JP2001194323A - Method and device for inspecting pattern defect - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a method and a device for detecting a minute circuit pattern with high resolution to detect a defect. SOLUTION: This device is made up of an objective lens for detecting an image of a sample, a laser lighting means for lighting the pupil thereof, a means for reducing the coherence of laser lighting, a storage-type detector, and a means for processing a detection signal thereof.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a pattern inspection for detecting a defect (short circuit, disconnection, etc.) and a foreign substance of a pattern to be inspected, and a foreign substance inspection. The present invention relates to a method and an apparatus for inspecting a defect of an inspected pattern for inspecting foreign matter. Hereinafter, the defect includes foreign matter.

[0002]

2. Description of the Related Art Conventionally, this type of inspection apparatus is disclosed in
As described in Japanese Patent Publication No. 318326 (Prior Art 1), an image of a pattern to be inspected is detected by an image sensor such as a line sensor while moving the pattern to be inspected, and the detected image signal is delayed by a predetermined time. By comparing the shading of the signals, the mismatch was recognized as a defect.

Further, as a prior art relating to a defect inspection of a pattern to be inspected, Japanese Patent Application Laid-Open No. 8-320294 (prior art 2) is known. According to the prior art 2, in a pattern to be inspected such as a semiconductor wafer where a region having a high pattern density such as a memory mat portion and a region having a low pattern density such as a peripheral circuit are mixed in a chip, the brightness on a detected image is high. The detected image signal is converted into A so that the brightness or contrast between the high-density region and the low-density region of the pattern to be inspected is determined from the frequency distribution of
The digital image signal obtained by the / D conversion is subjected to gradation conversion, and the gradation-converted image signal is compared with the gradation-changed image signal for comparison. There is described a technique for inspecting a high precision.

A conventional technique for inspecting a photomask pattern is disclosed in Japanese Patent Application Laid-Open No. Hei 10-78668 (prior art 3). In the prior art 3, a UV laser light such as an excimer laser is used as a light source,
Rotating the diffuser inserted on the optical path to uniformly illuminate the mask with UV light whose coherence has been reduced, calculating the feature quantity from the obtained mask image data, and determining the quality of the photomask Is described. A projection exposure apparatus using an excimer laser is disclosed in Japanese Patent Application Laid-Open No. 59-226317.
And Japanese Patent Application Laid-Open No. 62-231924.

[0005]

In the recent LSI manufacturing, circuit patterns formed on wafers have been miniaturized to meet the needs for high integration, and the pattern width has been reduced to 0.1.
The diameter has become smaller than 25 μm, and has reached the resolution limit of the imaging optical system. For this reason, an increase in NA of the imaging optical system and application of the optical super-resolution technology are being promoted. However, high NA has reached the physical limit. Therefore,
The essential approach is to shorten the wavelength used for detection to the range of UV light or DUV light. In addition, since it is necessary to perform inspection at a high speed, a method of scanning a laser beam with a narrow beam on a sample cannot be used. Conversely, if the laser beam is illuminated so as to cover the entire field of view, speckles occur, and overshoots and undershoots called ringing occur at the edges of the circuit pattern, so that a high-quality image cannot be obtained.

SUMMARY OF THE INVENTION An object of the present invention is to solve the above problems and to provide a pattern defect inspection method and apparatus for detecting a fine circuit pattern at high speed with high resolution and detecting defects. It is another object of the present invention to provide a method of manufacturing a semiconductor device capable of manufacturing an ultra-fine semiconductor device using the above-described pattern defect inspection method and apparatus.

[0007]

In order to achieve the above object, according to the present invention, a UV light source or a UV laser light source is used as a light source, and the generation of speckles of UV light or UV laser light in an optical path is suppressed. Means are provided so that the surface of the object is irradiated with UV light having reduced coherence to detect an image of the object. Here, the UV light includes DUV light. As means for suppressing the generation of speckles of the UV light, in the present invention, 1) the UV light from the UV light source is focused on one or more points on the pupil of the objective lens, and the focused point is detected. 2) The UV light emitted from the UV light source is incident on the bundle of optical fibers whose optical axis has been shifted, and the emitted light is projected on the pupil of the objective lens. 3) Focusing on the optical fiber group whose optical path length is changed more than the coherent distance of the UV light source, and focusing the emitted light on the pupil of the objective lens. 4) Arranging the diffusion plate, Means for causing relative movement with the light flux in a direction substantially perpendicular to the axis, and 5) illuminating the pupil with a combination thereof are provided.

In order to improve the pattern contrast, attention is paid to the fact that the polarization state of the laser can be freely controlled, and the polarization direction and the ellipticity of the illumination light are controlled to detect a partial polarization component of the detection light. Made it possible. That is, in the present invention, in order to achieve the above object, a pattern defect inspection apparatus includes a light source that emits UV light, laser light, UV laser light, or the like, and a UV light, laser light, or UV laser light that is emitted from the light source. Irradiating means for irradiating the sample with reduced coherence, and image detecting means for detecting an image signal by imaging the sample irradiated with laser by the irradiating means,
A defect detection unit configured to detect a defect of a pattern formed on the sample based on information on an image signal of the sample detected by the image detection unit.

Further, the present invention provides a light source for emitting UV light, irradiation means for reducing the coherence of the UV light emitted from the light source and irradiating the sample via an objective lens, and the irradiation means. Image detecting means for capturing an image of the irradiated sample with reduced coherence through the objective lens to detect an image signal, and forming the image on the sample based on information on the detected image signal detected by the image detecting means And a defect detecting means for detecting a defect of the selected pattern. Further, according to the present invention, in the above-described pattern defect inspection apparatus, a polarization unit for controlling a polarization state of the laser light or the UV laser light is provided. Further, according to the present invention, in the above-described pattern defect inspection apparatus, there is provided an image processing means for processing an image signal of the sample detected by the image detecting means, which corresponds to a throughput of three or more wafers equivalent to 200 mm in diameter per hour. Processed at a speed, the pattern formed on the sample
It was configured to detect including a defect of 00 nm. In addition, the present invention includes a plurality of image processing units (PE units) equipped with a plurality of algorithms and capable of selecting these algorithms and determination thresholds automatically or by designating, and providing a pattern formed on a sample. A pattern defect inspection device characterized by detecting a defect.

In the present invention, in order to achieve the above object, a pattern defect inspection method includes irradiating a pattern-formed sample with UV light or UV laser light.
Imaging a sample irradiated with V light or UV laser light,
The image signal of the sample obtained by the imaging is compared with a reference image stored in advance to detect a pattern defect. Further, according to the present invention, in the pattern defect inspection method,
By condensing and scanning the UV light or the UV laser light on the pupil of the objective lens, the coherence is reduced and the occurrence of speckle is suppressed. In the present invention,
In the pattern defect inspection method, a UV laser is irradiated on a sample on which a pattern is formed, an image of the sample irradiated with the UV laser is taken, and the brightness of a normal portion of a detected image signal of the sample obtained by the imaging is determined in advance. The brightness correction of the image signal is performed so that the brightness of the normal part of the stored reference image signal is substantially the same, and the detected image signal subjected to the brightness correction is compared with the reference image signal to determine a pattern defect. Detected.

Further, according to the present invention, a pattern defect inspection method includes the steps of: reducing the coherence of laser light emitted from a laser light source; Is irradiated through the objective lens while changing the irradiation direction with time, an image of the sample irradiated with the laser light is taken, and a detection image signal of the sample obtained by this imaging and a reference image signal stored in advance are taken. A pattern defect is detected by comparison. Further, according to the present invention, in a method for inspecting a defect of a pattern formed on a sample, the surface of the sample is irradiated with UV laser light with reduced coherence, and the surface of the sample irradiated with the UV laser light is irradiated. An image signal is obtained by imaging the sample, a defect of 100 nm or less on the sample is detected by processing the image signal, and information on the position of the detected defect of 100 nm or less on the sample is output.

Further, according to the present invention, in the pattern defect inspection method, the UV laser light having reduced coherence is irradiated with a laser beam having a diameter of 20 mm.
Irradiation is performed on a wafer equivalent to 0 mm, an image of the wafer irradiated with the UV laser light is captured to detect an image of the wafer, and the detected image of the wafer irradiated with the UV laser light is processed to be formed on the wafer. 100 nm or less of the resulting pattern was detected at a throughput of three or more wafers per hour. Further, according to the present invention, in the above-described pattern defect inspection method, a pattern having repeatability is formed on the sample.

Further, in the present invention, a sample on which a pattern is formed by irradiating UV light with reduced coherence, an image of the irradiated sample is obtained to obtain a detection image signal, and the obtained detection signal is obtained. Create a scatter diagram showing the correspondence between the feature amount in the normal part of the image signal and the feature amount in the normal part of the reference image signal,
The gradation value of the image signal is corrected based on the created scatter diagram, and the corrected detected image signal and the reference image signal are compared with a determination threshold obtained from the scatter diagram to determine a pattern defect. It is characterized by detecting. In particular, the scatter diagram is divided into a plurality of pieces in accordance with, for example, the contrast, so that the spread (variance) in the normal part can be reduced, and the determination threshold can be reduced.

According to the present invention, in the above-described pattern defect inspection method, a pattern having a design rule of 0.07 μm or less is formed on the sample, and the defect can be inspected. That is, 0.07 μm
The semiconductor device is manufactured by inspecting a pattern of the following design rule. In the present invention,
When comparing the detected image signal with the reference image signal, the false alarm is significantly reduced even in a portion where the pattern of the sample is not formed.

[0015]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of a method and an apparatus for inspecting a defect of an inspected pattern according to the present invention will be described with reference to the drawings. FIG. 1 is a diagram showing an embodiment of the apparatus according to the present invention. The stage 2 is composed of an X, Y, Z, θ (rotation) stage, on which a semiconductor wafer (sample) 1 as an example of a pattern to be inspected is placed. Illumination light source 3
Is a light source configured of, for example, a UV laser light source having a wavelength of 266 nm or 355 nm to illuminate the semiconductor wafer 1. The UV laser light source is configured by a device that generates a third harmonic (355 nm) or a fourth harmonic (266 nm) of a fundamental wave by converting the wavelength of a solid YAG laser using a nonlinear optical crystal or the like. If it exists as a UV laser light source, it may be 100 nm or less, and the resolution will be further improved. Also, the oscillation mode of the laser may be continuous oscillation,
Although pulse oscillation may be used, continuous oscillation is preferable because the stage is continuously driven and an image from the object 1 is detected.

The beam expander 21 expands, for example, a UV laser beam to a certain size. The coherence reduction optical system 4 reduces the coherence of the laser light emitted from the illumination light source 3. The coherence reduction optical system 4 only needs to reduce temporal or spatial coherence. The coherence reduction optical system 4 can be constituted by, for example, a scanning mechanism for scanning a laser beam from the illumination light source 3 on a pupil of the objective lens 7.
The beam splitter 5 is constituted by a polarization beam splitter in some cases, and is configured to reflect UV illumination light from the UV illumination light source 3 and to perform, for example, bright field illumination on the sample (semiconductor wafer) 1 through the objective lens 7. ing. When the beam splitter 5 is constituted by a polarizing beam splitter, the beam splitter 5 has a function of reflecting when the polarization direction of the UV laser light is parallel to the reflection surface and transmitting the light when the polarization direction is perpendicular. Therefore, UV
Since the laser light is originally a polarized laser light, it is possible for the polarized light beam splitter 5 to totally reflect the UV laser light.

