
Since the binarization output is the output from only one of the binarization circuits 52 and 552, it is not output from the AND circuit 53, and therefore only the foreign matter 70 can be detected separately from the circuit pattern. . Also, since the amount of diffracted light from the circuit pattern is small and the amount of scattered light from foreign matter is large,
When such a logical product detecting means is not necessary, if a logical sum circuit is used instead of the logical product circuit 53, the detection result by illumination from two directions can be obtained, and more stable detection can be performed. Further, at that time, if only detection sensitivity enough to use an OR circuit is always required, a wavelength separation mirror 42, a linear spatial filter 444, a variable spatial filter 447, an imaging lens 445, a detector. It is also possible to omit the 551 and binarization circuits 552 and obtain equivalent performance with a simpler configuration. Then, when the scattered light from the foreign matter is detected, the microcomputer 54 detects the position information of the X stage 10 and the Y stage 11 at the time of detection,
The position information of the foreign substance 70 calculated from the pixel position in the element of the detector 51 (551) and the detection output value of the detector are stored in the memory as foreign substance data, and the stored contents are subjected to arithmetic processing to perform CRT or the like. Is displayed on the display means 55.
As a result, the status of the foreign matter 70 is displayed.

FIG. 12 shows a conventional technique of analyzing the deflected state of scattered light and a conventional technique of comparing the outputs of a plurality of detectors, both of which show diffracted light generated from a circuit pattern. In order to avoid the influence, an optical system having a small aperture of NA 0.1 is arranged in a direction avoiding the diffracted light from the circuit pattern. In such a conventional configuration,
This causes a problem that it is easy to overlook an irregularly shaped foreign substance. That is, the NA used in the related art is a numerical value that is determined by the aperture diameter of the lens and the distance to the target object and expresses the characteristics of the lens. Specifically, θ shown on the right side of the figure is used, and NA
= Numerical value obtained by sin θ. Another problem is a pattern removal technique that is used as an auxiliary in various inspection techniques in response to the miniaturization of circuit patterns. Many pattern removal techniques automatically reduce the detection sensitivity of the foreign matter detector when a circuit pattern is found during inspection. Such a method can reduce false detection of the circuit pattern. However, on the other hand, Fig. 11
As shown in (A) to (C), there arises a problem that foreign matter attached near the edge of the circuit pattern is missed. The solution measures of the present invention for these two problems will be described below. Photograph parts 1004 and 1005 in FIG.
Is an observation of scattered light generated when a foreign substance is irradiated with a laser from above. This photo section 1004,10
What should be noted in 05 is that the scattered light (e) from the foreign matter has a directional distribution. For this reason,
In the conventional low NA detector 1001, if the detector is not installed at an appropriate position as shown in FIG. 13, the scattered light (e) generated from the foreign matter may enter the low NA optical system in a proper condition. Not to be overlooked. Moreover, since the distribution of the scattered light varies depending on the size and shape of the foreign matter, it is virtually impossible to properly arrange the optical system having a low NA for all the foreign matter. The result of experimentally measuring this is shown in FIG. FIG. 13 shows a scattered light distribution when a foreign substance is illuminated with laser light having an incident angle of 60 °,
Low NA (NA = 0.1) detection optical systems 1001, 10
The level of scattered light from the foreign matter was measured and indicated while changing the detection angle at 02. The figure shows that the detection level at the point A by the detection optical system 1001 exceeds the detection threshold, whereas the detection level at the point B by the detection optical system 1002 does not exceed the detection threshold and cannot be detected. Since the actual scattered light distribution of foreign matter is not uniform,
It is shown that the detection performance is not stable with such a low numerical aperture detection method. Therefore, in this embodiment, the high NA detection optical system having a large aperture is used to effectively collect the scattered light from the foreign matter having various scattering distributions. That is, as shown in FIG. 15, a high NA objective lens 41 is used as a part of the detection optical system, and the high NA objective lens 41 collects scattered light from foreign matters having various scattering distributions. I am trying to let you. However, in this case, if the NA of the objective lens 41 is too large, diffracted light from the pattern enters, so it is important to select the size of the NA appropriately. FIG. 16 shows the result of foreign matter inspection when the high NA objective lens 41 is used in the apparatus. In FIG. 16, the total of the foreign matters 70 detected by the five reticles 6 is shown on the vertical axis, and the dimension of the detected foreign matters is shown on the horizontal axis.
In addition, among the foreign substances, the foreign substances detected by the conventional technique are shown in different colors. From the figure, the detection capability of the prior art was 0.8 μm. Therefore, it can be understood that there is a difference from the present invention in the detection capability in the area of foreign matter smaller than 1 μm. However, even in the area of foreign matter larger than 1 μm, the number of detected particles is significantly improved in the present invention. This detection rate is approximately 1 in the ratio of the number of detections in the prior art.
It will be 0 times. This is because the high NA detection optical system adopted by the present invention corresponds well to irregularly shaped foreign matter and stably detects scattered light from the foreign matter. Note that FIG.
In Fig., Although the variable spatial filter is not shown,
It is assumed that they are arranged at substantially the same position as the linear spatial filter 44. Next, the state of detection of foreign matter attached to the edge portion of the circuit pattern will be described with reference to FIG. FIG. 17 shows the result of classifying the foreign matters detected in FIG. 16 by the attachment position. The reticle 6 is used as the foreign matter attachment position.
The upper glass portion (transparent portion), the chrome portion of the circuit pattern 80 (light-shielding portion that is often formed of a metal thin film such as chrome), and the circuit pattern 8 that is the boundary portion between them.
It was classified into three areas of 0 edge portion. From FIG. 17, it can be seen that since the foreign matter adheres much at the edge portion, the edge portion is greatly affected by the foreign matter adherence. Further, although the foreign matter adheres relatively little to the chrome portion, it does not affect the transfer as long as the foreign matter remains on the chrome portion.

