JPH09213756A - Semiconductor manufacturing method and manufacturing line thereof - Google Patents

Abstract

(57) 【Abstract】 PROBLEM TO BE SOLVED: To realize 100% inspection of processed substrates on a semiconductor manufacturing line or a sampling inspection at a sufficient inspection frequency.
A semiconductor manufacturing method and a manufacturing line for efficiently managing the manufacturing line are provided by combining the inspection data and other data. SOLUTION: A foreign matter inspection device composed of a means for detecting a regular array of patterns formed on a substrate and a means for extracting an irregular component other than the regular array and its detection data are collected. It is composed of means, means for collecting data on process parameters or work expected to be related to the generation of foreign matter, and means for processing such information.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention is a manufacturing process for manufacturing an object by forming a pattern on a substrate, such as a semiconductor manufacturing process, a liquid crystal display device manufacturing process, a printed circuit board manufacturing process,
BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method and apparatus for analyzing foreign matter and defect occurrence status in a manufacturing process for detecting foreign matter and defects that occur and taking countermeasures, or a manufacturing method and manufacturing line using an inspection apparatus for inspecting foreign matter and defects on a substrate .

[0002]

2. Description of the Related Art In the conventional semiconductor manufacturing process, the presence of foreign matter on a semiconductor substrate (wafer) causes defects such as poor insulation of wiring and short circuits. Further, semiconductor elements are miniaturized and minute foreign matter remains in the semiconductor substrate. When present, the foreign matter also causes damage to the capacitor insulating film and gate oxide film. These foreign substances have various causes such as those generated from the operating part of the transfer device, those generated from the human body, those generated by reaction in the processing device by the process gas, and those mixed with chemicals or materials. Are mixed in various states.

Even in the same liquid crystal display element manufacturing process, if a foreign substance is mixed into the pattern or some kind of defect occurs, it cannot be used as a display element. The situation is the same in the manufacturing process of the printed circuit board, and the mixture of foreign matter causes a short circuit of the pattern and a defective connection.

As one of the conventional techniques for detecting a foreign substance on a semiconductor substrate of this type, as described in JP-A-62-89336, the semiconductor substrate is irradiated with a laser so that the semiconductor substrate is exposed. By detecting the scattered light from the foreign matter generated when the foreign matter adheres to the substrate and comparing it with the inspection result of the same type of semiconductor substrate that was inspected immediately before, false alarm due to the pattern is eliminated, and high sensitivity and high reliability are achieved. What enables foreign matter and defect inspection is also disclosed in Japanese Patent Laid-Open No. 63-1358.
As disclosed in Japanese Patent Laid-Open No. 48-48, the semiconductor substrate is irradiated with a laser to detect scattered light generated from a foreign substance adhering to the semiconductor substrate, and the detected foreign substance is laser-photographed. Some are analyzed by analytical techniques such as luminescence or secondary X-ray analysis (XMR).

Further, as a technique for inspecting the above-mentioned foreign matter, the wafer is irradiated with coherent light, and the light emitted from the repetitive pattern on the wafer is removed by a spatial filter so as to emphasize the non-repeating foreign matter and defects. A method of detecting is disclosed.

Further, Japanese Patent Application Laid-Open No. 5-218163 discloses a small foreign matter monitoring device capable of detecting foreign matter on a wafer in real time in a semiconductor or the like manufacturing line.

[0007]

The above-mentioned prior art does not distinguish between the mass production line and the mass production line in the semiconductor manufacturing process, and the inspection device used in the mass production startup work is directly applied to the mass production line. Therefore, it is necessary for mass production lines to quickly detect the occurrence of foreign matter and take countermeasures. However, since the conventional inspection device has a large device scale and must be installed independently, the semiconductor substrate, the liquid crystal display element substrate, and the printed circuit board processed in the manufacturing line are brought to the inspection device. The inspection was for foreign matters and defects. Therefore, it takes time to carry these substrates and inspect for foreign matters and defects, and it is difficult to inspect all of them, or it is difficult to obtain a sufficient inspection frequency even in the sampling inspection. Further, such a construction requires manpower.