The polarizing element group 6 controls the polarization directions of the UV laser illumination light and the reflected light so that the reflected light does not reach the image sensor 8 with uneven brightness due to the pattern shape and density difference. It has a function of arbitrarily adjusting the polarization ratio of the illumination light, and is composed of, for example, a 、 or 波長 wavelength plate. Then, a mirror 28 and a lens 2 provided in an optical path between the beam splitter 5 and the image sensor 8 are provided.
9 and the detector 30, the pupil plane 7 of the objective lens 7
It is configured to detect the aerial image of a. in this way,
The CPU 290 sets the rotation angle by controlling the rotation of each of the half-wave plate and the quarter-wave plate around the optical axis based on the aerial image of the pupil plane detected by the detector 30. The polarization state of the reflected light emitted from the circuit pattern formed on the semiconductor wafer 1, that is, the diffracted light of the reflected light is controlled. For example, the 0th-order diffracted light is attenuated and the high-order diffracted light is almost attenuated. And the contrast of the pattern can be significantly improved, and a stable detection sensitivity can be obtained. Since UV light or DUV light is used as the objective lens 7, the influence of chromatic aberration can be reduced by using a reflective objective lens. The image sensor 8 has a pixel size of about 0.05 μm to 0.3 μm on a sample basis, and has a grayscale image corresponding to the brightness (shade) of light reflected from the semiconductor wafer 1 as an example of the pattern to be inspected. It outputs a signal.

With the above structure, the UV light (for example, UV laser light) L1 emitted from the illumination light source 3 is
, Is expanded by the beam expander 21, and the coherence reduction optical system 4, the lens 22, the beam splitter 5,
Then, the light enters the objective lens 7 via the deflecting element group 6 and is irradiated onto the sample (semiconductor wafer) 1. That is, UV light L
1 is condensed by the lens 22 near the pupil 7 a of the objective lens 7, and is then Koehler-illuminated onto the sample 1. The reflected light from the sample 1 is detected by the image sensor 8 from above the sample 1 via the objective lens 7, the polarizing element group 6, the beam splitter 5, and the imaging lens 23. That is, while the stage 2 is scanned and the semiconductor wafer 1 as an example of the pattern to be inspected is moved at a constant speed, the position of the surface to be inspected of the semiconductor wafer 1 in the Z direction is always detected by a focus detection system (not shown). Then, the stage 2 is controlled in the Z direction so that the distance from the objective lens 7 becomes constant, and the image sensor 8 outputs the brightness information (shade image signal) of the pattern to be inspected formed on the semiconductor wafer 1. Detect with high accuracy. The signal processing circuit 19 includes an A / D converter 9, a gradation conversion unit 10, a delay memory 11, a positioning unit 286, and a local gradation conversion unit 2.
87, comparison section 288, CPU 290, image input section 29
2, scatter diagram creation unit 293, storage device 291, display means 2
94, an input unit 289, an output unit 295, and the like.

The A / D converter 9 converts the gray image signal 8a obtained from the image sensor 8 into a digital image signal and outputs an image signal of the sample. For example, 10
Bits are used. The gradation converter 10 performs gradation conversion on a 10-bit digital image signal output from the A / D converter 9 as described in JP-A-8-320294. That is, the gradation converter 10
Is used to correct an image by performing logarithmic conversion, exponential conversion, polynomial conversion, and the like, and is configured to output an 8-bit digital signal, for example. The delay memory 11 is a storage unit that stores a reference image signal, and constitutes a semiconductor wafer in which an output image signal from the gradation converter 10 is repeated.
This is to store and delay a cell or a plurality of cell pitches or one chip or a plurality of chips. Here, a cell is a repeating unit of a pattern in a chip. Positioning unit 2
Reference numeral 86 denotes a gradation-converted image signal (detected image signal obtained from the sample) 285 output from the gradation converter 10
And a delay image signal (reference image signal) 284 obtained from the delay memory 11 to detect the amount of positional deviation by a normalized correlation and perform alignment in pixel units.

The local tone conversion unit 287 performs tone conversion of signals having different feature amounts (brightness, differential value, standard deviation, texture, etc.) for both or one of the image signals so that the feature amounts match. Part. Comparison section 288
Is a part for detecting a defect based on a difference in feature amount by comparing the detected image signals subjected to gradation conversion by the local gradation conversion unit 287. That is, the comparing unit 288 compares the reference image signal delayed by an amount corresponding to the cell pitch or the like output from the delay memory 11 with the detected image signal. By inputting the coordinates of the array data and the like on the semiconductor wafer 1 by the input means 289 including a keyboard, a disk, and the like, the CPU 290 can perform the defect inspection based on the coordinates of the array data and the like on the semiconductor wafer 1. Data is created and stored in the storage device 291. This defect inspection data is provided to a display means 29 such as a display as required.
4 and output to the output means 295.

The details of the comparison processing unit 288 may be those disclosed in Japanese Patent Application Laid-Open No. 61-212708, for example, an image registration circuit, a difference image detection circuit of a registered image, a difference image A mismatch detection circuit for binarizing
It comprises a feature extraction circuit that calculates the area, length (projection length), coordinates, etc. from the quantified output. The image input unit 292
This is a part for synchronously or asynchronously inputting both images that have been aligned in pixel units by the alignment unit 286 in order to create a scatter diagram. The scatter diagram creation unit 293 includes the image input unit 2
For both images input at 92, a scatter diagram is created between the feature amount of the detected image and the feature amount of the reference image for each category, and is displayed on, for example, the display unit 294.

These detailed embodiments will be described later.

Next, the illumination light source 3 will be described. For higher resolution, it is necessary to shorten the wavelength, but it is difficult to obtain illumination with high illuminance in the UV wavelength region where the effect can be obtained most. Discharge lamps are excellent as a UV light source. In particular, mercury-xenon lamps have a stronger emission line in the UV region than other discharge lamps. FIG. 2 shows an example of the radiation intensity with respect to the wavelength of the mercury xenon lamp.
Compared to the conventional wide wavelength range of visible light, the emission line in the DUV region is only 1-2% of the total output light (30% in the visible region).
%). In addition, there is no directivity in light emission, and the efficiency of guiding the light emitted from the discharge lamp to the sample cannot be increased even with a carefully designed optical system. It is not possible to secure a sufficient amount of light for use in image detection. In addition, even if a high-output discharge lamp is used to improve the illuminance (luminance) on the sample, the size of the luminescent spot is only larger than that of a low-output discharge lamp. Per optical power). Therefore, it is considered that using a laser as a light source is suitable for performing high-luminance illumination effective in the UV region.

As described above, there is a great merit in using a UV laser beam as a light source. In the present invention, UV
A method for performing illumination by laser light will be described. FIG. 3 shows the illumination state of the objective lens pupil and the visual field when illuminated with normal white light. In the figure, AS shown in (a) indicates a pupil, and FS shown in (b) indicates a visual field. At the position of the pupil AS, an image 31 of the light source is formed, and at the position of the visual field FS, the entire visual field 32 is almost uniformly illuminated.

Next, FIG. 4 shows a case where illumination is performed with a UV laser light source. In this case, the light source image 41 at the position of the pupil AS shown in FIG. In the field of view FS shown in FIG.
The circuit pattern illuminated as shown by 2, for example, in the case of a pattern having a cross section as shown in FIG. When the circuit pattern is illuminated with laser light to obtain an image of the circuit pattern in this manner, overshoot, undershoot, and speckle occur at an edge portion. This is because the σ of the illumination is small. This means that illumination is not performed on the visual field FS below the objective lens 7 from various angles. In normal white light illumination, the pupil A
Illumination with a certain size is performed on S, and illumination is performed on the field of view FS from a direction having an angle range equivalent to the NA (numerical aperture) of the objective lens 7.

For coherent (having coherence) light such as laser light, σ (proportional to the size of the light source on the pupil) is zero. This is because the image on the pupil becomes a point because the light source image of the coherent light is a point. Of course, as shown in FIG. 5A, the light beam 51 expanded by another lens system can be projected onto the pupil 7a (AS) of the objective lens 7, but σ = 0 since the UV laser light has coherence. The same result (52 shown in FIG. 5B) as when all light is emitted from the position is obtained, and this does not solve the problem. Therefore, means for reducing the coherence of the UV laser light is required. To reduce coherence, either temporal coherence or spatial coherence may be reduced.

Therefore, in the present invention, an image of the light source is formed on the pupil 7a of the objective lens 7 of the inspection apparatus, and the pupil-light scanning mechanism (pupil) constituting the light modulator as the coherence reducing optical system 4 is formed. As an upper light scanning optical system), for example, first, 6 in FIG.
It is proposed to illuminate the position 1 and then scan the position 62 and then the position 63... To illuminate the field of view shown in FIG. Further, as shown in FIG. 6C, a spiral scan may be performed on the pupil 7a. Also,
As shown in FIG. 6D, scanning may be performed two-dimensionally on the pupil 7a. During this time, images of speckle, overshoot, and undershoot are obtained at each position, but since the obtained times are different, there is no interference with each other. Therefore,
Addition of them on the detector 8 results in the same image as with an incoherent light source. In order to perform the addition on the detector 8, the detector 8 has a pixel size of about 0.05 μm to 0.3 μm in terms of on the sample (on the visual field) and is a storage type detector (such as a CCD). Specifically, TDI
Sensors) are suitable.

That is, consider the use of a one-dimensional sensor as the storage type detector 8. As shown in FIG. 7, even if the one-dimensional sensor 71 illuminates the entire visual field, only the region 72 contributes to the detection even when the entire field of view is illuminated, and does not contribute to the detection in the region 73 occupying most of the optical power. become. Thus, in order to improve the illuminance, it is preferable to perform linear illumination on the one-dimensional sensor 71 as shown in an area 82 as shown in FIG. (A two-dimensional image is obtained by scanning the CCD in the Y direction on the visual field). In this case, by illuminating the pupil with the length in the Y direction in the figure as shown at 91 in FIG. Scanning on the pupil is performed in the X direction. The scanning period Ts is the CCD
Is shorter than the accumulation time Ti. As a result, the images can be added.

However, in this scanning, the scanning in the Y direction cannot be performed because the pupil extends from the beginning in the Y direction. For this reason, it is not possible to reduce overshoot and undershoot due to light interference generated in the Y direction of the CCD in the visual field. Conversely, if the length in the Y direction is reduced to perform scanning in the Y direction on the pupil, the width in the Y direction on the visual field is increased, and the illuminance is reduced. In order to solve this problem, in the present invention, as the image sensor 8, a time delay integration type, that is, TDI (Time Delay & In) among CCD sensors as shown in FIG.
The problem can be solved by using a sensor of the (tegration) type. In the case of the TDI sensor, N stages (several tens to 256 stages) of light receiving portions are arranged in the field of view, so that even if the area illuminated in the field of view is N times wider, the UV slit illumination light is short. 102 can be used effectively for detection.