Although the foreign matter inspection apparatus of the above embodiment is applied to the reticle having the chrome portion as the circuit pattern, the transfer resolution between the chrome portions as the circuit pattern 80 is increased as shown in FIG. The same can be applied to the reticle 6 having a pattern (shifter pattern) 1003 formed of a phase shifter film for the purpose of improvement. That is, although the shifter pattern 1003 is transparent, it is several times as thick as the chrome portion (thickness is about 0.1 μm) (when the exposure light in the stepper is the i-line of the mercury spectrum, the thickness is about 0.4 μm).
Since the structure has a size of about m), the diffracted light from the edge portion is considerably larger than the diffracted light from the edge portion of the chrome portion. Therefore, there is a possibility that the foreign matter cannot be detected. However, the exposure technique using the shifter pattern 1003 basically uses the repeatability of the circuit pattern, and conversely, by utilizing the repeatability of the circuit pattern, that is, the repeatability of the shifter pattern 1003. The scattered light from can also be removed. When an image illuminated by a coherent light source having a wavelength λ is subjected to Fourier transform by a Fourier transform lens having a focal length f with respect to a circuit pattern having a repeating period T, it is discretely condensed at a position of an integral multiple of λf / T. Will be done. On the other hand, non-repeatable foreign matters are widely distributed. Therefore, if only the image of the foreign matter can be detected by blocking the light-condensing position of the circuit pattern image on the Fourier transform surface with a filter as in the prior art, many proposals have been made in the past. However, in the prior art, the following two are common. (1) Illumination is performed by a parallel beam in order to increase the degree of the focus position of the Fourier transform image. (2) Inverse Fourier transform is performed after Fourier transform, and in order to restore a complete image when restoring (re-imaging) the image on the inspection sample, epi-illumination or transmitted illumination perpendicular to the sample surface is performed. The diffracted light from the illumination including the 0th-order light is incident on the Fourier transform lens. However, if it is only necessary to detect the foreign matter, that is, if it is possible to determine the presence or absence of the foreign matter, it is not necessary to perform the inverse Fourier transform, and the re-imaging image after the inverse transformation is complete. There is no need at all. In this embodiment, as described above, the detection optical system 4 has the spatial filter including the linear spatial filter and the variable spatial filter, and the incident angle θ is 20 ° ≦ θ ≦ 80 ° with respect to the reticle 6. Since the first and second illumination systems 2 and 3 are configured to perform obliquely, the illumination light that is remarkably brighter than the diffracted light of the first or higher order (generally tens to hundreds of times or more) can be obtained. The specularly reflected light (0th order diffracted light) does not enter the detection optical system 4. Therefore, it has the advantages described below. First, as compared with the prior art, it is not necessary to provide a light-shielding portion that shields the 0th-order diffracted light on the spatial filter, so that the proportion of the light-shielding portion in the aperture (NA) of the detection optical system 4 is reduced. Since it is possible to secure a substantial NA, 0.
It is possible to detect foreign matter in the submicron order, which is required after the 5 micron LSI, with stability. Secondly, the influence of the positional deviation of the spatial filter is reduced as compared with the conventional technique in which the 0-th order diffracted light that is extremely bright is incident on the detection optical system. This is not only an important advantage in the present method of changing the position of the spatial filter adapted to the shape of the circuit pattern 80, but also an unstable phase shift of the film shape as compared with the circuit pattern formed by the metal thin film. Enables detection of foreign matter on the reticle. Thirdly, it is not necessary to illuminate with a parallel beam in order to narrow the range of the 0th-order diffracted light as much as possible, and it is possible to realize illumination with a focused beam capable of detecting foreign matter on the order of submicron with low S / N. Therefore, even if this device is applied to a reticle having a phase shifter pattern, not only can foreign matter be detected reliably, but also
Stable detection can be secured.

By the way, there is also a demand for observing a complete image of a foreign substance. In such a case, although not shown in detail in the apparatus of the present embodiment, the detection position of the foreign matter is stored in the microcomputer 54, the position is reproduced by the display means 55 at the time of observation, and it is incorporated in the apparatus. It is also configured to observe with oblique illumination. In that case, for observation, white illumination such as a halogen lamp is cheaper and simpler than coherent light such as laser illumination, and there is no influence of interference. However, this is not the case when the configuration of the confocal microscope is required to enhance the resolution during observation.

Next, a modification of the spatial filter is shown in FIG.
9 to 34. The embodiment shown in FIG. 19 is an example in which the photographic dry plate 1901 is used as the spatial filter because the spatial filter acts as a spatial light modulator. The most convenient is a photographic dry plate. in this case,
Before the inspection, the photographic plate is placed on the Fourier transform surface in an unexposed state, exposed to scattered light (diffracted light) from the pattern to be inspected, and then taken out, and after development, the photographic plate is again Fourier-transformed. Used by positioning on the conversion surface and returning. When a negative type photographic plate is used as the photographic dry plate, the portion exposed to the scattered light from the circuit pattern becomes the light-shielding portion as it is. For example, using an instant camera type negative film simplifies the development process. For positioning after development, as shown in the figure, a holder 1902 of the photographic dry plate 1901 is provided with a positioning means 1903 composed of positioning pins and the like, and the photographic dry plate 1901 is positioned by using this means. Will be easier to position. The embodiment shown in FIG. 20 uses means for confirming the position condition of the variable spatial filter 47 (447). That is, the half mirror 2002 is arranged behind the Fourier transform surface 2003 in the variable spatial filter 47, the insertion mechanism 2001 for tilting the half mirror 2002 is provided, and the half mirror 2002 is positioned at a desired position by driving the insertion mechanism 2001. Then, the optical path branched by the half mirror 2002 can be observed with the eyes 2004, and the positioning state of the variable spatial filter 47 can be confirmed. This insertion mechanism 2001 and half mirror 2002
Is used when positioning the variable spatial filter 47, and is removed when the positioning is completed.

Alternatively, as an alternative to direct visual observation,
As shown in FIG. 21, if the TV camera 2102 captures an image of the position of the variable spatial filter 47 via the imaging lens 2101, the positioning situation can be surely grasped. Such a means for confirming the positional condition of the variable spatial filter 47 is used by inserting the variable spatial filter 47 only at the time of observation without branching the optical path by the mirror 2002 when sufficient space exists behind the filter. You can also do it. 20 and 21, if the position moving mechanism 2005 that moves the position of the variable spatial filter 47 is used, delicate position adjustment can be easily performed. In the embodiment shown in FIG. 22, the holder 2201 of the photographic dry plate 1901 is formed in a water tank shape made of an optically transparent material such as glass, and a developing solution, a stop solution, a fixing solution, and a washing solution are supplied to the holder 2201. The supply path 2202 is connected to the discharge path 2203 for discharging the liquid so that development can be performed without removing the photographic dry plate 1901. In this case, since it cannot be dried, if the aquarium-shaped holder 2201 is exposed to light with the developer filled at the time of photographing, it can be used with the liquid filled at the time of inspection, and the drying is unnecessary. In this case, if the holder 2201 is made of an optically transparent material, not only the exposure becomes easy, but also the exposed state of the photo signboard 1901 can be confirmed from the outside.