On the other hand, Japanese Unexamined Patent Publication No. 5-218163 discloses a compact foreign matter monitoring apparatus capable of detecting foreign matter on a wafer in a manufacturing line for semiconductors in real time. However, it does not disclose what to do with semiconductors in relation to process parameters or data relating to operations.

An object of the present invention is to realize 100% inspection of processed substrates on a semiconductor manufacturing line or a sampling inspection with a sufficient inspection frequency, and to combine this inspection data with other data efficiently. An object of the present invention is to provide a semiconductor manufacturing method and a manufacturing line capable of managing the manufacturing line.

[0010]

In order to achieve the above object, according to the present invention, a foreign matter / defect inspection apparatus is arranged in a mass production line, and foreign matter / defect inspection is performed on all or almost all substrates. It was configured to do.

As a concrete configuration for realizing this, a small foreign matter monitoring device capable of detecting foreign matter on a wafer in a line in real time is arranged in a substrate manufacturing line for semiconductors, etc. It was configured to be placed in the transport system between the devices.

That is, according to the present invention, in a substrate manufacturing line of a mass-produced semiconductor or the like having a plurality of processing devices, an illumination system, an imaging optical system, and a spatial filter arranged on a Fourier transform surface of the imaging optical system. And a detector arranged at the image forming position of the image forming optical system for detecting the occurrence of foreign matter on the semiconductor substrate, and a foreign substance monitoring device is provided at the inlet of the predetermined processing device, or the outlet thereof, or a plurality of outlets. A method and an apparatus comprising a foreign matter generation condition analyzing apparatus in a semiconductor manufacturing process, which is installed in a transfer system between processing apparatuses to detect a foreign matter generation state on a semiconductor substrate by the processing apparatuses.

Further, in the present invention, specifically, in a substrate manufacturing line for a semiconductor or the like having a plurality of processing devices, an illumination system coherent in at least one axis direction and means for detecting light from the illuminated substrate. A means for acquiring information on the repetitive pattern on the substrate from the detected light, an imaging optical system, a spatial filter arranged on a Fourier transform surface of the imaging optical system, and the imaging optical system. A foreign matter monitoring device for detecting the generation state of foreign matter on a semiconductor substrate, which includes a detector arranged at an image forming position of the system.

Further, according to the present invention, in an apparatus for inspecting a foreign substance on a semiconductor substrate, an illumination system for illuminating a semiconductor substrate with plane wave light having a substantially single wavelength in a linear shape, and the illumination system is used for illumination. An imaging optical system for forming a reflected light image from the semiconductor substrate, a spatial filter installed in the middle of the imaging optical system so as to block diffracted light from a repetitive pattern on the semiconductor substrate, and an imaging Detector for detecting the generated light image, erasing means for erasing a signal repeatedly generated on the semiconductor substrate among signals detected by the detector, and a semiconductor substrate based on the signal not erased by the erasing means A foreign matter and defect inspection apparatus comprising: a foreign matter detecting unit for detecting the upper foreign matter.

Further, the present invention is characterized in that, in the above-mentioned foreign matter and defect inspection apparatus, the image forming optical system is constituted by a refractive index changing type lens array.

Further, the present invention is an apparatus installed on a transport system for transporting a sample, comprising an illumination optical system for illuminating the sample transported on the transport system, and an illumination on the illuminated sample. A detection optical system that forms an image on a detector by detecting an image at a predetermined position, and a plurality of linear optical devices that are installed at the imaging positions of the illumination optical system and the light source of the illumination by the detection optical system and whose mutual intervals are variable. The foreign matter and defect inspection apparatus is characterized by comprising a spatial filter composed of the light-shielding object and a processing circuit for processing a signal detected by the detector.

Further, according to the present invention, the above-mentioned detection optical system is composed of two Fourier transform lens groups, and when an image of an object is formed, the image side is telecentric. It is a device.

The present invention also has a structure in which the two Fourier transform lens groups are composed of lens groups having different numerical apertures, and one of the Fourier transform lens groups can be replaced with a lens group having different numerical apertures. It is a characteristic foreign matter and defect inspection apparatus.