For this reason, the length of the short UV slit light 102 on the pupil 7a in the Y direction can be reduced to about 1 / N of that of the CCD shown in FIG.
a can be scanned in both X and Y directions. As a result, overshoots and undershoots occurring in the X and Y directions of the TDI in the visual field can be reduced, and a good detected image can be obtained. The scanning period Ts on the pupil is
It suffices if the storage time is shorter than N times the storage time of one TDI stage. However, in consideration of the illuminance distribution generated in the visual field, it is preferable that Ts be shorter than 1/2 of N times Ti in order to perform more uniform detection. In order to form the short UV slit spot light 102, as shown in FIG. 1, a cylindrical lens or a cylindrical lens group (FIG. 12) is provided between the beam expander 21 and the pupil optical scanning mechanism 4. What is necessary is just to arrange | position the UV slit light formation optical system 25 which consists of multi UV slit light) 25a etc. By the way, since the short UV slit light 102 spreads in one direction to some extent, it causes some light interference. Therefore, by placing the diffusion plate 26 as an optical modulator after the pupil light scanning optical system 4,
Light interference can be completely eliminated.

Further, in order to perform uniform illumination, the UV light from the UV laser light source 3 is not directly condensed on the pupil as the short UV slit spot light 102, but as shown in FIG. It is preferable to condense light after passing through a fly-eye having a lens group or an integrator 25b. As described above, even when a UV multi-slit spot beam is used, the optical interference can be completely eliminated by disposing the diffusion plate 26 as an optical modulator after the pupil optical scanning mechanism 4. That is, for example, a plurality of UV light sources may be created from one UV laser light source by a glass rod lens group (fly-eye lens) having a uniform length as shown in FIG. Further, a multi-cylindrical lens array having a simpler configuration than the glass rod lens group shown in FIG. 12 may be used. In the case of a multi-cylindrical lens array, since a plurality of UV light sources are generated only in one direction, a plurality of UV light sources can be generated two-dimensionally by arranging two orthogonal UV light sources. Further, in that case, by changing the respective pitches, it is possible to generate UV light source groups having different vertical and horizontal pitches between the light sources.

Further, as shown in FIG. 13 (a), a light source group 252 is formed on the pupil plane 7a of the objective lens 7 with a different magnification, and, for example, as shown by an arrow 254 in FIG.
Is circularly rotated, as a result, a distribution 243 of illumination on the pupil plane as shown by oblique lines in FIG. 13B is obtained, which becomes annular illumination and improves the resolution of the detected image. Will have. Further, the annular illumination condition can be changed only by changing the magnification of the light source group.
It is also possible to illuminate the entire pupil plane so that σ = 1. When a TDI image sensor having N stages with a scan rate of 1 kHz is used, there is an additional advantage. The basic period of the galvanomirror may be 1 kHz / N, and the entire pupil plane can be scanned at this period. Galvanometer mirrors with a frequency of several kilohertz are commercially available.
If an image sensor is combined, pupil scanning can be realized at a practical speed, and high-speed image detection becomes possible. Where T
The number of stages (the number of stages) of the DI image sensor may be prepared according to the speed of the galvanomirror. If a TDI image sensor with a variable number of stages (number of stages) is used, the accumulation time can be changed by a pupil scanning method.

FIG. 19 shows a U-type lens using the above lens array.
1 shows a schematic diagram of a V illumination system. Correctly, three-dimensionally
However, it is not possible to show the important light condensing relationship here, so it is schematically shown. The UV parallel light flux 235 from the UV laser light source 3 is incident on the lens array 25, and generates a plurality of bright spots (new light sources) on the second pupil conjugate plane 233 conjugate with the pupil plane 7a of the objective lens 7. From here, a plurality of UV luminous fluxes are emitted.
It is shown inside. The light emitted from the new UV light source group is converted into a substantially parallel light beam by the second projection lens 232 and projected on the second scanning mirror surface 198.

The light reflected here is transmitted to the second condenser lens 19.
9, through the first pupil conjugate plane 231, the first projection lens 1
The light is converted into substantially parallel light by 910 and projected on another first scanning mirror surface 195. Then, the light is condensed on the pupil plane 7 a by the first condenser lens 22, converted into substantially parallel light by the objective lens 7, and illuminated on the sample surface 1.
The good point of this method is that the generated plural bright spots have corresponding outputs of the intensity distribution of the incident Gaussian beam 235, and therefore, they overlap on the sample 1 so that the illumination with a small illuminance distribution can be obtained. The point is that it can be obtained.

Next, a description will be given of a method of improving the pattern contrast by controlling the deflecting element group 6 as described above, in addition to increasing the resolution by UV light.
To improve the pattern contrast, the deflection element group 6
Focusing on the fact that the polarization state of the UV laser light can be freely controlled based on the control of (i), the polarization direction and the ellipticity of the illumination light are controlled, and the image sensor 8 detects a partial polarization component of the detection light. Made it possible. The characteristics of illumination by UV laser light include a single wavelength and linearly polarized light. Therefore, the half-wave plate provided in the optical path
The polarization state can be controlled with high efficiency by the polarizing element group 6 such as a quarter-wave plate. The control may be performed by, for example, rotating a half-wave plate and a quarter-wave plate around the optical axis.
Since the pattern contrast greatly changes depending on the polarization state of illumination, the performance of the optical system can be improved by controlling the polarization state (positioning by rotating the wave plate). More specifically, the direction of linearly polarized light can be controlled by a 波長 wavelength plate, and the ellipticity can be changed by a 波長 wavelength plate.

As shown in FIG. 18, a desired polarization component can be extracted by an analyzer 242 provided on the detection side, and a component that does not contribute to defect detection, for example, a zero-order light is further reduced, and a diffracted light is reduced. For example, more light components including pattern edges and contributing to defect detection can be captured. Thereby, the detection sensitivity can be improved. The analyzer 242 may also be rotatable according to the polarization state. With these combinations, parallel Nicols and orthogonal Nicols can be realized. Of course, a circularly polarized state can also be realized. These do not depend on the illumination wavelength itself. If the above concept holds,
Arbitrary configuration may be used for realization. Objective lens 7
By observing the diffracted light from the pattern on the pupil plane by the detector 30 and selecting the polarization state, the diffracted light is 0% higher than the higher-order diffracted light.
It can be confirmed that the next light is attenuated. Thereby, the low frequency component can be attenuated, and the pattern contrast can be improved. Of course, a spatial filter may be arranged at a position conjugate with the pupil of the objective lens 7 to attenuate the zero-order light (the diffracted light from the pattern is blocked by the spatial filter,
The scattered light from the foreign matter can be guided to the image sensor.)

However, by controlling the polarization, higher-order diffracted light can be extracted more efficiently. Experiments by the inventors have shown that the contrast is improved by about 20-300%. The polarizing element 241 can be installed at a position where desired performance can be obtained (for example, between the beam splitter 5 and a polarizing element group 6 such as a quarter-wave plate).

The scanning of the UV slit spot on the pupil may be a spiral scan 66, a television (raster) scan 67, or another scan as shown in FIGS. 6B and 6C. . However, it is desirable that one unit of scanning be performed within the accumulation time of the image sensor 8. Therefore,
Scanning is preferably performed in synchronization with the operation of the image sensor 8. For example, as shown in FIG.
In the case of circular scanning on a, assuming that the accumulation time of the image sensor 8 is 1 ms, the galvanometer mirrors 195 and 198 for two-dimensional scanning may be driven at a basic period of 1 kHz. Further, it is preferable to synchronize the stage 2, the image sensor 8, and the on-pupil scanning. In this case, stage 2 has the greatest inertia and is therefore most difficult to synchronize. The above-pupil optical scanning optical system is easy to synchronize at a wide frequency or easily at a limited frequency depending on the type of the mechanism. Further, since the sensor is an electric circuit, synchronization is easy. Therefore, it is easy and desirable to generate a basic synchronization signal from the position of the stage and synchronize the other two with it.

That is, as shown in FIG. 1, the synchronization signal generator 163 is a sensor based on the position of the stage detected by a position detecting mechanism (not shown) such as a linear encoder attached to the XY stage 2. And a synchronizing signal 165 for the above-pupil optical scanning mechanism.
The optical scanning mechanism on the pupil includes an A / O deflector and an E / O
Synchronization is easiest when an electric signal from a deflector or the like is directly converted into a deflection angle of light. Furthermore, a mirror-based deflector such as a galvanometer mirror or a polygon mirror can also be used as the above-pupil light scanning mechanism. As described above, the F shown in FIG. 6B and FIG.
An image of the illumination 65 of the UV light or DUV light on the entire field of view, such as S, or the illumination 101 of the UV light or DUV light on the light receiving surface of the TDI sensor can be obtained without causing optical interference.

Next, an embodiment of a TDI sensor capable of detecting UV light, especially DUV light, will be described.
When a DUV laser light source is used as the illumination light source 3, it is necessary to use an image sensor that is sensitive to DUV. In the front-illuminated image sensor, since incident light passes through the gate and enters the CCD, the incident light having a short wavelength is attenuated and has little sensitivity to wavelengths of 400 nm or less, and DUV light cannot be effectively detected. In order to obtain DUV sensitivity with a front-illuminated image sensor, there is a method of reducing the attenuation of short wavelengths by making the gate thinner. Another method is to apply an organic thin film coating to the cover glass,
There is a method of detecting DUV light with an image sensor having sensitivity only to visible light by emitting visible light when DUV light is incident. In addition, since light is incident on the back side without the gate structure, the back-illuminated image sensor has a high quantum efficiency (for example, 30% or more) and a large dynamic range (for example, 3000 or more).
It is sensitive to wavelengths below 400 nm and is particularly advantageous for short wavelength illumination below 200 nm. In the case of such an image sensor, even when several illumination wavelengths are used, one image sensor can be used.

The image sensor 8 is set to TDI (Time D
elay Integration (time delay integration type) can increase the sensitivity. Further, by providing the anti-blooming characteristic, it is possible to solve a problem that when an amount of detected light more than necessary is obtained, electric charges overflow to surrounding pixels. As explained above, 266 nm and 24
By using DUV light with a wavelength of 8 nm, 0.0
Defect inspection for devices with a rule of 7 μm or less can be realized. Further, the present invention can be applied to inspection of a Cu damascene as an object to be inspected. In addition, even if there is no pattern as an object to be inspected, no spec is generated. Therefore, even if the detected image is compared with the reference image, the inspection can be performed without generating a false report.

Next, an embodiment of the above-pupil light scanning mechanism (up-pupil light scanning optical system) 4 will be described. FIG. 16 shows a configuration for scanning the pupil 7a of the objective (detection / illumination) lens 7 with a UV laser spot according to the present invention. In FIG. 1, the configuration on the illumination side is shown, and the configuration on the detection side is omitted. In order to show the principle, the optical scanning mechanism on the pupil is shown only for one dimension. The UV beam emitted from the laser light source 3 (parallel light because it is a UV laser beam) is shaped into a required beam shape by a beam shaping mechanism 21, and the shaped beam is a light that is a coherence reducing optical system 4. The light is deflected by an on-pupil light scanning mechanism 195 that constitutes a modulator. Here, a polygon mirror is shown as an embodiment of the above-pupil optical scanning mechanism 195. The scanned parallel beam is converted into a change in the deflection angle by an f-θ lens 22 called a condenser lens. Therefore, the lens 22 is moved from the scanning mirror surface to the lens 2.
It is arranged at a position separated by a focal length of 2. And
The light is focused on the pupil plane 7a of the objective lens 7 by the lens 22. Therefore, the distance between the lens 22 and the pupil plane 7a is
2 focal length. Thus, on sample 1,
The UV laser beam emitted from the objective lens 7 illuminates the sample 1 as a parallel beam while changing the angle.