FIG. 23 shows a variable spatial filter 47 (44).
As 7), a spatial light modulator that does not require development is used. That is, in this embodiment, the glass 2301 made of a photochromic material is used as the spatial light modulator. This is a glass doped with silver salt, and its light transmittance can be reversibly changed by light irradiation, and it has a semi-permanent life. As shown in FIG. 24, the photochromic glass 2301 has a light transmittance TR (%) characteristic that the light transmittance is reduced to 20% (blackened) in about 1 minute by irradiation with light having a wavelength of 488 nm. When the irradiation of No. 2 is stopped, the original transmittance is restored (fading) in about 5 to 10 minutes. Further, the fading of the photochromic glass 2301 has a characteristic that it is promoted when overheated. By utilizing such a property, it is possible to detect a foreign substance by the following procedure. (1) Call the periodic pattern in the sample to be inspected into the inspection field of view. (2) Irradiation light is irradiated for about 1 minute while the inspection stage is stationary to expose the Fourier transform image of the periodic pattern on the photochromic glass 2301. As a result, the portion corresponding to the Fourier transform image of the periodic circuit pattern is blackened. (3) The sample to be inspected is scanned and inspected. In this case, the scattered light from the foreign matter is photochromic glass 2301.
Is detected by passing through the non-blackened area. Further, since the scattered light from the foreign matter is only detected momentarily during the scanning of the sample to be inspected, the photochromic glass is not blackened and the detection of the next foreign matter is not affected.
Moreover, since the diffracted light from the periodic circuit pattern occupying most of the area of the sample to be inspected continues to hit the photochromic glass 2301 most of the time during the inspection, the Fourier transform image of the periodic circuit pattern on the glass is blackened. The state is maintained. (4) At the end of the inspection, the photochromic glass 2301 that has been inspected is overheated by the overheating mechanism 2302 to be discolored, and the next inspection is started. In this example, the example in which the overheating mechanism 2302 is used to promote the fading of the photochromic glass 2301 has been shown, but it may be performed without using the overheating mechanism. That is, a plurality of photochromic glasses 2301 are provided on the same plane, and a position moving mechanism 2303 for moving these plural things is prepared, and the position moving mechanism 2303 retracts one of the photochromic glasses 2301 whose inspection has been completed. At the same time, if the other photochromic glass 2301 is moved to the inspection position and inspected, the image of the previously used portion is naturally fading during the inspection, so that the overheating mechanism 2302 becomes unnecessary. As described above, the use of the photochromic glass 2301 can be simplified as compared with the above-described embodiment using the photographic plate, because there is no development process. Also, when creating (exposing) a variable spatial filter,
If the image of the linear spatial filter is written on the variable spatial filter by multiple exposure or the like, both can be integrally formed, and the linear spatial filters 44 and 444 as in the embodiment shown in FIG. 1 can be formed. The fact that it is not necessary to use separately is the same as in the embodiment using a photographic plate.