Further, according to the present invention, the direction of the linear structure of the spatial filter is parallel to the repeating direction of the pattern formed on the sample, and the mechanism for setting the sample substantially parallel to the sample immediately before the inspection is performed. And a mechanism for measuring the tilt of the sample after being installed substantially parallel, and a pattern formed by tilting the illumination optical system and the detection optical system according to the measured tilt so that the direction of the linear structure of the spatial filter is formed on the sample. It is a foreign matter and defect inspection apparatus characterized by having a mechanism for making it parallel to the repeating direction.

Further, in the present invention, the above-mentioned detection optical system has a linear detector, and the detection optical system has at least two or more repeating patterns such as a chip pattern transferred onto the sample within its field of view. With a field of view of this size
A foreign matter and defect inspection apparatus having means for comparing and erasing pattern information other than foreign matter or defects by comparing two or more repeating patterns.

Further, according to the present invention, the comparing and erasing means determines the signal level of one pixel on the detector when judging whether the signal of one pixel on the detector is a foreign substance or a defect. And comparing the signal level of the pixel at the corresponding location of the adjacent repeating pattern captured by the same detector as the detector with the signal level of a plurality of pixels close to the corresponding location, When there is a pixel having the same value as the signal level of the one pixel in the signal levels of the one place or the adjacent place, the signal detected by the one pixel on the detector is the signal from the repeating pattern. The apparatus for inspecting foreign matter and defects is characterized by having a processing means for judging.

Further, according to the present invention, the spatial filter is arranged so that the repeating pitch of the repeating pattern shielded by the spatial filter is larger than several times the width of the area in which the corresponding pixels or adjacent pixels are combined. The foreign matter and defect inspection device is characterized in that the pitch of a linear pattern is set.

The present invention further comprises means for Fourier transforming the signal detected by the detector during the transport,
Foreign matter and defects characterized by having means for calculating the pitch of the Fourier transform image by the repetitive pattern formed on the sample from the calculation result by the Fourier transform means, and means for changing the pitch of the spatial filter according to the calculation result. It is an inspection device.

The present invention further comprises means for continuously changing the pitch of the spatial filter, means for acquiring the signal of the detector while continuously changing the pitch, and pitch for minimizing the acquired signal. The foreign matter and defect inspection apparatus is characterized by having a calculating means and a means for changing the pitch of the spatial filter so as to obtain the calculated pitch.

[0025]

In order to achieve the above object, a foreign matter and defect inspection apparatus is arranged in a mass production line to realize real-time sampling, the foreign matter monitoring apparatus is downsized, and a processing apparatus for a substrate manufacturing line for semiconductors, etc. It is configured so that it can be placed in the transfer system between the input / output port or the processing device. That is, the present invention is directed to an image forming optical system including an oblique illumination system including an illumination array and a lens array or a microlens group, and an image forming optical system in a substrate manufacturing line for mass-production semiconductors including a plurality of processing devices. A foreign matter monitoring device for detecting the occurrence of foreign matter on a semiconductor substrate, which is provided with a spatial filter arranged on the Fourier transform surface and a detector arranged at the image forming position of the image forming optical system. By installing the processing apparatus at the entrance or the exit, or at the transport system between a plurality of processing apparatuses, it is possible to detect the occurrence state of foreign matter defects on the substrate by the processing apparatus in the plate manufacturing process. The cause of the foreign matter defect can be determined by processing the situation of the foreign matter defect occurrence, classifying the generation modes, and analyzing the components of the generated foreign matter defect.

This is because the conventional apparatus has a large scale and requires a long inspection time. In order to realize a real-time monitor using these conventional apparatuses, it is necessary to arrange a large number of large scale apparatuses. It was difficult. In reality,
The limit is to inspect one lot, several lots, or one semiconductor substrate every day. It cannot be said that the foreign matter and defect inspection with such a frequency has detected the occurrence of the foreign matter sufficiently early. That is, it was far from ideal real-time sampling for a mass production line. Therefore, the foreign matter defect inspection apparatus having the configuration of the present invention can reduce the number of processes and equipment in the mass production line by installing necessary and sufficient monitors in necessary and sufficient places.