FIGS. 17 and 18 show a case where the UV laser beam is two-dimensionally scanned on the pupil 7a. In FIG. 1, a plate-like mirror such as a galvanometer is shown as an embodiment of an optical scanning mechanism on the pupil which is an optical modulator. The mirror 1911 in the figure is used to bend the optical path and is not essential. Therefore, what is different from FIG. 16 is the f-θ lens 199, the scanning mirror 198, which is the optical scanning mechanism on the pupil of the other axis, and the lens 1 that enters the scanning mirror 195.
910 has been added. In FIG. 18, an analyzer 242 is arranged between the beam splitter 5 and the imaging lens 23. In this embodiment, the UV laser light source 3 emits UV linearly polarized laser light. Therefore, for example, the analyzer 242 can shield the linearly polarized light component of the zero-order light UV illumination light reflected from the sample.

The above-described optical modulator on the pupil constituting the optical modulator converts the UV illumination light to the light shown in FIGS.
By scanning the pupil 7a of the objective lens 7 two-dimensionally as shown in (d), the sample 1 can be illuminated with a wide field of view without interference. Note that the NA of the objective lens 7 in the example is 0.75. As this NA is larger, the effect of the pupil scanning becomes larger, and the influence of the thin film interference of the sample pattern (the brightness of the patterns having different thicknesses is different, the difference between the normal portions becomes larger in the pattern comparison described later,
Minute defects become difficult to detect. (Even if the film thickness changes in a minute range called grain or hillock, the difference in brightness is large.)
Is to be reduced.

FIG. 18 shows an example in which a diffusion plate 26 as an optical modulator is arranged in the optical path, similarly to FIG. The arrangement position of the diffusion plate 26 is a position conjugate with the pupil 7a of the objective lens 7. In this embodiment, since the UV laser beam is scanned on the diffusion plate 26, the effect of reducing coherency is further increased. Of course, the diffusion plate 26 for reducing the spatial coherence may be reciprocated or rotated at high speed in a direction intersecting the optical axis of the UV laser beam. In particular, placing the diffusion plate 26 in the optical path for the UV slit spot light 102 shown in FIG. 10 and the multi UV slit spot light 252 shown in FIG. 13 is effective in reducing optical interference. The image of the light source image on the pupil 7a is formed by the illumination light source 3
This is performed by condensing the UV light from the pupil upper surface 7a of the objective lens 7 (in the case of detection by epi-illumination, both the illumination or irradiation lens and the detection lens are used) by the condenser lens 22. Here, in the case of the UV laser light whose light source is a point light source,
The points and spots are narrowed down to the diffraction limit. That is, UV
All the power of the laser is concentrated on this spot,
The power at that point will be considerable.

The actual objective lens 7 is composed of a very large number of lens groups (10 or more in many cases) in order to correct aberration, and the position of the pupil plane 7a is also determined by the design of the objective lens. In addition to the separated position, the position may be inside the lens (glass material portion) or near the lens surface. In this case, damage due to exposure of the coating (such as antireflection) applied to the lens to high-power laser light becomes a problem. This is because, in the present invention, a UV light spot is formed on the pupil plane. Therefore, in the present invention, no damage is caused on the sample.

The power of this spot becomes more serious in the case of a small spot with the same total power.
Therefore, an explanation will be given by using an average power density = total power / spot area defined by using the spot area obtained from the spot diameter and the total power of the spot. FIG. 15 shows a cross section of the intensity distribution 181 of the UV laser beam. This is a typical laser beam shape, and has a distribution called a Gaussian distribution in which the center is high and becomes lower toward the periphery. It is difficult to define the diameter of the spot for a gradually decreasing distribution, but here, the point 182 where the intensity is 13.5% of the intensity at the center is defined as the diameter. According to this definition, in the case of a Gaussian distribution, the power density at the center is twice the average power density of the whole. The value that damages the coating due to this average power density is 200 W / sq mm in experiments by the inventors (thus the center power density is 400 W / sq mm). If this value is not exceeded, no damage is caused to the coating. With respect to this problem, in the present invention in which the role of the pupil 7a in the objective lens 7 is large, the problem can be avoided by designing the position of the pupil plane at a position away from the surface of the lens glass material in advance. When separated, the spot becomes blurred from the focused state, the diameter increases slightly, and the average power density decreases. It is known from the results of experiments and studies by the inventors that the separation distance is required to be about 5 mm or more.

If a sufficient separation distance cannot be obtained due to the configuration of the objective lens 7, only that lens may be left uncoated. The inventor believes that if only some of the lenses are uncoated, the influence on the transmittance of the entire objective lens is small, and the problem of the proof stress of the coating can be addressed. Also,
As described above, the stage 2 is used as the UV laser light source 3.
Is driven, and the image sensor 8 captures an image. Therefore, a continuous wave laser is suitable. As described above, by using the continuous oscillation type as the laser light source 3,
The peak power can be kept low and the objective lens 7
Can be prevented from being damaged. The reason is that in a pulse oscillation type laser, even if the average output is suppressed, a very large power is applied at the peak value (peak value) of the pulse output, and at this time, the objective lens 7 and the like are damaged. Is generated. Of course, a pulse oscillation type laser may be used for a small output laser which does not have to worry about damage.

Next, another embodiment for reducing spatial coherence in the coherence reducing optical system 4 will be described. To reduce spatial coherence,
It is sufficient to obtain light having an optical path difference longer than the coherent distance of the UV laser. More specifically, as shown in FIG.
When the output light of the laser is made incident on the optical fiber 111 or the glass rod bundled with its length changed, the output light becomes incoherent (no coherence) light. If these are arranged on a position conjugate with the pupil 7a of the objective lens 7, an image without overshoot, undershoot, and speckle can be obtained. In this method, the shorter the coherence length of the UV laser light source, the better.
To this end, the oscillation wavelength band Δλ1 as shown in FIG. 20A is narrow, and a plurality of longitudinal modes as shown in FIG. Those having a wide Δλ2 are suitable.

Another method for reducing spatial coherence is to change the transverse mode (spatial distribution, light intensity I with respect to space) of emitted light when the optical fiber is incident on the optical fiber with its optical axis shifted. Some use phenomena. Usually, such a mode change is considered to be a disadvantageous phenomenon for industrial use, and it is general to make an effort to reduce the change in the transverse mode. As shown in the figure, the emitted light (a), (b), (c), (d), (e),. produce. As a result, the obtained emission lights become incoherent with each other, and they are arranged on a position conjugate with the pupil 7a of the objective lens 7. In the case of this method, a very large number of light sources (bright spots on the pupil) can be obtained by bundling a plurality of fiber strands.

Next, the two UV light sources 3
An embodiment for obtaining light will be described with reference to FIG. That is, in the present embodiment, the UV emission light emitted from the UV laser light source 3 is separated into two UV lights 133/134 having polarization planes orthogonal to each other by the polarization beam splitter 131, and
This is a method of obtaining virtual light sources of two UV lights 133 and 134 at a position conjugate with the pupil 7a of the objective lens 7. 132 is a mirror for changing the direction. Since UV light having polarization planes orthogonal to each other has no coherence, a virtual UV light source having no coherence can be obtained with a very simple configuration. Although only two virtual UV light sources can be obtained by this method, a virtual UV light source having no coherence can be obtained in half the labor by combining with the method described in the above.

Further, since the independent UV light sources do not have coherence, as shown in FIG.
Each point of the pupil of the objective lens 7 may be illuminated using 41, 142, 143, 144,... Further, as described above, by combining the polarization beam splitters 151 to 154 to form the configuration shown in FIG. 24, the number of virtual UV laser light sources can be reduced to half, and the cost can be reduced. As described above, there have been described a plurality of embodiments in which the coherence of the UV laser light is reduced, thereby illuminating a plurality of points on the pupil 7a of the objective lens 7 and condensing the image with the objective lens 7 to obtain an image. Combinations may be used, or a reduction method equivalent to these methods may be used. Further, as in the above embodiment, UV laser illumination is performed by partially changing the illumination optical path by a mirror or the like that vibrates (or oscillates) in a part of the optical path, and further, the image by the illumination of those optical paths is temporally accumulated. In the case of detecting an image by performing such a process, a temporal coherence reducing action is included in the process, so that it is not necessary to reduce spatial coherence as strictly as described above. Next, an embodiment of the signal processing circuit 19 will be described. Since the inspection target 1 has a repetitive pattern, the inspection extracts defect candidate points by comparing the pattern with an adjacent pattern. An output signal from the image sensor 8 is converted into a digital signal by the A / D converter 9.
In order to create a reference image to be compared, the delay memory (storage unit) 11 corresponds to the pitch of one or more repeated chips, or the cell region corresponds to the pitch of one or more repeated cells. Delay by the shift amount. Thus, the output of the delay memory 11 is an image obtained by shifting the inspection image by the repeated pitch.

Comparison section 288a in comparison processing section 288
Then, two images, that is, a detected image and a reference image, which have been locally gradation-converted by the local gradation conversion unit 287, are compared to obtain a difference (difference image) between corresponding pixel values. Then, the comparison unit 288
In step a, the obtained difference image is determined using a determination threshold for detecting defects, and defect candidate points are extracted. Judgment of defect candidate
The determination is performed on the entire difference image using a preset threshold value or a determination threshold value obtained from the brightness of the image to be inspected. As another determination method, a method may be considered in which a determination threshold is calculated for each coordinate or brightness of an image, and determination is performed with a different determination threshold at each point of the image.

Since the binarized image indicating the defect candidate also includes a false report, in order to extract only the defect as much as possible,
The feature amount extraction unit 288b extracts a feature amount from the detected image of the detected defect candidate point, and finds a defect. The feature amount extraction unit 288b calculates feature amounts such as the area, coordinates, and projection length of the defect candidate point, determines whether the defect candidate point is a defect or a false report based on the calculated feature amount, and detects the defect 288c. I do. Next, processing of two images to be compared will be described. In particular, here, in order to compare two images having different brightnesses, the local tone converter 287 positively corrects the brightness, which is one of the feature amounts, for one of the detected images. . The image sensor 8, which has sensitivity to UV or DUV, outputs a brightness image signal corresponding to the brightness of the reflected light from the semiconductor wafer 1, which is the pattern to be inspected, that is, a shading image signal. The digital image signal from the grayscale image signal is output from the unit 9.

The alignment unit 286 aligns the detected image obtained from the gradation conversion unit 10 and the reference image obtained from the delay memory 11 in pixel units based on the normalized correlation. The normalization is performed to reduce the influence of the difference in brightness between images to be aligned. That is, the positioning unit 286 moves the stored image (reference image) g (x, y) with respect to the detected image f (x, y), and the position (Δx) at which the correlation value R (Δx, Δy) becomes maximum. , Δy) is calculated by the following equations (1) to (5) (Δx, Δ
y: integer).