In the embodiment shown in FIG. 25, a thermoplastic is used as the variable spatial filter. This is an element formed by sequentially laminating a transparent conductor 2502, a photoconductor 2503, and a thermoplastic resin 2504 on a transparent substrate 2501, and after charging the entire surface with a charger 2505,
By exposing and overheating, the thermoplastic resin 2504 is deformed according to the exposure pattern by electrostatic force and is regenerated, which forms a kind of electrophotography. An element having an image storage function, such as a thermoplastic, is called a photo-addressable spatial light modulator, and a photoconductor film (electrical resistance that changes in electric resistance depending on input light (a Fourier transform image to be recorded in the case of the present invention) ( CdS, amorphous Si, BSO, Z
nS, amorphous Se, PVK, etc., and a film (liquid crystal, electro-optic crystal (DKDP, LiNbo, PLZ) whose birefringence changes depending on the electric field intensity distribution formed by the film.
T, BSO, etc.), and the output light (detection light in the present invention) to be modulated is passed, the deflection state is modulated, and if the detector is passed, an amplitude-modulated image is obtained. Various elements have been proposed depending on the combination of the photoconductive film, its film thickness, and the birefringent material, and the same can be applied to the present invention. Further, when the variable spatial filter is created (exposed), if the image of the linear spatial filter is written in the variable spatial filter by multiple exposure, the linear spatial filters 44 and 444 shown in FIG. 1 can be changed. It can be formed integrally with the spatial filter. Among the elements having an image storage function such as the above thermoplastics,
In some devices, it is necessary to change the wavelength of recording light and the wavelength of detection light. That is, when an image illuminated by a coherent light source having a wavelength λ on a circuit pattern having a repeating period T is Fourier transformed by a Fourier transform lens having a focal length f,
Since the light is discretely condensed at a position of an integral multiple of f / T, in order to perform recording and detection at different wavelengths, the wavelength λ at the time of recording
1 and the wavelength λ2 at the time of detection, the Fourier transform lens is set so that the relationship between the focal length f1 of the Fourier transform lens at the time of recording and the focal length f2 of the Fourier transform lens at the time of detection is f1: f2 = λ2: λ1. Need to be replaced. In the case of foreign matter inspection, the optical properties change depending on the wavelength.
Since there are a wide variety of samples such as the refractive index reflectance and scattering characteristics of the sample to be inspected and the foreign substance, strictly speaking, it is difficult to match the Fourier transform images of the recording light and the detection light only by the focal length of the Fourier transform lens. Considering the purpose, the recording light and the detection light are configured to be emitted from the light source having the same wavelength. Further, since the birefringence, that is, the rotation phenomenon of the deflecting surface is used, special consideration is required for detecting the foreign matter by the deflecting illumination. FIG. 26
28 shows an example thereof. FIG. 26 shows a case where the reticle 6 is inspected by an unpolarized oblique illumination 2601 as an illumination system. In this case, since the detection light taken into the detection optical system is also in the non-deflected state, the analyzer 2603 linearly deflects the light before it enters the modulation element 2602, and the modulation element 260 is used.
2 and the analyzer 2604. Therefore, in this embodiment, the analyzer 2603 and the modulation element 260 are
2 and the analyzer 2604, the variable spatial filters 47 and 447.
Are configured. FIG. 27 shows a case where the reticle 6 is inspected by the oblique illumination 2701 of S deflection (linear deflection having a deflection surface perpendicular to the paper surface). The diffracted light from the circuit pattern preserves the P deflection, whereas the scattered light from the foreign matter changes part of the P deflection to S deflection, and as a result, the scattered light in which P deflection and S deflection are mixed is generated. Enter the detection optics. Therefore, as in the case of FIG. 27, at the rear position of the field lens 43,
The variable spatial filters 47 and 447 are configured by arranging the modulation element 2602 and the analyzer 2604 so that the analyzer 2603 shown in FIG. 26 is unnecessary and the P-polarized component before the incident is shielded. In FIGS. 26 to 28, since the recorded Fourier transform image is used as it is as the data of the spatial filter, the configuration is simple, but the spatial filter and the periodic pattern on the reticle to be shielded are exactly the same. , There is no room for alignment. Further, since the detected scattered light does not include the 0th-order light, it is more tolerant of positional deviation than the conventional method, but if higher sensitivity detection is desired, a margin can be provided by the optical processing of the following procedure. It is also possible to form a spatial filter having the same. That is, the Fourier transform image from the periodic pattern when illuminated by the focused beam is discretely focused as bright points, and the size of the bright points is such that the NA of the focused beam is large. It is large and small as NA is small. For example, when the focal length of the condenser lens is constant, as shown in FIG. 44, when a beam 3501 having a large aperture formed by a beam expander or the like is condensed by the condenser lens 3502 and used as illumination light,
The Fourier transform surface image 3504 from the periodic pattern has a large bright point 350 at a discrete point as enlarged and clearly shown in FIG.
It will be 3. On the other hand, as shown in FIG. 45, the operation of the lens system 3503 or the light shielding plate 360 having a small opening is performed.
5, when the beam 3601 having a smaller aperture is focused and used as illumination light, the Fourier transform surface image 3504 from the periodic pattern becomes discretely small bright spots 3603. Since this relationship depends on the NA of the condensing optical system as described above, even if the focal length of the lens is changed while keeping the beam aperture constant, or both the beam aperture and the lens focal length are changed. realizable. In this way, when the size of the bright spots is controlled, the size of the bright spots is increased when recording on the spatial filter, and when the size of the bright spots is decreased during the inspection, the difference in the size of the bright spots causes misalignment. There is a margin of error.

Next, an electrically addressed spatial light modulator can be used instead of the optically addressed spatial light modulator. In FIG. 29, a variable spatial filter 47 is configured by a liquid crystal spatial light modulation element 2901 that uses liquid crystal as an electrically addressed spatial light modulation element. That is, this liquid crystal spatial light modulator (Liquid Crystal Spatial)
al Light Modulator: Liquid crystal SLM,
Or Liquid Crystal Display:
The LCD) 2901 is considered to be a matrix of minute shutters of several hundreds × several hundreds, but since the birefringence characteristics of the liquid crystal are used, the relationship between the deflecting surface of the illumination light and the deflecting surface to be shielded by the liquid crystal is shown. The precautions are the same as for the optically addressed spatial light modulator described above. Liquid crystal SLM 29
Although each shutter 01 is controlled by an electric signal from the outside through the liquid crystal driver 2902, the liquid crystal SLM alone does not have a light detection function storage function. Therefore, the Fourier transform image to be recorded is the liquid crystal S as shown in FIG.
Insertion mechanism 2 inserted only at the time of recording behind LM2901
An optical path is branched by a mirror 2002 having 001, an image at a spatial filter position is formed on a TV camera 2102 or the like by an image forming lens 2101, and a Fourier transform image to be recorded is detected by the TV camera 2102.
After being temporarily stored in 03, it is sent to the liquid crystal driver 2902. Alternatively, as shown in FIG. 30, the optical path is branched by a mirror 2002 having an insertion mechanism 2001 that is inserted in front of the liquid crystal SLM 2901 only during recording, and a TV camera 2102 or the like is provided at a position conjugate with the spatial filter position for detection.
The Fourier transform image to be recorded may be temporarily stored in the image memory 2903 and then sent to the liquid crystal driver 2902. 29 and 30, a half mirror may be used instead of the mirror 2002. The liquid crystal driver 2902 inverts (negative) the detected image and outputs a control signal for displaying on the liquid crystal SLM 2901. Although the data can be inverted before being stored in the image memory 2903, in this case, the liquid crystal driver 2902 does not need the image inversion function, and the liquid crystal driver 2902 has only the function of sending a control signal. Therefore, if the liquid crystal SLM is composed of the liquid crystal SLM,
It has a liquid crystal driver 2902 for driving the LM, a TV camera 2102 for detecting a conjugate image of an image on the liquid crystal SLM, and an image memory 2903 as a storage unit for the detected image.