By implementing a foreign matter inspection apparatus that can be placed on a fully automatic transport system, it is possible to construct an inspection line without human intervention.

One of the main operations for mass production of LSI
One of these is the work of investigating the cause of these foreign substances and taking countermeasures, and detecting the generated foreign substances and analyzing the elemental species is a major clue to the search for the cause. On the other hand, in a mass production line, it is necessary to quickly detect the occurrence of these foreign substances and take countermeasures. When a time elapses from the generation of a foreign substance to the detection of the generation of the foreign substance, the number of defects generated increases and the yield decreases. Therefore, in order to maintain a high yield, it is essential to shorten the elapsed time from the generation of a foreign substance to the detection thereof. That is, the foreign matter defect inspection apparatus that can shorten the sampling time of the monitor enables real-time sampling in a mass production line, and maximizes the effect of the foreign matter and defect inspection.

According to the present invention, the state of generation of a foreign substance on a semiconductor substrate can be detected in real time by installing it at the entrance of the processing apparatus, the exit thereof, or the transfer system between a plurality of processing apparatuses.

Further, according to the present invention, the cause of generation of foreign matter is investigated by using a foreign matter detection / analysis system in which a sampling semiconductor substrate is devised so that the evaluation at the time of mass production start-up proceeds smoothly and quickly. Or take measures against the dust source of the equipment, and the results are fed back to each material, process, device, etc., and the process specifications that easily generate dust are set as the design specifications of the element that is strong against dust generation. ,
Used to make specifications for mass production line inspections and evaluations, install a foreign matter monitor on the semiconductor substrate as needed at locations where foreign matter is likely to occur, or set specifications to monitor only the increase or decrease of specific foreign matter at specific locations. It is a thing. As a result, at the time of mass production startup of the semiconductor manufacturing process, each process, equipment, etc. are evaluated by expensive and high-performance evaluation equipment in order to evaluate and improve (debug) materials, processes, equipment, design, etc. It is possible to reduce the number of line processes and equipment as much as possible, and particularly to reduce the number of inspection and evaluation items to shorten the equipment cost and the time required for inspection and evaluation.

As described above, by separating the mass production line from the mass production line, foreign matter detection, analysis, and
The evaluation device can be operated efficiently, mass production can be started up quickly, and the weight of the mass production line can be reduced by inspecting foreign substances used in the mass production line and using a simple monitoring device with the minimum necessary evaluation equipment. .

Further, in the monitoring device of the above-mentioned mass production line of the present invention, attention was paid to the following method in order to solve the inspection device having the same function as the conventional large-sized device at high speed and small size with the current technology. First, we focused on the repeatability of the memory. Conventionally, a method of removing a repetitive pattern and detecting a defect is known. This method can reliably ensure the detection performance. However, this method is not mentioned to be advantageous in realizing the above monitoring device. Further, the monitor in this case does not need to monitor all points on the semiconductor substrate, and it suffices to monitor the semiconductor substrate at a certain ratio. In the manufacture of a memory having many repeating patterns, the repeating unit of this memory is used. We paid attention to the fact that just monitoring just the effect is great.

In the repeating pattern, when coherent light is emitted, the light is emitted only in a specific direction. That is, in the case of the memory, the light emitted from the repeated portion in a specific direction can be blocked by the spatial filter, and the foreign matter that does not repeatedly occur can be detected with high sensitivity. At this time, if a liquid crystal is used as the spatial filter, the shape of the spatial filter can be arbitrarily changed by turning the liquid crystal on and off, so that it is possible to automatically inspect an arbitrary repetitive pattern.

The reason why the yield at the time of semiconductor manufacturing is improved by the above means is as follows. Through a strict detection experiment on the number of foreign particles on a semiconductor substrate, it was newly found that the number of foreign particles does not increase or decrease gradually but increases or decreases suddenly. In the past, it was thought that the number of foreign substances would gradually increase or decrease, so as described above, one to one lot
Foreign substances and defects were inspected at a frequency of once a day. However, with this inspection frequency, a sudden increase in foreign matter may be overlooked or may be detected after a while while increasing, and a considerable number of defects will occur. That is, in a mass production line, it is necessary to quickly detect the occurrence of foreign matter and take countermeasures, and if time elapses from the generation of foreign matter to the detection of foreign matter generation, the number of defectives increases and the yield decreases. Therefore, it is possible to maintain a high yield by shortening the elapsed time from the generation of a foreign substance to the detection thereof. That is, by shortening the sampling time of the monitor, ideally by sampling in real time, it is possible to maximize the effect of the foreign matter and defect inspection.