[0056]

(Equation 1)

[0057]

(Equation 2)

[0058]

(Equation 3)

[0059]

(Equation 4)

[0060]

(Equation 5)

Here, the image is continuously detected by the image sensor 8, but the image is divided into small areas, and positioning is performed in this unit. In the above equation, the detected image is X ×
This is the size of the Y pixel. The division into small areas is for dealing with the distortion of the image. That is, the size of the image distortion is determined so that the image distortion is at a level that can be almost ignored within the small area. Although not shown, the above-described normalized correlation for obtaining the image position shift does not need to be performed for all the images. For example, the image is divided into K in the longitudinal direction of the image sensor, and each divided small image is divided. This may be performed on a small image having information among images (the size of X / K × Y pixels). To determine whether or not there is information, for example, each small image is differentiated, the presence or absence of an edge is detected, and a small image having many edges is selected. For example, when the image sensor 8 is a linear image sensor capable of parallel output in a multi-tap configuration, each tap output image corresponds to a small image. This idea is
Images output in parallel are based on the fact that the displacements are equal. Further, the normalized correlation may be obtained independently for each of the divided small areas, and the positional deviation amount obtained for the maximum area may be used. The image sensor 8 used here is a parallel output type time delay integration type TDI CCD which is sensitive to UV or DUV.
It may be an image sensor.

The local tone converter 287 detects the detected image signal f (x, y) and the reference image signal g (x, y) having different brightnesses.
Is a part for converting the gradation of both or one of the image signals so as to match the brightness of the normal part. Here, the gain a (x, y) and the offset b are set for each pixel (x, y) based on the following equations (6) to (9).
A linear conversion is performed by (x, y) to match the brightness of the normal part.

[0063]

(Equation 6)

[0064]

(Equation 7)

[0065]

(Equation 8)

[0066]

(Equation 9)

The comparison processing section 288 compares the detected image signal obtained by the local gradation conversion with the reference image signal, and detects a mismatch as a defect or a defect candidate.

Here, the gradation conversion of the image signal may be performed based on scatter diagrams shown in FIGS. 27 and 28 described later. On a scatter plot, linearly approximate the data with the highest frequency,
The slope and intercept are given by the above equations a (x, y) and b (x, y)
Is equivalent to In the example of FIG. 28, an approximate straight line of an area centered on each point of interest is obtained on a scatter diagram, and a gain a (x,
y) and offset b (x, y) are calculated. In addition, linear approximation is performed for the entire scatter diagram, but multiple scatter diagrams are created according to the contrast value and variance value of the detected image in a local region, and the data on each scatter diagram is linearly It may be approximated. That is, for example, a range filter (rang) provided inside the scatter diagram creation unit 293.
e filter) calculates and outputs a difference between a maximum value and a minimum value of brightness in a neighboring region of the target pixel sequentially cut out from the detected image (a value of a category called contrast) and outputs the difference. The values are sequentially provided to the CPU 290 and stored in the internal memory. As the contrast detection, a percentile filter other than the range filter can be used according to the S / N.

Accordingly, the scatter diagram creator 293 performs the pipeline image processing based on the two images of the predetermined area input from the image input unit 292 as shown in FIG.
A scatter diagram for each contrast value can be created and provided to the CPU 290. As described above, by decomposing the scatter diagram according to the contrast value, the spread of data on each scatter diagram can be suppressed, and as a result, the determination threshold can be reduced. In many cases, as seen in the case of FIG. 32 described later, the reason is that the higher the contrast, the smaller the variance. Then, the CPU 290
Is the slope 45 in each scatter diagram as shown in FIG.
The data is divided into segments (categories of light and shade differences) parallel to the straight line of the degree, and linear approximation is performed using normal categories (lumps with the highest frequency of occurrence) in the vicinity of each of the divided data. That is, as shown in FIG. 31, the CPU 290 sets the target pixel of each scatter diagram to
In accordance with the “nearest neighbor decision rule”, an approximate straight line is determined using data of the neighboring majority (normal) category group.

The method of dividing the value of the contrast, the value of the variance, and the value of the density difference is determined in advance by experiments. By the way, as shown in FIG. 32, the scatter diagram creator 293 calculates the frequency of occurrence in the classification category between the determined contrast value and the contrast value, and calculates it as C
By the PU 290 displaying on the display means 294,
Enables quick view of sensitivity. In particular, from FIG. 32, it can be seen that the variance decreases as the contrast increases, and the determination threshold can be reduced. As described above, the category information can be mapped and displayed on the screen of the display unit 294 as shown in FIG. Further, the scatter diagram is decomposed according to the contrast values shown in FIG. Then, the local gradation conversion unit 287
Is, for example, for a detected image obtained from the positioning unit 286, the pixel f (x, y) is converted into a least square approximation straight line (gain
Based on gain (x, y) and offset offset (x, y)), a corrected detection image f '(x, y) is obtained by performing local gradation conversion by the following equation (Equation 10). As described above, for example, the detected image is subjected to local gradation conversion based on each scatter diagram (FIG. 3).
As shown in FIG. 1, the distribution is rotated. ) To perform the brightness correction (color unevenness correction), it is possible to further suppress the dispersion of data in the normal part, it is possible to further lower the determination threshold, and to reduce the defect in the normal part. And can be detected separately. That is, the distance between the normal part and the defect can be increased by the scatter diagram decomposition and the gradation conversion.

F ′ (x, y) = f (x, y) × gain
(X, y) + offset (x, y) (Equation 10)
(X, y) indicates the slope of the approximate straight line, and offset (x, y)
Indicates the intercept of the approximate straight line. Thereafter, in the comparison processing unit 288, for example, a difference image between the gradation-converted detected image and the reference image is extracted, and the difference image is determined with the lowest determination threshold to detect a defect or a defect candidate. Finally, the defect and its characteristic amount are output.

As described above, the scatter diagram is decomposed according to the output of the range filter, and the detected image is subjected to gradation conversion based on each scatter diagram to perform brightness correction (color unevenness correction). It is possible to suppress the spread of the data in the normal part and set the lowest lower limit as the determination threshold value, and detect an ultra-fine defect of about 50 nm buried in the scatter diagram without erroneous detection.

Next, the simultaneous correction of the position adjustment in the position adjustment unit 286 and the brightness in the local gradation conversion unit 287 will be described. That is, the scatter diagram data having a large output (contrast) of the range filter shown in FIG. 33A indicates the correspondence of the pattern edges to be compared. The reason is that a large contrast is detected from the edge of the pattern formed on the inspection object. Therefore, if the positioning in the positioning unit 286 is poor, the scatter diagram data spreads as shown in FIG. Accordingly, the CPU 290 checks the spread of the scatter diagram data, calculates the minimum amount of positional shift as shown in FIG. 33B, and uses the calculated positional shift amount as a positioning unit. By performing the sub-pixel matching by feeding back to 286, it is possible to simultaneously perform the alignment correction in the alignment unit 286 and the brightness correction in the local tone conversion unit 287 described above. As a result, the variance can be minimized, the determination threshold value can be minimized, and the sensitivity can be further improved, as compared with the brightness correction alone.

The calculation of the displacement amount may be performed for each disassembled scatter diagram and integrated.

Next, the operation of the inspection apparatus having the above configuration will be described. In FIG. 1, the image sensor 8 illuminates the semiconductor with an image sensor 8 while illuminating with UV light so as not to cause speckle, moving the stage 2 in the X direction and moving the target area of the semiconductor wafer 1 of the pattern to be inspected at a constant speed. A pattern to be inspected formed on the wafer 1, that is, brightness information (shade image signal) of a memory mat portion and a peripheral circuit portion in a chip is detected. Naturally, in the inspection object 1,
Even if the pattern is not formed, the detection image and the reference image to be compared are effective because speckle noise does not occur. When the movement for one row is completed, the head is moved to the next row in the Y direction at a high speed and positioned. That is, the inspection is performed by repeating the constant speed movement and the high speed movement. Of course, a step-and-repeat inspection can be used. And
The A / D converter 9 converts the output (shade image signal) of the image sensor 8 into a digital image signal 285. This digital image signal 285 has a 10-bit configuration. Of course, if there are about 6 bits, there is no particular problem in image processing, but a certain number of bits is required to detect a minute defect.

Semiconductor wafer 1 obtained based on design information
By inputting the coordinates of the array data and the like in the chip on the input unit 289 composed of a keyboard, a disk, and the like, the CPU 290 can input the coordinates of the array data and the like in the chip on the semiconductor wafer 1. , The defect inspection data is created and stored in the storage device 291. Defect reliability, which indicates the likelihood of a defect described later, is also added to the defect inspection data obtained from the comparison processing unit 288 and stored in the storage device 291.

The defect inspection data can be displayed on the display means 294 such as a display as necessary, together with the defect reliability, or can be output on the output means 29 such as a printer.
5 can also be output. In addition, various inspection devices, optical review devices, SEM-type review devices, and defect classifiers (devices that classify defects into defect categories by focusing on their feature amounts. Devices that use neural networks, etc. And the like, defect inspection data and defect reliability can also be sent. Of course, only the defect reliability may be displayed and output. Image input unit 292
Inputs a detected image and a reference image to be compared obtained from the positioning unit 286, and obtains a scatter diagram in the scatter diagram creating unit 293. FIG. 27 shows how to obtain a scatter diagram. In the scatter diagram, the vertical axis and the horizontal axis represent two detected images f (x,
y), and the brightness of the reference image (x, y). The scatter diagram includes, as shown in FIG. 39, a local brightness gradient (differential value) and a brightness variation (brightness) as shown in FIG. (Standard deviation), texture, and the like may be respectively set as the vertical axis and the horizontal axis.

As described above, in the scatter diagram, the frequency is converted into a gray scale value and displayed as shown in FIG. Here, the frequency is 0
Is displayed in gray, low frequency in white, and high frequency in black. Of course, the scatter diagram may display only the presence or absence of data. As shown in FIG. 26, the calculation unit 2 in the CPU 290
90a is based on the scatter diagram created by the scatter diagram creating unit 293, based on the frequency of the defect on the scatter diagram, the position of the defect on the scatter diagram, or relative distance information on the defect on the scatter diagram, or The information referring to the look-up table 290b is calculated. The information calculated in this way is used as the defect reliability as information on the defect obtained from the comparison processing unit 288 (the coordinates of the defect and the difference between the defect area, the defect length, and the defect brightness, which are the characteristic amounts of the defect). Etc.) and stored in the storage device 291.

Here, a high frequency in the scatter diagram indicates that the point is not likely to be defective. For example, in FIG. 27, pixels corresponding to black data on a scatter diagram have a high frequency, and these pixels have a high probability of being normal portions. On the other hand, a pixel corresponding to white data has a low frequency, indicating that the brightness is only a small number, and the frequency of a defect is high. Thus, it can be said that the frequency information is an important parameter indicating the probability of the defect. Similarly, regarding the position on the scatter diagram, if the two detected images to be compared and the reference image have the same brightness, each point is distributed on a straight line having a slope of 45 degrees. It is a parameter of the probability of an important defect. In FIG. 27, it can be seen that the pixels corresponding to the data that are separated from the straight line (not shown) having a slope of 45 degrees have a low frequency and are highly likely to be defective. That is, the CPU 290
By obtaining a straight line that minimizes the square error on the scatter diagram obtained according to the feature amount (for example, brightness, differential value, standard deviation, etc.) of each point created by the scatter diagram creating unit 293, Defect reliability information (relative distance of defect) can be obtained from the distance from the straight line. That is, as shown in FIG. 28, a scatter diagram is obtained according to the feature amount of each point,
An approximate straight line is obtained for the data of each plane. In particular, by using the fact that frequency is a parameter indicating the likelihood of a defect, in two compared images, each point having a frequency equal to or higher than a certain value is weighted by using a plurality of surrounding pixels that are determined. A straight line that minimizes the squared error is obtained. The size of the area is locally variable according to the frequency of the scatter plot. The variable method is desirable because the method of inputting the frequency and outputting the area size with reference to the look-up table is flexible.