When the variable spatial filter is created (exposed), the method of writing the image of the linear spatial filter on the variable spatial filter by multiple exposure is the TV camera 2102.
It is generally not possible with such detectors. Therefore, the data of the linear spatial filter is stored in another area of the image memory 2903 and is combined with the data of the variable spatial filter by software, or when the data is sent to the liquid crystal driver 2902, superimposing or the like is performed. By synthesizing the data of the variable filter and the data of the linear spatial filter by the function of 1, the same linear spatial filters 44 and 444 shown in FIG. 1 can be integrally formed. Liquid crystal SLM
The alignment of the above spatial filter and the Fourier transform image from the periodic pattern is performed by moving the liquid crystal SLM 2901 by the position moving mechanism 2005 or by the TV camera 21 at the time of recording.
02 by the position moving mechanism 2105, or both of them. Alternatively, the image can be aligned by shifting the image on the image memory 2903 by electric means or software means. In the case of FIG. 29, the mirror 2 having the insertion mechanism 2001
002 or half mirror to split the optical path to TV
The camera 2102 can confirm the alignment status. Further, by processing the image on the image memory 2903, the size of the recorded bright spots is increased by an electric means or a software means, thereby making a margin of alignment error in the light shielding portion of the spatial filter. be able to. The advantage of such an electric address type spatial light modulator is that the leaf image converted image is stored in the image memory. Therefore, as shown in FIG. 31, once the image is recorded, it is stored in an arbitrary image memory 2903 unless it is erased. The above image data 3001, 3002, 3003 ... Can be switched at any timing. Therefore, it is not necessary to recreate the periodic pattern for which image data has been created once. This has the advantage that when the linear spatial filter is omitted by combining the data as described above, the data of the linear spatial filter need only be adopted once. Further, even when there are regions having different periods in one reticle to be inspected, it is possible to cope with this by switching the image data as shown in FIG. At that time, when switching based on the reticle design data 3201 and when the rate of change 3202 in the number of detected foreign substances is monitored and the number exceeds a set value (set in the binarization circuit 3203) and suddenly increases. , It is determined that the spatial filter is no longer suitable, and the image data 3001, 300
2, 3003 ... may be switched. Then, the image data 3001, 3002, 3003 on the image memory
It is possible not only to switch the ..., but also to suspend the inspection and replace the spatial filter by newly capturing image data. Furthermore, as shown in FIG. 33, reticle design data 3301, oblique illumination wavelength 3302,
NA value 3303 of oblique illumination, direction 3304 of oblique illumination
The reflectance of the substrate glass of the reticle for oblique illumination light 33
05, the reflectance of the reticle circuit pattern portion with respect to the oblique illumination light 3306, the film thickness 3307 of the reticle circuit pattern portion, and the NA value 3308 of the detection lens, the Fourier transform image is obtained by the computer 3309, and the image memory 290 is obtained.
3 may be configured to generate image data. FIG. 34
[Fig. 4] is an example of a detection result according to the embodiment of the present invention. In the figure, the vertical axis V indicates the detection value (relative value) of scattered light, and when the detection value from the foreign matter is larger than the detection value from the circuit pattern, the foreign matter cannot be detected. The solid line shows the detection value from the foreign substance (size 0.5 μm), the broken line shows the detection value from the periodic pattern with the phase shifter, and the dashed line shows the detection value from the aperiodic pattern. Further, the horizontal axis indicates the detection condition, A is a case where no spatial filter is used, B is a linear spatial filter only, C is a variable spatial filter only, and D is a variable spatial filter and a linear spatial filter. It is a detection result when the spatial filter which consists of both is used. From the figure, the linear spatial filter alone cannot support the phase shifter, and the variable spatial filter alone cannot support the non-periodic pattern. Therefore, by using both the variable spatial filter and the linear spatial filter, , It can be seen that the entire surface of the reticle can be detected with high sensitivity.

On the other hand, it has been described above that the pixel size of the detector 51 for detecting foreign matter may be about the same as the smallest reticle-like pattern size, but when an array type detector 51 is used, Causes the following inconveniences. That is, in the case where the foreign matter is detected by the pixel size of the detector of 2 × 2 μm ² and the determination is made as an example, as shown in FIG. 35, the foreign matter 70 is spread over a plurality of (2-4) pixels. May be detected. In such a case, the foreign matter 70
The scattered light from is also dispersed into a plurality of pixels, and as a result, the detection output of one pixel is reduced to 1/2 to 1/4, and the detection power of the foreign matter 70 is reduced accordingly. Actually 1 due to the effect of crosstalk between detector pixels
/ 3). In addition, the pixels of the detector and the minute foreign matter 7
The positional relationship with 0 is very delicate due to its size, and changes with each inspection. Therefore, even if the same sample is used, the results will differ from inspection to inspection, and the reproducibility of detection will deteriorate. Therefore, FIG.
As shown in, the detection pixel is reduced to 1 × 1 μm ² ,
The detection outputs of four adjacent 1 × 1 μm ² pixels of each pixel are electrically added to simulate the detection output of 2 × 2 μm ² pixels. This value is duplicatively obtained by 1 μm (a, b, c, d in the figure), and the obtained maximum value a is 2 × 2.
The detection determination of the foreign matter 70 is performed as a representative output by μm ² pixels. Therefore, a microcomputer is provided with a function of a four-pixel addition process capable of representatively outputting the maximum value obtained by electrically adding the detection outputs of four adjacent pixels and simulating the detection output of 2 × 2 μm ² pixels. Put it on 54. As a result, the fluctuation of the detection output from the same foreign matter is actually within the range of ± 10%, and the detection reproducibility of 80% or more can be secured for all the foreign matter. FIG. 16 shows a result (detection reproducibility of 80% or more) when the 4-pixel addition processing function is applied. Incidentally, FIG. 37 shows an example of the detection reproducibility to which the 4-pixel addition processing function is not applied, and it is clear from this figure that sufficient detection reproducibility cannot be ensured without the 4-pixel addition processing.