Further, in the conventional apparatus, the semiconductor substrate is taken out and inspected, and at this time, a new foreign substance is attached to the semiconductor substrate, which also reduces the yield.
With the foreign matter and defect inspection apparatus according to the present invention, the semiconductor substrate can be inspected without being taken out, so that the yield reduction due to the foreign matter adhered to the semiconductor substrate can be eliminated.

In order to realize a high-speed compact foreign matter inspection device,
The reason why the method using this spatial filter is more suitable than the polarization detection method shown in the prior art (Japanese Patent Laid-Open No. 62-89336) will be described with reference to FIGS.

In the method of illuminating the sample with light and detecting the scattered light from the foreign matter, the scattered light from the pattern formed on the sample surface becomes noise. This noise increases as the size of the pixel (minimum unit detected as one signal) of the detector 2006 increases as shown in FIG. Since the pattern serving as the noise source is formed on almost the entire surface of the sample, the noise increases in proportion to the pixel size.

On the other hand, the larger the number of pixels, the longer the inspection time. Therefore, it is necessary to increase the pixel size in order to realize high-speed inspection. Therefore, it is necessary to increase the pixel size and reduce the noise level. As a method for reducing this noise level, a method using polarization is analyzed in “Analysis of Reflected Light from LSI Wafer Pattern”, Proceedings of the Society of Instrument and Control Engineers, 17-2, 77/82 (1981). Has been done. According to this, the scattered light (noise) from the pattern is utilized by using the polarized light.
Can be attenuated. However, the attenuation rate of scattered light by this method depends on the direction of the detector, as analyzed in the above paper. Therefore, when condensing light emitted in various directions like using an imaging optical system, the attenuation rate is 0.1% to 0.01% when the respective attenuation rates are integrated.
About.

On the other hand, in the method using the spatial filter of the present application, the attenuation rate is 0.001% to 0.0001.
%. The reason for this will be described with reference to FIGS. The wafer 2001 on which the repeated pattern is formed is illuminated with illumination light 2002, and the illuminated region is illuminated by the lens system 200.
3 and 2005 are used to form an image on the detector 2006. Here, FIG. 36 shows the intensity distribution of the emitted light from the pattern on the Fourier transform surface on which the spatial filter 2004 is mounted.
Light emitted from the repeating pattern is concentrated at a position corresponding to the pitch of the pattern. As an example of calculating this concentration ratio, the diffracted light intensity distribution in the case of multiple slits is written by Hiroshi Kubota,
It is explained in "Applied Optics" (Iwanami).

According to this, as the number of slits (in the present application, the number of repetitive patterns simultaneously illuminated) increases, the concentration ratio increases. This ratio can also be calculated using the Fourier transform F []. When the shape of the illuminated pattern is a (x, y), the light intensity distribution at the position of the spatial filter is F [a (x, y)]. If the shape of the spatial filter is p (u, v), then p (u, v) * F [a
(X, y)] is the light that passes through the spatial filter.
In addition, the shape of the figure complementary to the spatial filter is ￣p (u,
v), p (u, v) * F [a (x, y)]
Is a light component shielded by the spatial filter.
The ratio of these two components becomes the previous attenuation rate. When this attenuation rate is calculated when the number of pattern repetitions is 3, it is 0.00
It is about 1%. When the number of repetitions is 5, it becomes about 0.0001%, and if the number of repetitions is further increased, the attenuation rate decreases. Therefore, it is possible to lower the attenuation rate than using polarized light,
The pattern noise can be reduced.

The above calculation is for the case where the pattern shape and other conditions are ideal, and may not necessarily match the actual experimental result. However, the experimental results show that the attenuation rate is reduced by one to three digits compared with the polarization method, and the pattern noise can be reduced.