The CPU 290 obtains a gain (slope of the approximate line) and an offset (intercept of the approximate line) from the approximate line thus obtained, and obtains a local tone conversion unit 287.
Feedback to The local tone conversion unit 287 performs brightness correction on the detected image f (x, y) based on the gain and the offset. Next, the comparison processing unit 288
Extracts a difference image between the brightness-corrected detected image and the reference image in this way, determines the extracted difference image with a determination threshold value, and outputs a defect.
Numeral 290 is for obtaining a distance from an approximate straight line for a defect in the scatter diagram, and outputting or displaying the distance as a probability of the defect. The smaller this distance is,
It is closer to the normal part, and the larger is the closer to the defect. That is, in the scatter diagram, the frequency decreases as the distance from the approximated straight line increases, and the accuracy of the defect increases. Each point having a frequency equal to or higher than a certain value means, for example, a point having a frequency equal to or lower than 1 is regarded as having a high degree of accuracy of a defect and is excluded from linear approximation. Further, the variations Vr and Ve of the entire image from the approximate straight line are calculated by, for example, the following (Equation 11).
It is obtained by the equation and the equation (12). Here, it is assumed that the approximate straight line is Y = m · f (x, y) + n. m is the slope and n is the intercept.

[0081]

[Equation 11]

[0082]

(Equation 12)

The information on these variations can be used as a measure of the degree of coincidence of the entire image. As described above, it is possible to determine the likelihood of the mismatch information output by the inspection device using the information obtained from the scatter diagram. Further, using the input means 289, for example, a determination threshold value for determining whether or not the absolute value of the difference image is a defect is input, and a line segment of the input determination threshold value is displayed on the display unit 29.
By plotting on the scatter plot shown in 4,
The validity of the inputted determination threshold can be determined on the scatter diagram. In addition, a determination threshold suitable for an image can be determined with reference to the information of the displayed scatter diagram. That is, by determining the determination threshold based on the probability of the above-described defect, it becomes possible to detect the defect with higher reliability. For example, the determination threshold is determined adaptively for each pixel, and the determination threshold is determined according to the frequency of the scatter diagram. The conversion between the frequency and the determination threshold is performed using a look-up table (LUT) 290b as shown in FIG. It is assumed that the contents of the lookup table 290b, that is, the way of conversion are determined before the inspection.

The images used in the scatter diagram are two detected images to be compared and a reference image, for example, an image after pixel-by-pixel alignment, but at each stage of image processing, two images are used. Can be input to the image input unit 292. FIG.
2 illustrates an example in which two detected images and a reference image (stored image) are processed based on the method illustrated in FIG. 1. The target is a pattern to be inspected subjected to a flattening process such as CMP (Chemical Mechanical Polishing), and a line and space pattern is detected at the lower right of the image. The upper left is an area without a pattern. The histogram of the image during each process is also shown. As can be seen from the histogram, in the first stage, the brightness of the two images does not match. First, in the positioning unit 286, correlation values of these images are obtained by normalized correlation,
By determining a position where the correlation value is high, positioning is performed in units of pixels. Next, in the local tone conversion unit 287, based on the gain and offset obtained from each scatter diagram provided from the CPU 290, local brightness correction as local tone conversion is performed on the two aligned images. carry out.

FIG. 35 shows a scatter diagram of an image. At the stage of alignment in units of pixels, since the brightness of the two images does not match, the scatter diagram does not follow a straight line at an oblique angle of 45 degrees, but shows variations from the straight line. However, after the processing of the local tone conversion (the formula (6) and the formula (7) or the formula (10)) according to the present invention, the scatter diagram is distributed near a straight line, It can be seen that there is an effect in terms of equalizing the brightness of the two images. The slope and intercept are the slope and intercept of the line segment fitted to the scatter diagram data. The slope, which is a measure of the degree of coincidence between the two images, was initially 0.705, but becomes 0.986 after local brightness correction, which is local gradation conversion, and the degree of coincidence of brightness is improved. You can see that it is. Further, the value of Ve, which indicates the degree of coincidence between two images, was initially 40.02, but after the local brightness correction as a local gradation conversion, 8.5.
98, indicating that the degree of coincidence of brightness is improved.

In these methods, the numerical values of the entire image are calculated for each image to be compared. However, in the method shown in FIG. 27, the above-mentioned Ve or the like may be obtained for each local size to be converted. Such brightness correction processing is particularly effective when two images to be compared have a difference in brightness. The difference in brightness occurs due to the difference in the film thickness of the corresponding patterns to be compared. When the film thickness is different, if the band of the illumination wavelength is narrow, the difference in brightness increases due to thin-film interference. In the present invention, such an influence is reduced by controlling the polarization state, but the difference in the brightness that still remains is solved by the above correction. As a result, extremely small defects of 100 nm or less can be detected. In the example of FIG. 33, using the scatter diagram after the local brightness correction, information on the likelihood of the defect is given to the mismatch according to the above-described procedure. In the scatter plot, the pixels distributed and distributed around are
High accuracy of defects. The determination threshold can be set using a straight line having a slope of 45 degrees so as to sandwich the distributed data. Of course, even at the stage where the alignment is performed in units of pixels in the alignment unit 286, information on the likelihood of a defect can be similarly extracted from the scatter diagram. However, in this case, since the brightness correction has not been performed, the variance of the data in the normal part increases, and as a result, the determination threshold value must be set large so as to sandwich data with a large variance distribution. And cannot be set to high sensitivity. Therefore, it may be more desirable to use the scatter diagram after the local brightness correction to determine the determination threshold.

The scatter diagram creation and display, or the calculation of the judgment threshold value using the data of the scatter diagram is performed in synchronization with the image detection, for each image or for each pixel of the image. Inspection can be realized. Note that, as described above, the image processing is realized by pipeline-type processing, but any other configuration can be applied. FIGS. 36A to 36C show examples of defect output lists.
Shown in The gradation-converted images are compared with each other by the comparison operation unit 288, and the difference image exceeds the determination threshold and is output as a defect or a defect candidate. This output is an example in which defect reliability is added in addition to numerical values representing the characteristics of a defect, such as a defect number, coordinates, length, and area. Here, the defect numbers are numbers assigned in the order in which the chips to be inspected are scanned. The defect coordinates are positions at which a defect is detected in a coordinate system provided with reference to, for example, a mark or an origin of the chip to be inspected or the like. The length of the defect is the X axis
The length of the defect along the axis. Of course, the length along the long axis and the short axis may be calculated.

These units are, for example, microns, depending on the precision required. The defect reliability is information obtained from the scatter diagram described above. For example, the frequency of the defective pixel on the scatter diagram, the distance from the approximate straight line, and the like are shown. FIG. 36A is based on the frequency of defective portions in the scatter diagram. The lower the frequency, the higher the reliability value of the defect. FIG. 36B is based on the distance of the defective portion from the approximate straight line in the scatter diagram. The longer the distance, the higher the reliability value of the defect. FIG. 36 (c)
This is based on the position coordinates of the defective part in the scatter diagram.
The further away from the straight line having the inclination of 45 degrees, the higher the reliability value of the defect. Of course, the defect reliability may include a plurality of frequencies of the defective portion on the scatter diagram, a distance from the approximate straight line, and the like. When the defect has a plurality of pixels, a statistical value such as an average value or a maximum value of the frequency of each pixel or a median is calculated. In this way, by adding the reliability information to the defect information, the reliability information can be used for calculating the fatality of the defect.

A method of detecting a defect while allowing a difference in brightness between two images to be compared will be described. As shown in FIG. 37, the brightness of the two images f and g to be compared is corrected to obtain images F (f) and g. Here, F (•) is a function for converting brightness. Then, these are compared to detect a mismatch as a defect. Next, a method of obtaining a function for converting brightness will be described. As shown in FIG. 37, each point of the image is mapped to “category space”. The category space is a space centered on the feature amount. As the feature amount, for example, a local density difference between two images (for example, a difference in brightness between a target pixel and a corresponding pixel) and a local image contrast (for example, 2 × 2 including a target pixel of an image f including the target pixel). (Maximum value-minimum value in a pixel). A set of points is defined as a segment in the category space. In FIG. 37, segment A,
Segment B exists. Such division into segments is generally called segmentation. Various techniques have been developed for segmentation. Here, based on the method shown in FIG. 28, division is performed based on frequency data in the category space. Needless to say, the division may be made based on the categories of the contrast and the light and shade difference by the method shown in FIG. In the category space, a window (for example, 3 × 3) surrounding each data is set, and the window size (for example, 5 × 3) is determined so that the maximum frequency in the window becomes the set threshold value. An upper limit (for example, 9 × 5) is provided for the window size. Then, the data in the same window is determined to be the same segment. A point with a high frequency belongs to another category, but since a high frequency indicates a normal pattern portion, there is no problem in determining that the segment is another segment. The less frequent points are segments with a wider range (however, the upper limit is set on the window, which corresponds to the largest range), and the possibility of defects increases accordingly.

Next, as shown in FIG. 37, a scatter diagram (here, brightness is set as an axis) of two images is created for each segment. Then, the set of points is approximated by a straight line.
The approximation need not be linear, but may be a polynomial higher-order approximation. This approximate straight line (curve) corresponds to a conversion formula F (•) for converting brightness (gradation). Therefore, an approximate straight line (curve) is obtained for each segment. The brightness of f among the two images f and g to be compared is shown in FIG.
Conversion is performed based on the approximate straight line (curve) of the segment to which each pixel belongs to obtain F (f). Here, since the data in the narrow window belongs to the same segment, the same approximate line is used, and the data in another adjacent window uses another approximate line. Since these are frequent, the distance of each data from the approximation line will also be short.

On the other hand, in the segment to which the window of the less frequent point belongs, the scatter diagram data also varies, the distance of each data from the approximate straight line becomes longer, and the possibility of defects increases accordingly. The segmentation method is not limited to the above-described method, and may be another method without departing from the spirit of the invention. In addition, in the case of linear approximation, if the infrequent data is ignored and not handled, it is possible to prevent the approximation accuracy from being deteriorated by the small number of data. The obtained images F (f) and g are brightness-corrected images to be compared, and these are compared to detect a defect or a defect candidate. Here, since the brightness is corrected, the threshold value for determining the difference between the two images can be set to a small value, and the defect detection sensitivity can be improved accordingly. Note that the two images f and g to be compared have been registered in advance by the positioning unit 286. Although the brightness of one image is corrected here, both may be corrected.

Here, the criticality of a defect indicates the criticality of the defect to the pattern to be inspected, and is determined by, for example, the size of the defect and the coordinates (area) where the defect exists.
The smaller the size of the pattern, the higher the criticality for the same defect size. By using such a fatality determination together with the reliability, it is possible to perform the criticality determination with higher accuracy. Thereby, the process diagnosis of the pattern to be inspected can be performed more accurately.