Here, a specific example of the 4-pixel addition processing circuit will be described with reference to FIG. In this example, the detector 51 is 1 μm
This is a one-dimensional type image pickup device in which 512 pixels when reduced to m are arranged, and an output 2503 from an odd-numbered pixel and an output 2502 of an even-numbered pixel of the one-dimensional type image pickup device are separately output. It is a general one. 256-stage shift register 2501, 1-stage shift register 2504, and 1 pixel (1 μm reduced by adders 2505 to 2508).
4 pixels (2 × 2 pixels) shifted in 4 directions by m) are added, and the average value of each average value is obtained by the dividers 2509 to 2512. Then, the maximum value determination circuit 2513 finds the maximum value in the four directions, and outputs the maximum value as the detection value 2514 from the foreign matter. In this example, only the foreign matter is made bright and visible by optical processing, and when the detected signal is larger than the threshold value set for the detection, the foreign matter is detected by determining that there is a foreign matter. It is the one. However, in the detection signal at that time, (1) variations in sensitivity characteristics (about ± 15%) among pixels of the one-dimensional image sensor detector, and (2) sensitivity unevenness due to illuminance distribution of the illumination light source ( Shading) exists. As a result, as shown in FIG. 39, the size of the detection signal varies depending on the detection pixel (position in the Y direction) even for the same foreign matter, and there is a possibility that the foreign matter cannot be stably detected by binarization with a threshold value. Therefore, in the embodiment, FIG.
As shown in (A), the shading including the sensitivity unevenness for each pixel was measured with the standard sample 111 shown in the embodiment of FIG. 1 in advance, and then, as shown in (B) of FIG. The shading correction data obtained by calculating the reciprocal of the measurement data is obtained, the amplifier gain of the detection signal of the detector is changed for each pixel by this, and the influence of shading is eliminated as shown in FIG. I'm trying to detect. The standard sample 111 is placed on the inspection stage unit 1 as in the embodiment shown in FIG. 1 or installed in the vicinity of the inspection stage unit 1. However, the standard sample 111 is placed on the sample table instead of the reticle only during shading measurement. A configuration in which it is placed is also possible. The standard sample 111 may be any one as long as it has fine irregularities on the surface and has uniform scattering characteristics. For example, a glass substrate with fine processing marks is polished, or aluminum is sputtered on the substrate. A film having a thin film with minute irregularities such as a film is used. However, if the minute unevenness on the standard sample 111 is 1 × 1μ pixel
Since it is practically difficult to process uniformly with respect to m ² , the shading measurement is performed many times, for example, 1000 times.
The correction data is obtained from the average value that is repeated twice. In addition, since scattered light from minute unevenness has unevenness in intensity, a simple average value indicates that repeated data of 1000 times is 10 times, for example.
If it is divided by 00, the value becomes too small, and the accuracy of the calculation may decrease. Under these conditions,
The value to be divided may be set to a value that is a fraction of the number of times of repetition, and is set to 200 after 1,000 times of repetition. By doing this, when the shading before correction and the shading before correction are compared, as shown in FIGS.
It can be seen that the shading that was present at about 0% is corrected to 5% or less. It should be noted that if the above-mentioned correction data is remeasured / updated for each inspection, the optical fluctuation component can be surely removed even if the illumination / detector and the like are temporally unstable.

Next, a concrete example of the shading correction circuit will be described with reference to FIG. In the figure, the detection value of the one-dimensional image sensor is A / D converted (here, 256 gradations, 8 bi
The subtractor 3209 subtracts the value of the dark current portion of the one-dimensional image sensor from the value 3212 obtained in t). The subtractor 32
Reference numeral 09 denotes a synchronization circuit 3205 for each pixel of the one-dimensional image sensor.
Subtract based on memory 3206 controlled by. Next, the subtracted shading correction magnification is multiplied by the multiplier 3219. The multiplier 3219 is a memory 32 controlled by the synchronization circuit 3205 for each pixel.
Based on the data from 07. Then, the multiplication result is returned to the original bit by the middle-order bit output circuit 3211. Therefore, the middle bit output circuit is
The result of multiplication in which the number of bits (16 bits) is twice the detection value 3212 of the one-dimensional image sensor that has been D / D converted is the original bi.
It is set back to t (8 bits). As can be seen from the figure, although the correction is performed by digital processing in this example, it goes without saying that the same result can be obtained even if the correction is performed in an analog manner before the A / D conversion. For example, 2 x
In the case of the pixel unit of 2 μm ² , if a foreign substance having a size of 2 μm or more exists, the number of pixels detecting the foreign substance is different from the actual number of foreign substances. 10μ
When there is one foreign matter of about m, (10μm / 2μ
m) ² = 25 pixels will be detected,
In this state, if you try to observe the detected foreign matter,
It is necessary to confirm all five detection results, which causes inconvenience. In the prior art, this inconvenience was avoided by the grouping processing function that checks the connection relationship between pixels that have detected a foreign substance and determines that "one foreign substance has been detected" when the pixels are adjacent. However, this conventional technique requires software-like processing, and thus when a large number of detection signals are involved, a large amount of time, for example, about 10 minutes is required for 1000 detection signals, which causes inconvenience. Therefore, in the embodiment, a visual field range (for example, 3
It is divided into blocks of 2 × 32 μm ² ) and the block processing is performed in which all detection signals in the same block are determined as the same foreign matter and processed. This way,
Regardless of the shape of a large foreign object, it is possible to put it in the field of view once and observe and confirm it. The block process is a simple grouping process in terms of function, but has a feature that it can be easily implemented as hardware. In the embodiment, the processing is performed in real time by implementing the block processing in hardware, and the throughput of the apparatus including the inspection time can be significantly improved. For example, the detection signal 1
In the case of 000 pieces, it could be shortened to 2/3 or less as compared with the conventional technique. FIG. 43 shows the processing contents for determining foreign matter detection based on the detection signal. Next, a specific example of the block processing circuit is shown in FIG. In the figure, not only the determination of the same foreign matter but also the number of detection signals that are the basis of the determination can be classified and counted according to a preset large / medium / small threshold value. You can also know the maximum value of. From these data, it is devised so that it is possible to estimate the approximate size of a foreign substance and the situation in which a plurality of foreign substances are contained in the same block. A circuit for outputting an inspection stop signal when the number of signals for detecting foreign matter reaches a preset number is also incorporated. As described above, if the lock process is performed for each block, it becomes possible to make the determination criterion of the foreign matter finer, and the foreign matter can be detected and determined more precisely. In the illustrated embodiment, the detectors 51, 55
As an example, the example in which the one-dimensional solid-state image sensor is used is shown because of the advantage that the detection field of view can be widened while maintaining the resolution. Of course, it can be used.

[0038]

As described above, the claims 1 and 2 of the present invention are as follows.
According to Nos. 2 and 3, the sub-micron order fine foreign matter adhering to the circuit pattern substrate such as the photomask or reticle can be detected by the spatial filter, so that the foreign matter can be detected easily, easily and stably. According to claim 4, not only the same effects as in claims 1 to 3 are obtained, but also in a reticle having a phase shift film particularly for the purpose of improving transfer resolution. It can be dealt with easily and is beneficial.

[Brief description of drawings]

FIG. 1 is an overall configuration diagram showing a first embodiment of a foreign matter inspection device according to the present invention.

FIG. 2 is an explanatory diagram showing a reticle inspection state by scanning of X and Y stages.

FIG. 3 is an explanatory diagram showing a configuration example of an illumination system for a sample.