[0042]

EXAMPLES The structure of a specific example of the online monitor of the present invention will be described below with reference to FIGS.

In this embodiment, as shown in FIG. 1, illumination means 102, detection optical system 103, rotation adjusting mechanism 105, spatial filter unit 106, detector 107, rotation detection means 108, operational amplifier 201, A / D conversion. A detection head 101 composed of a vessel 202, a pitch detection means 212,
The operator processing system 203, the foreign matter data memory 206, the large foreign matter data memory 207, the pattern memory 208, the software processing system 210, the parameter transmitting means 209, the foreign matter memory 211, the coordinate data creating means 232, the microcomputer 229, and the displaying means 230 are included. It

Further, as shown in FIG. 2, the illumination means 102.
Is a beam expander including a semiconductor laser 112, a collimator lens 113, a concave lens 114, and a receiver lens 115, a cylindrical lens 116, and a mirror 11.
8 and the detection optical system is a Fourier transform lens 1
08, spatial filter unit 106, rotation detecting means 1
08, and a Fourier transform lens 111.

As shown in FIG. 3, the spatial filter unit 108 includes coil springs 121 and 122, a plurality of linear spatial filters 141, a coil spring support 119, and a coil spring support 119.
120, guide 125, right-hand thread 127, left-hand thread 12
Screw 126 having eight, worm gears 129, 130,
It is composed of a motor 140. Further, the spatial filter unit 108 includes detectors 123 and 124 for rotation detection.
Is installed.

Further, as shown in FIG. 4, the operator processing system 203 has a 4-pixel adding means 214 and an octalizing means 21.
5. Cut-out means 2 including a plurality of line memories 216
04, buffer memory 217, determination pixel cutout unit 2
18, operator cut-out means 219, 231, a comparison circuit group composed of a plurality of foreign matter comparison circuits 220, a threshold value setting circuit 221, a comparison circuit group composed of a plurality of noise comparison circuits 222, a threshold value setting circuit 223, an OR circuit 224. , 22
5, AND circuits 226, 227, 228.

Further, as shown in FIG. 5, the pitch detecting means 212 comprises an FFT circuit 242, an operator pitch calculating means 241, a filter pitch calculating means 244 and a spatial filter control system 243.

Further, as shown in FIG. 6, the rotation adjusting mechanism 105 includes a rotation guide 151, a rotation bar 152, and a spring 1.
53, piezo element 154, piezo element controller 1
55 and a mount 156.

(Relationship) The substrate 1 is illuminated by the illuminating means 102, and the scattered light or diffracted light from the surface foreign matter, defect or pattern is taken in, and the spatial filter unit 1
An optical filtering process is performed at 06, and the light is detected by the detector 107 in the detection optical system 103. The detected signal is amplified by the operational amplifier 201 in the detection head 101 into a signal having a large impedance and less susceptible to noise,
It is converted into a digital signal by the A / D converter 202 and transmitted to the operator processing system 203. The rotation detecting means 108 measures the rotation direction of the substrate 1, and the rotation adjusting mechanism 105 adjusts the rotation direction in advance. The detection optical system is 10
Since No. 3 has a sufficiently large depth of focus, if the substrate is carried by a carrying system with mechanical accuracy, automatic focusing is basically unnecessary. Specifically, at a wavelength of about 800 nm and a numerical aperture of 0.08, the depth of focus is about ± 100 microns. Of course, there is no problem with having an autofocus mechanism.

The pitch detecting means 212 measures the repeating pitch of the pattern on the substrate 1 and the chip pitch from the detection signal. In the operator processing system 203, the pattern information is removed by utilizing the repeatability of the chip pitch based on the information such as the chip repeat pitch transmitted by the parameter transmitting means 209. The result is stored in the foreign matter data memory 206, the large foreign matter data memory 207, and the pattern memory 208, and further, the parameter transmission means 20.
Based on the information such as the position coordinate chip repeat pitch of the test element group transmitted by 9, the pattern information such as the test element group having no repeatability between chips is removed by the soft processing system 210, and the foreign substance memory 211 stores the pattern information. Is stored. Here, the coordinate data creating means 2
Coordinate data is created by 32 and stored as necessary at the same time as the foreign substance information. The above processing is managed by the microcomputer 229 and displayed by the display means 230.