Next, another embodiment of the comparison operation processing section 288 will be described with reference to FIG. That is, the defect extraction circuit 2881 is a circuit that extracts a defect or a defect candidate based on the binarized image obtained from the comparing unit 288a (the image binarized by the judgment threshold value for the difference image). is there. The gate circuit 2882 includes a defect extraction circuit 2881
This is a circuit that gates the detected image and the reference image output from the local gradation conversion unit 287 based on the signal of the defect or the defect candidate extracted in step (a). Therefore, the detected image of the defect or defect candidate and the reference image are stored in the defect image memory 28.
83. Defect image memory 288
3 and the display memory 2884 are connected to a PCI adapter 2885 via a bus. The processor element (PE) is composed of a plurality of groups (PB0 to PB3). These processor element groups (PE groups)
The above-mentioned 288b and 288c are constituted,
It is connected to a PCI adapter 2885 via a PCI bus, and is connected to a defective image memory 2883 and a display memory 2883 via a link. Further, the management CPU 28
86 is connected to the PCI bus.

With the above-described configuration, each processor element (PE) performs a detailed analysis on a local image including a defect candidate, and calculates a characteristic amount (dimension and area, etc.) of a subpixel of a fatal defect. And categories (foreign matter defect, disconnection defect, short circuit defect, scratch, etc.) are given and extracted. That is, with the configuration shown in FIG. 38, it is possible to execute complicated processing such as false alarm removal in real time. Further, as shown in FIG. 39, regarding the defect determination based on the scatter diagram, the horizontal axis and the vertical axis adopt the above-described feature amounts other than the brightness, and the comparison processing unit 288 realizes an advanced defect determination. For example, a large number of defect candidates (binary images) are obtained by reducing the determination threshold, and are extracted by the defect extraction circuit 2881. Image processing unit shown in FIG. 38 (configured by arranging a plurality of PE groups in parallel)
Realizes high-sensitivity defect determination, such as sieving based on the detected image signals and reference image signals of a large number of these extracted defect candidates, while detecting small defects while preventing erroneous detection. Further, the image processing unit (PE
Group), it is possible to classify defects such as short-circuits, disconnections, or fatalities thereof in parallel with defect detection.

That is, as shown in FIG. 38, an image processing section (PE section) having a plurality of algorithms mounted thereon and capable of selecting these algorithms and judgment thresholds automatically or by designating a plurality of PCI buses. And connected to the link. That is, a plurality of image processing units (PE
Section) forms a defect candidate extracted by the defect extraction circuit 2881 on the sample 1 based on the selected algorithm and the judgment threshold for the detected image signal and the reference image signal obtained through the gate 2882. A defect to be detected of the pattern thus detected is detected. Software (algorithm) installed in the PE unit (processor element) has a function of determining a defect to be detected specified by the user or a false report not desired to be detected. That is, the PE unit can mount a plurality of algorithms according to the user's request. For example, it is realized by selecting and replacing mounted software for each kind of sample and each manufacturing process. This selection may be performed automatically or manually. Each image processing unit (each PE unit) can also change the setting of the determination threshold value according to the selected algorithm. The algorithm has a configuration that can be mounted later.

Here, the above-described image detection and processing are performed at 10 MHz or more for each pixel. Then, at a speed corresponding to a throughput of three or more wafers equivalent to 200 mm in diameter per hour, 100 nm or less (particularly, 5 mm or less).
(0 nm or less), and effective inspection information can be output in an appropriate time in a semiconductor manufacturing line.

Further, the above-mentioned scatter diagram can be used for evaluating the machine difference of the inspection apparatus. That is, the same position of the inspection object 1 is detected by the inspection device A and the inspection image of the inspection device B.
The brightness with respect to the detected image is plotted on the vertical and horizontal axes, and the brightness of the image is plotted. If there is no machine difference, the data will be on a 45 ° straight line, and if there is a machine difference, it will deviate from the straight line. The degree of the machine difference can be evaluated by looking at the degree of the deviation. Judgment can be made based on the magnitude of the variation from the straight line having an inclination of 45 degrees.
Alternatively, the determination may be made based on the magnitude using the above-described variation scales Vr and Ve.

According to the present invention described above, high brightness U
V or DUV illumination can be obtained, a high-resolution image can be captured in a short time, and as a result, a high-speed and high-sensitivity inspection device can be obtained. The detected pattern defect outputs its position and size. In particular, as the inspection object (sample) 1, a conductive metal such as Cu is buried in a via (contact) hole or a groove of a wiring formed in an insulating film such as SiO ₂ by film formation, and CMP or the like is performed. There is a damascene such as Cu in which an excess deposited portion is removed by polishing to form a hole buried wiring. Therefore, the inspection method and apparatus according to the present invention can be applied to damascene such as Cu. In addition, the inspection method and apparatus using the DUV light (light of 266 nm or 248 nm) according to the present invention is 0.07 μm
0.0 when applied to devices below design rules
This is very effective in that an ultrafine defect smaller than 7 μm can be detected.

[0099]

According to the present invention, an image having a quality equal to or higher than that of ordinary discharge tube illumination can be obtained with a short wavelength illumination which is essential for high resolution and a laser light source which is advantageous for its practical use. The effect of being able to obtain at high speed and to be able to detect minute defects with high sensitivity is obtained.

[Brief description of the drawings]

FIG. 1 is a perspective view showing a schematic configuration of a defect inspection apparatus for a pattern to be inspected according to the present invention.

FIG. 2 is a diagram illustrating an emission spectrum of discharge tube illumination.

FIG. 3 is a diagram showing an illumination state on a pupil and a visual field of a detection objective lens by discharge tube illumination.

FIG. 4 is a diagram showing an illumination state on a pupil and a field of view of a detection objective lens by laser illumination, a pattern on the field of view, and a detection signal therefrom.

FIG. 5 is a diagram showing an illumination state on a pupil and a field of view of a detection objective lens by laser illumination expanded on a pupil.

FIG. 6 is a diagram showing an illumination state on a pupil and a field of view of a detection objective lens by laser illumination according to the present invention.

FIG. 7 is a diagram showing a relationship between a CCD detector and an illumination area in a visual field according to the present invention.

FIG. 8 is a diagram showing a relationship between a CCD detector and an illumination area in a visual field according to the present invention.

FIG. 9 is a view showing a CCD detector on a pupil and a field of view of a detection objective lens by laser illumination according to the present invention and an illumination state.

FIG. 10 is a diagram showing a TDI detector on a pupil and a field of view of a detection objective lens by laser illumination according to the present invention and an illumination state.

FIG. 11 is a perspective view showing a schematic configuration of a glass rod lens group according to the present invention.

FIG. 12 is a perspective view showing a schematic configuration of a multi-cylindrical lens array according to the present invention.

FIG. 13 is a diagram illustrating a situation in which annular illumination is performed by laser illumination according to the present invention.

FIG. 14 is a side view of the TDI image sensor according to the present invention.

FIG. 15 is a diagram illustrating an intensity distribution of a beam from a laser light source.

FIG. 16 is a schematic sectional view of a laser illumination optical system according to the present invention.

FIG. 17 is a perspective view showing a schematic configuration of a laser illumination optical system according to the present invention.

FIG. 18 is a perspective view showing a schematic configuration of a laser illumination optical system according to the present invention.

FIG. 19 is a schematic front cross-sectional view showing a schematic configuration of a mechanism for scanning a pupil with laser illumination according to the present invention.

FIG. 20 is a front view illustrating a schematic configuration for reducing spatial coherence of laser illumination according to the present invention.

FIG. 21 is a front view illustrating a schematic configuration for reducing spatial coherence of laser illumination according to the present invention.

FIG. 22 is a front view illustrating a schematic configuration for reducing spatial coherence of laser illumination according to the present invention.

FIG. 23 is a front view illustrating a schematic configuration for reducing the spatial coherence of laser illumination according to the present invention.

FIG. 24 is a front view illustrating a schematic configuration for reducing spatial coherence of laser illumination according to the present invention.

FIG. 25 is a block diagram showing a schematic signal flow for detecting a defect according to the present invention.

FIG. 26 is a block diagram showing a schematic signal flow for determining a determination threshold value according to the present invention.

FIG. 27 is a diagram illustrating a scatter diagram according to the present invention.

FIG. 28 is a diagram illustrating brightness correction according to the present invention.

FIG. 29 is a diagram illustrating an embodiment of the present invention in which the spread of data in a normal part is suppressed by decomposing a scatter diagram according to categories (contrast values).

30 is a diagram for explaining linear approximation using a normal category based on a category classification based on the nearest neighbor decision rule for each scatter diagram shown in FIG. 29;

FIG. 31 is a diagram for explaining gradation conversion correction (brightness correction) of a detected image signal based on a least square approximation straight line obtained from each scatter diagram according to the present invention.

FIG. 32 is a diagram showing an example of classification categories and their frequencies according to the present invention.

FIG. 33 is a diagram for describing simultaneous correction of displacement and brightness (gradation value) in sub-pixel units according to the present invention.

FIG. 34 is a diagram illustrating a flow of defect determination including displacement correction and local brightness correction according to the present invention.

FIG. 35 is a diagram illustrating an example of a scatter diagram according to the present invention.

FIG. 36 is a diagram illustrating a defect output according to the present invention.

FIG. 37 is a diagram illustrating a method for obtaining a function for converting brightness according to the present invention.

FIG. 38 is a diagram showing a specific configuration of a comparison processing unit according to the present invention.

FIG. 39 is a diagram illustrating an embodiment of disassembling a scatter diagram with various feature amounts as axes according to categories (for example, contrast values) according to the present invention.

[Explanation of symbols]

1. Inspection sample (semiconductor wafer), 2. Stage, 3.
Laser light source (UV light source), 4 ... Coherence reducing optical system (Coherence reducing optical system), 5 ... Beam splitter (Polarizing beam splitter), 6 ... Polarizing element group, 7 ... Objective lens, 7a ... Pupil (pupil plane) , 8 ... detector (image sensor:
TDI sensor), 9 A / D converter, 10 gradation converter, 11 delay memory, 19 signal processing circuit, 21 beam forming mechanism (beam expander), 22 first condenser lens (f) -Θ lens), 25 ... Lens array (cylindrical lens or its array), 26 ... Diffusion plate,
195, 198 scanning mechanism (scanning mirror), 199 second condenser lens, 231 first pupil conjugate plane, 232 second projection lens, 233 second pupil conjugate plane, 235 UV parallel light flux, 241 Polarizing element, 242: analyzer, 252: light source group, 253: annular illumination, 285: positioning unit, 28
7 Local gradation conversion unit, 288 Comparison processing unit, 289 Input means, 290 CPU, 291 Storage device, 292
Image input unit, 293 scatter diagram creation unit, 294 display unit (display), 295 output unit, 2881 defect extraction circuit, 2882 gate circuit, 2883 defect image memory, 2884 display memory, 2885 PCI adapter , PE: Pro sensor element (image processing unit).