FIG. 4 is an explanatory diagram showing a relationship between scattered light and diffracted light from respective angle patterns in a circuit pattern and a linear spatial filter.

FIG. 5 is an explanatory plan view showing the relationship between the illumination direction of the illumination system and the angle pattern of the circuit pattern.

FIG. 6 is a Fourier transform image (a) showing diffracted light from an angle pattern parallel to the irradiation direction of the illumination system in the circuit pattern (a), a Fourier transform image (d) showing diffracted light from each pattern corner portion in the circuit pattern, Fourier transform image (e) showing scattered light from the circuit pattern.

FIGS. 7A and 7B are explanatory views (A) of an essential part showing a corner portion of a circuit pattern and an enlarged explanatory view (B) of an “A” portion in FIG.

FIG. 8 is an explanatory diagram showing detection output values of scattered light from a foreign substance and diffracted light from respective corners.

FIG. 9 is a diagram showing the shape of a circuit pattern having a fine structure.

FIG. 10 is an explanatory diagram showing an output value level of a detection signal detected from a foreign matter and a corner portion of a circuit pattern.

FIG. 11 is an explanatory view showing foreign substances missed in the conventional technique.

FIG. 12 is an explanatory diagram showing a foreign matter detection method in the related art, its problems, and detection performance.

FIG. 13 is an explanatory diagram showing a detection optical system according to an example and a conventional technique.

FIG. 14 is an explanatory diagram comparing the detection of scattered light between the case of using the high NA detection optical system of the example and the case of using the conventional low NA detection optical system.

FIG. 15 is a configuration diagram showing a main part of FIG.

FIG. 16 is an explanatory diagram comparing the relationship between the size of detected foreign matter and the number of detected foreign matter in each of the embodiment and the prior art.

FIG. 17 is an explanatory view showing the detected foreign matters classified according to the foreign matter adhesion positions according to the embodiment.

FIG. 18 is an explanatory cross-sectional view of a main part showing a reticle having a pattern of a phase shift film.

FIG. 19 is a perspective view using a photographic plate as a variable spatial filter.

FIG. 20 is a schematic configuration diagram of positioning status confirmation showing another example as a variable spatial filter.

21 is a schematic configuration diagram showing another example for confirming the positioning situation in FIG. 20. FIG.

FIG. 22 is a schematic configuration diagram showing another example using a photographic plate as a variable spatial filter.

FIG. 23 is a schematic configuration diagram using a photochromic material as a variable spatial filter.

FIG. 24 is an explanatory diagram showing characteristics of a photochromic material.

FIG. 25 is a schematic configuration diagram in which a thermoplastic is used as a variable spatial filter.

FIG. 26 is an explanatory diagram of an example of inspecting a foreign substance by non-deflected illumination.

FIG. 27 is an explanatory diagram of an example in which foreign matter is inspected by S-polarized illumination.

FIG. 28 is an explanatory diagram of an example of inspecting a foreign substance by P-polarized illumination.

FIG. 29 is an explanatory diagram when a liquid crystal spatial light modulator is used as a variable spatial filter.

FIG. 30 is an explanatory diagram showing another example in which a liquid crystal spatial light modulator is used as a variable spatial filter.

FIG. 31 is an explanatory diagram showing an example of switching of image data of a liquid crystal spatial light modulator.

FIG. 32 is an explanatory diagram showing another example of switching of image data, which corresponds to a reticle having a different cycle.

FIG. 33 is an explanatory diagram corresponding to various data showing still another example of switching of image data.

FIG. 34 is an explanatory diagram showing inspection results depending on the presence / absence of a spatial filter.

FIG. 35 is an explanatory diagram when a foreign substance is detected across a plurality of detection pixels of the detector.

FIG. 36 is an explanatory diagram when performing a 4-pixel addition process on a foreign substance extending between detection pixels.

FIG. 37 is an explanatory diagram of the relationship between the number of detected foreign matters and the detected foreign matter size, which shows the reproducibility of the foreign matter detection result when the 4-pixel addition processing is not performed.

FIG. 38 is a block diagram showing a specific example of a 4-pixel addition processing unit.

FIG. 39 is an explanatory diagram showing sensitivity unevenness (shading) in which the detection signals of the same foreign matter have different magnitudes depending on the pixels to be detected.

FIG. 40 is an explanatory diagram sequentially showing a state in which the measured shading is corrected.

FIG. 41 is a block diagram showing a specific example of a shading correction circuit.

FIG. 42 is a block diagram showing a block process of dividing the entire inspection visual field into blocks and performing a confirmation process of foreign matter detection.

FIG. 43 is a block diagram showing a process of determining foreign matter detection based on a detection signal from a detector.

FIG. 44 is an explanatory diagram showing the size of a bright spot of a Fourier transform image from a periodic pattern with high NA illumination.

FIG. 45 is an explanatory diagram showing the size of a bright spot of a Fourier transform image from a periodic pattern by low NA illumination.

FIG. 46 is an explanatory diagram showing the theoretical value of the intensity of scattered light from a foreign substance with respect to the wavelength of laser light and a dimensionless number depending on the particle size of the foreign substance.

FIG. 47 is an explanatory diagram showing the directions of diffracted light from a foreign substance by oblique illumination.

FIG. 48 is an explanatory diagram showing the definition of the aperture (NA) of the optical system.

FIG. 49 is an explanatory diagram showing the cross-sectional area of scattered light proportional to the intensity of scattered light from a foreign matter and the size of the foreign matter diameter.