Further, as shown in FIG. 2, the illumination means 102.
Then, the light from the semiconductor laser 112 is collimated as a plane wave by the collimator lens 113, the concave lens 114, and the receiver lens 115, and illuminates the substrate through the cylindrical lens 116 and the mirror 118. Here, as shown in the figure, the illumination is collimated only in the x direction by the cylindrical lens 116, and is condensed on the substrate in the y direction. In the detection optical system, the light beam Fourier-transformed by the Fourier transform lens 108 is optically filtered by the spatial filter unit 106, and further, the Fourier transform lens 111 causes the detector 107 to perform the filtering.
An image on the substrate is formed on top.

Further, as shown in FIG. 3, in the spatial filter unit 108, while being guided by the guide 125, the screw 12 having the right-hand threaded portion 127 and the left-hand threaded portion 128 is provided.
Coil spring supports 119 and 120 that move by rotation of 6
As a result, the pitch between the black linear spatial filters 141 passed between the coils of the coil springs 121, 122 is changed. Power is supplied from the motor 140 via the worm gears 129 and 130.

Further, the inclinations of the substrate 1 with respect to the detection optical system head 101 of the substrate are measured by the detectors 123 and 124 for detecting rotation which are installed on the spatial filter unit 108. In FIG. 3, it is described that the detection light is incident from the upper direction 2 from the viewpoint of viewability.

Further, the operator processing system 203 shown in FIG.
Then, the pixels 2 × 2 around the detection signal are the 4 pixel addition means 2
14 is added and averaged. This process is intended for stable detection by averaging, but the detection performance (detection sensitivity)
Since it itself falls a little, bypass means are installed so that it can be bypassed if necessary. The added signal is 8
It is octal-valued by the digitizing means 215 and is stored in the buffer memory 217 as two-dimensional image data through the clipping means 204 composed of a plurality of line memories 216. After being stored, necessary data is cut out from the two-dimensional image data by the determination pixel cutout means 218 and the operator cutout means 219, 231, and the cutout data is sent to the comparison circuit.

Here, the detector 107 uses a one-dimensional linear sensor so as to enable high-speed detection by high-speed stage scanning. The line memory 216 and the buffer memory 21 convert the data from this detector into a two-dimensional image.
7, and every time the signal from the detector is sent pixel by pixel, the entire image moves pixel by pixel in the x direction. This is so-called pipeline processing. A comparison circuit group including a plurality of foreign matter comparison circuits 220, a threshold value setting circuit 221, a comparison circuit group including a plurality of noise comparison circuits 222, a threshold value setting circuit 223, OR circuits 224 and 225, AND circuits 226 and 22.
7, 228 extracts the foreign matter signal by the logic described later.

In the pitch detecting means 212 shown in FIG. 5, the FFT circuit 242 subjects the detected image to the Fourier transform processing, and from this result, the operator pitch calculating means 24
1, the operator pitch is calculated by the filter pitch calculation means 244, and the spatial filter pitch is calculated by the filter pitch calculation means 244.
It is sent to 19, 231.

Further, in the rotation adjusting mechanism 105 shown in FIG. 6, the rotation bar 151 is used as a guide for the rotation bar 152.
The detection optical system head 101 in which the
Based on information from 08, piezo element controller 15
The rotation is controlled by the expansion and contraction of the piezo element 154 controlled by 5. The spring 154 is installed so that the piezo element 154 and the rotary bar 152 are in contact with each other.
It is possible to directly fix the piezo element 154 and the rotary bar 152 without 54. With this configuration, the pedestal 1
The detection optical system head 101 is rotationally controlled with respect to the rotation guide 151 installed on the reference numeral 56. Although the drive system using the piezo element is shown here, the piezo element is not necessarily used, and even if the linear movement mechanism using the rotary motor is configured, a motor that can be rotationally driven is used as the rotation guide 151 itself. However, any other rotary or linear drive mechanism using ultrasonic waves may be used. The piezo element is used here because the piezo element is small and has high precision driving performance.