Continued on the front page (51) Int.Cl. ⁷ Identification symbol FI Theme coat II (reference) G01B 11/24 F (72) Inventor Yukihiro Shibata 292 Yoshida-cho, Totsuka-ku, Yokohama-shi, Kanagawa Pref. Hitachi, Ltd. Production Technology In the laboratory (72) Inventor ▲ Yoshi ▼ Minoru 292 Yoshida-cho, Totsuka-ku, Yokohama-shi, Kanagawa Prefecture Inside the Hitachi, Ltd. (72) Inventor Toshihiko Nakata 292, Yoshida-cho, Totsuka-ku, Yokohama-shi, Kanagawa Prefecture Incorporated Hitachi Manufacturing Co., Ltd. (72) Hiroaki Shishido 292, Yoshida-cho, Totsuka-ku, Yokohama, Kanagawa F-term in Hitachi, Ltd. Production Technology Research Laboratory (Reference) 2F065 AA49 AA56 BB02 BB18 CC18 CC19 CC25 DD04 DD06 DD09 DD16 EE00 EE08 EE09 FF42 FF48 FF49 FF67 GG04 GG22 LL05 LL12 LL30 LL34 LL35 LL36 LL37 LL47 CA04 CB01 CC13 DA07 DA08 EA04 EA08 EA09 EA11 EA12 EA14 EB01 EB02 EC03 EC06 ED07 FA10 4M106 AA01 BA05 BA07 CA39 CA50 DB04 DB07 DB08 DB13 DB14 DB19 DB20 DB21 DB30 DJ04 DJ06 DJ18 DJ20 DJ21 5B007 AA03 BA01 BA21 DA02 DB02 DB09 DC31

Claims (37)

[Claims] 1. A laser light source for emitting laser light, irradiation means for reducing the coherence of the laser light emitted from the laser light source and irradiating the sample on a sample, and the coherence reduced by the irradiation means. Image detecting means for picking up an image of the irradiated sample and detecting an image signal; and defect detecting means for detecting a defect in a pattern formed on the sample based on information on the detected image signal detected by the image detecting means. A pattern defect inspection device, comprising: 2. A light source for emitting UV light, irradiation means for reducing the coherence of the UV light emitted from the light source and irradiating the sample on a sample, and irradiation with reduced coherence by the irradiation means. Image detecting means for capturing an image of the sample and detecting an image signal, and defect detecting means for detecting a defect of a pattern formed on the sample based on information on the detected image signal detected by the image detecting means. A pattern defect inspection device characterized by the above-mentioned. 3. A UV laser light source for emitting a UV laser light, irradiation means for reducing the coherence of the UV laser light emitted from the UV laser light and irradiating the sample onto a sample, and polarized light for controlling the state of polarization. A control unit, an image detecting unit that irradiates the sample with the coherence reduced by the irradiating unit and detects an image signal by capturing an image of a sample whose polarization state is controlled by the polarization controlling unit, and an image detecting unit. A defect detecting means for detecting a defect of a pattern formed on the sample based on information on the detected image signal detected. 4. A light source for emitting UV light, irradiation means for reducing the coherence of the UV light emitted from the light source and irradiating the sample via an objective lens, and the coherence is reduced by the irradiation means. Image detecting means for imaging the reduced and irradiated sample through the objective lens and detecting an image signal; and a pattern formed on the sample based on information on the detected image signal detected by the image detecting means. A pattern defect inspection apparatus comprising: a defect detection unit configured to detect a defect. 5. An irradiating means comprising: an objective lens for irradiating a sample, a condensing optical system for condensing the light on a pupil of the objective lens, and a light spot or light spot condensed by the condensing optical system. The pattern defect inspection apparatus according to claim 1, further comprising: an optical scanning unit configured to scan a light beam on the pupil. 6. The pattern defect inspection apparatus according to claim 5, wherein said optical scanning section is constituted by a mirror which rotates. 7. The pattern defect inspection apparatus according to claim 1, wherein said image detecting means comprises an accumulation type image sensor means. 8. The sample detecting device according to claim 1, wherein the defect detecting unit compares the detected image signal detected by the image detecting unit with the reference image signal stored in the storing unit. 5. The pattern defect inspection apparatus according to claim 1, further comprising: a defect detection unit configured to detect a defect of a pattern formed in the pattern inspection apparatus. 9. A defect detecting means, comprising: a storage unit for storing a reference image signal; a brightness of a normal part of the detected image signal detected by the image detecting means; and a normality of the reference image signal stored in the storage unit. A brightness correction unit that corrects the brightness of the image signal so that the brightness in the unit becomes almost the same, and compares the detected image signal corrected by the brightness correction unit with the reference image signal to the sample. The pattern defect inspection apparatus according to claim 1, further comprising: a defect detection unit configured to detect a defect of the formed pattern. 10. A defect detecting means, comprising: a storage unit for storing a reference image signal; a feature amount of a detected image signal detected by the image detecting unit in a normal part; and a reference image signal stored in the storage unit. A scatter diagram creator that creates a scatter diagram showing a correspondence relationship with the feature amount in the normal portion, and a tone converter that corrects the tone value of the image signal based on the scatter diagram created by the scatter diagram creator. A defect detection unit for detecting a defect of a pattern formed on the sample by comparing the detected image signal corrected by the gradation conversion unit with a reference image signal. Or the pattern defect inspection device according to 3 or 4. 11. A polarizing plate comprising: a quarter-wave plate or a half-wave plate and a quarter-wave plate disposed in an optical path connecting the UV laser light source and the surface of the sample; 4. The pattern defect inspection apparatus according to claim 3, further comprising one or both of an analyzer arranged in an optical path connected to a detector of the image detecting means. 12. A pattern defect inspection apparatus according to claim 11, wherein at least one of said half-wave plate or quarter-wave plate and said analyzer is rotatable. 13. The pattern defect inspection apparatus according to claim 2, wherein said image detecting means has a time delay integration type (TDI) image sensor having sensitivity to UV light. 14. The pattern defect inspection apparatus according to claim 13, wherein said time delay integration type (TDI) image sensor is an anti-blooming TDI sensor. 15. The pattern defect inspection apparatus according to claim 13, wherein said time delay integration type (TDI) image sensor is a backside illumination type TDI sensor. 16. The pattern defect inspection apparatus according to claim 1, wherein said defect detection means outputs information on a position and a size of a defect of a pattern to be detected. 17. A laser light source for emitting laser light, irradiation means for reducing the coherence of the laser light emitted from the laser light source, and irradiating the sample on the sample, and an image of the sample irradiated by the irradiation means. And an image processing unit for processing an image of the sample detected by the image detecting unit. The processing unit processes a wafer having a diameter of 200 mm at a speed corresponding to a throughput of three or more wafers per hour. A pattern defect inspection apparatus for detecting a pattern formed on the sample including a defect of 100 nm. 18. A method for detecting a defect in a pattern formed on a sample, comprising a plurality of image processing units equipped with a plurality of algorithms and capable of selecting these algorithms and judgment threshold values automatically or by designation. A pattern defect inspection apparatus characterized by the following. 19. A sample on which a pattern is formed by irradiating UV light with reduced coherence, an image of the sample irradiated is obtained to obtain a detection image signal, and the obtained detection image signal is referred to. A pattern defect inspection method, wherein a pattern defect is detected by comparing the pattern defect with an image signal. 20. A method in which the UV light is focused on a pupil of an objective lens and scanned to irradiate the UV light onto a sample on which a pattern is formed with reduced coherence, and the irradiated sample is imaged. A detected image signal, and comparing the obtained detected image signal with a reference image signal to detect a pattern defect. 21. Irradiating a sample on which a pattern is formed with UV light with reduced coherence, imaging the illuminated sample to obtain a detection image signal, and normalizing the obtained detection image signal. The brightness of the image signal is corrected so that the brightness in the portion and the brightness in the normal portion of the reference image signal are substantially the same, and the detected image signal subjected to the brightness correction is compared with the reference image signal to determine the pattern. A pattern defect inspection method characterized by detecting a defect. 22. Irradiating a sample on which a pattern is formed with UV light with reduced coherence, imaging the irradiated sample to obtain a detection image signal, and referencing the obtained detection image signal and a reference signal. Align the image signal with at least the pixel unit,
The tone value of the image signal is corrected so that the tone value in the normal portion of the aligned detected image signal and the tone value in the normal portion of the reference image signal are substantially the same, and the tone value correction is performed. A pattern defect inspection method comprising comparing a detected image signal and a reference image signal to detect a defect in a pattern. 23. Irradiating a sample on which a pattern has been formed with UV light with reduced coherence, imaging the irradiated sample to obtain a detection image signal, and normalizing the obtained detection image signal. A scatter diagram showing the correspondence between the feature value in the portion and the feature value in the normal portion of the reference image signal, and correcting the tone value of the image signal based on the created scatter diagram; A pattern defect inspection method comprising comparing an image signal and a reference image signal with a judgment threshold value obtained from the scatter diagram to detect a pattern defect. 24. The coherence of a laser beam emitted from a laser light source is reduced, and the laser beam having the reduced coherence is temporally changed on an object on which a pattern is formed via an objective lens. Irradiating while changing the pattern to obtain a detected image signal by imaging the irradiated sample, and comparing the obtained detected image signal with a reference image signal to detect a pattern defect. Defect inspection method. 25. The pattern defect inspection method according to claim 19, wherein when the sample is imaged to obtain a detection image signal, the state of polarization is controlled to image the sample. 26. The pattern defect inspection method according to claim 19, wherein when irradiating the sample with the UV light, the UV light is irradiated while controlling the state of polarization. 27. The pattern defect inspection method according to claim 19, wherein when the sample is imaged, the image is imaged by a time delay integration type (TDI) image sensor. 28. The pattern defect inspection method according to claim 19, wherein said sample is imaged by an anti-blooming time delay integration type (TDI) image sensor. 29. The pattern defect inspection method according to claim 19, wherein when imaging the sample, the image is imaged with a back-illuminated time-delay integration (TDI) image sensor. 30. The pattern defect inspection method according to claim 19, wherein when detecting a defect of the pattern, information on a position and a size of the detected defect is output. 31. When irradiating the sample with the UV light, NA
24. The pattern defect inspection method according to claim 19, wherein the irradiation is performed through an objective lens of 0.75 or more. 32. The pattern defect inspection method according to claim 19, wherein a pattern having repeatability is formed on said sample. 33. The pattern defect inspection method according to claim 19, wherein a pattern having a design rule of 0.07 μm or less is formed on said sample. 34. The apparatus according to claim 19, wherein when comparing the detected image signal with the reference image signal, false information is significantly reduced even in a portion where the pattern of the sample is not formed. The described pattern defect inspection method. 35. A method for inspecting a defect of a pattern formed on a sample, the method comprising irradiating a surface of the sample with UV laser light having reduced coherence, and irradiating the sample with the UV laser light. An image signal is obtained by imaging the surface of the sample, a defect of 100 nm or less on the sample is detected by processing the image signal, and information on a position of the detected defect of 100 nm or less on the sample is output. Pattern defect inspection method. 36. The pattern defect inspection method according to claim 35, wherein, when obtaining an image signal by imaging the surface of the sample, imaging is performed by controlling the state of polarization. 37. Irradiation of a UV laser beam having reduced coherence onto a wafer having a diameter of 200 mm, imaging of the irradiated wafer, detecting an image of the wafer, and processing the detected image of the wafer. 100 of the pattern formed on the wafer
2. A pattern defect inspection method comprising: detecting defects of 3 nm or less at a throughput of three or more wafers per hour.