[Explanation of symbols]

DESCRIPTION OF SYMBOLS 1 ... Inspection stage part, 2 ... 1st illumination system, 3 ... 2nd illumination system, 4 ... Detection optical system, 5 ... Judgment processing system, 6 ... Reticle, 9 ... Z stage, 10 ... X stage, 11 ... Y stage, 21, 31 ... Laser light source, 44, 444 ... Linear spatial filter, 47, 447 ... Variable spatial filter, 5
1, 551 ... Detector, 52 ... First detector, 552 ... Second detector, 53 ... AND circuit, 70 ... Foreign matter, 80 ... Circuit pattern, 111 ... Standard sample, 112 ... Block processing circuit, 113 ... first shading correction circuit, 123
Second shading correction circuit 114 First first pixel addition processing circuit 124 Second second pixel addition processing circuit

Claims (17)

[Claims] 1. A foreign matter inspection device for detecting foreign matter adhered to a transparent or semitransparent sample having a periodic circuit pattern, wherein the sample is orthogonal to each other on the same plane parallel to the sample surface. And an inspection stage that selectively moves in a direction perpendicular to the plane, an illumination system that obliquely irradiates a desired position on the surface of the sample, and a Fourier transform plane at a position vertically above the illumination area of the sample. In each of the
Moreover, it is possible to rewrite the first and higher order diffracted light generated from the pattern of the circuit pattern at a specific angle in the direction orthogonal to the irradiation direction of the illumination system, while locally blocking the scattered light generated from the circuit pattern. A detection optical system having a special spatial filter, a detector for detecting scattered light from the sample through the detection optical system, and a detection signal from the detector binarized to discriminate scattered light from foreign matter on the sample A foreign matter inspection device, comprising: 2. A foreign matter inspection apparatus for detecting foreign matter adhered on a transparent or semitransparent sample having a periodic circuit pattern and an aperiodic circuit pattern, respectively, wherein the sample is the same as the sample surface. Of the inspection stage that selectively moves in a direction perpendicular to each other and a direction perpendicular to the plane, an illumination system that obliquely irradiates a desired position on the surface of the sample, and an illumination area of the sample. Rewritable filter parts, which are respectively arranged at the same position on the Fourier transform plane in the vertically upper position and block the diffracted light of the first or higher order generated from the periodic circuit pattern, and the illumination of the illumination system in the non-periodic circuit pattern. A filter that locally blocks diffracted light of the first order and higher generated from an angle pattern in a direction orthogonal to the direction, and transmits scattered light generated from a non-periodic circuit pattern in the other part. A detection optical system having a spatial filter consisting of a detector, a detector for taking in scattered light and first-order or higher-order diffracted light from the sample through the detection optical system, and binarizing a detection signal from the detector, A foreign matter inspection device, comprising: a determination unit that discriminates scattered light from the foreign matter above. 3. A foreign matter inspection device for detecting a foreign matter adhered to a transparent or semitransparent sample having a periodic circuit pattern and an aperiodic circuit pattern, respectively, wherein the sample is the same as the sample surface. Of the inspection stage that selectively moves in a direction perpendicular to each other and a direction perpendicular to the plane, an illumination system that obliquely irradiates a desired position on the surface of the sample, and an illumination area of the sample. A filter unit that is arranged on the Fourier transform plane at the vertically upper position and blocks the diffracted light of the first order or more generated from the periodic circuit pattern,
And locally blocks the diffracted light of the first or higher order generated from the angular pattern of the direction orthogonal to the illumination direction of the illumination system in the aperiodic circuit pattern, and scatters other portions from the aperiodic circuit pattern. A detection optical system having a rewritable spatial filter integrally formed with a filter unit that transmits light on the same plane, and a detector that takes in scattered light from a sample and diffracted light of first or higher order through the detection optical system. And a determination unit that binarizes a detection signal from the detector and discriminates scattered light from a foreign substance on a sample. 4. The transparent or semi-transparent sample has a circuit pattern formed between the circuit patterns with a material that gives a phase change to the transmitted light of the sample.
3. The foreign matter inspection device according to item 1. 5. The foreign matter inspection apparatus according to claim 1, wherein the rewritable spatial filter is a photographic dry plate. 6. The observation in which the rewritable spatial filter can move and position the spatial filter and can observe a position of a scattered light generated from a sample and a Fourier-transformed image of diffracted light of a first order or higher order. 6. A means for holding a spatial filter, and further comprising any one of an optically transparent water tank provided with a mechanism for supplying and discharging a developing solution, a stop solution, a fixing solution, and a washing solution. The foreign matter inspection device described in. 7. The rewritable spatial filter is made of a photochromic material.
The foreign matter inspection apparatus according to item 1. 8. The rewritable spatial filter is composed of a photochromic material and has a position moving mechanism for moving the rewritable spatial filter to a desired position, and an overheating mechanism for heating the rewritable spatial filter. The foreign matter inspection apparatus according to claim 7, characterized in that. 9. The foreign matter according to claim 1, wherein the rewritable spatial filter is configured by either one of a thermoplastic element and an optical address type spatial light modulating element. Inspection device. 10. The rewritable spatial filter is configured by either one of a liquid crystal type spatial light modulating element and an electrically addressing type spatial light modulating element, and a conjugate of an image on the one light modulating element. It has a detector for detecting an image, a storage unit for storing the image detected by the detector, and a drive circuit for driving one spatial light modulator based on the contents of the storage unit. Item 1. The foreign matter inspection device according to any one of items 1 to 3. 11. The storage means has a switching means for storing the detection image by the detector each time the cycle of the sample is different, and selectively switching the storage contents of each to output to the drive circuit. The foreign matter inspection device according to claim 10. 12. The switching means is configured to perform a switching operation according to the design data content of the sample or the inspection data content based on the design data, and when the number of detected foreign matters exceeds a certain inspection time or at a certain number. 13. The foreign matter inspection apparatus according to claim 11, wherein the foreign matter inspected within the inspection area is one of means for performing a switching operation when the area of the inspected foreign matter exceeds a predetermined amount. 13. The storage content of the storage means includes design data of a sample, wavelength of oblique illumination light, NA value of oblique illumination,
Irradiation direction of oblique illumination, reflectance of the sample to the oblique illumination light on the substrate, reflectance of the circuit pattern of the sample to the oblique illumination light, film thickness of the circuit pattern of the sample,
The foreign matter inspection apparatus according to claim 10, wherein the foreign matter inspection apparatus is a Fourier transform image obtained based on each of the NA values of the detection optical system. 14. The foreign matter inspection apparatus according to claim 1, wherein the detection optical system has an NA value of 0.4 or more. 15. The foreign matter inspection apparatus according to claim 1, wherein the illumination system is a collective illumination. 16. The illumination system has an NA of 0.01 ≦ NA.
The foreign matter inspection apparatus according to claim 15, wherein ≦ 0.4. 17. The incident angle of the illumination system with respect to the sample is 2
2. The range is 0 ° to 80 °.
The foreign matter inspection device according to any one of items 5 and 16.