(Principle) The principle of the parameter compression type spatial filter (PRES filter) will be described.

Conventionally, there has been disclosed a technique for detecting foreign matters or defects having non-repeatability by using the repeatability of the pattern on the wafer surface. However, even though the pattern has repeatability, the shape of the diffraction pattern differs depending on the repetition period and the shape of the basic pattern. Therefore, the shape of the spatial filter, which is the light shielding plate, must be changed according to the shape of the target repeating pattern. As a method for changing the spatial filter, a method using a photographic plate has been disclosed. With these methods, it took a long time to create a spatial filter suitable for the target, and a large-scale device was required.

Specifically, as shown in the figure, when illuminated with coherent light, that is, a plane wave from an oblique direction, it is assumed that a diffraction pattern as shown in the figure is observed at the Fourier transform position. In this case, when the pitch of the pattern on the substrate changes, not only the pitches px and py of the diffraction pattern,
The overall phase φ changes. Further, when the basic shape of the pattern on the substrate changes, the arrangement of the dot patterns forming the diffraction pattern also changes. That is, it is difficult to deal with the pattern shape because there are many parameters that describe the Fourier transform plane diffraction pattern.

Here, let us consider a case where, instead of the plane wave as shown in the figure, the x direction as shown in the figure is narrowed down on the sample and the y direction is coherent, that is, a plane wave is illuminated. In this case, the Fourier transform plane does not form an image in the u-axis direction but has a compressed shape in the u-axis direction. As a result, the spatial filter is compressed into a one-dimensional parameter only in the v-axis direction.

Here, the pitch p in the v-axis direction of the compressed diffraction pattern is a pitch corresponding to the pitch in the y-axis direction of the area illuminated on the substrate surface. Further, the thickness w of the diffraction pattern on each line is determined by the numerical aperture sin β of the front Fourier transform lens on the Fourier transform surface. Specifically, it is determined by the exit side numerical aperture of the illumination system and the numerical aperture of the front Fourier transform lens.

Therefore, it is determined if the illumination system and the Fourier transform lens are determined, and is not affected by the pattern on the substrate to be inspected. However, in some cases, such as when changing the numerical aperture of the illumination, it may be better to make the width of the linear spatial filter variable.

Further, in practice, a one-dimensional image sensor capable of continuous scanning of the stage is suitable for realizing high-speed inspection. When this one-dimensional image sensor is used, in order to improve the efficiency of illumination, the one-dimensional sensor shape, that is, the illumination on a straight line on the sample surface is suitable. To achieve such lighting, at least 1
It is necessary to narrow down the direction. That is, the one-way coherent illumination has a great effect on improving the efficiency of illumination intensity.

As described above, the conventional spatial filter has various shapes depending on the pattern shape on the substrate, and in order to cope with the various patterns, the spatial filter corresponding to each pattern is required. Was there. According to the present invention, these different spatial filters can be considered as functions of only the pitch p from a different viewpoint, and the spatial filter having multidimensional parameters is compressed into one dimension. By compressing the dimension of the parameter of the spatial filter in this manner, it is possible to simplify the spatial filter that has a difficulty in responding to the shape change due to the complicated shape, and make it possible to deal with all repetitive patterns.

The above construction is applicable not only to detecting foreign matters or defects on a wafer or a liquid crystal display element, but also to any inspection object for detecting non-repetitive portions from repetitive patterns. is there. Specifically, it can be applied to a semiconductor mask, a reticle, a micromachining component using a semiconductor process, other micromachining components, a printed circuit board, and the like. The present invention, when inspecting these objects, realizes high-speed inspection by realizing illumination with high illuminance while applying spatial filtering technology without exchanging spatial filters for each object. .

(Spatial Filter Control, Operator Pitch Control) FIG. 9 shows a pattern signal removing method using the spatial filter and operator pitch processing. In the present invention, the repeatability of cells having a pitch of several hundreds of microns or less is erased by using a spatial filter, and the repetition of cells having a pitch of several hundreds of microns or more is eliminated between adjacent chips (in some cases, 1
Erasing is performed by using the repeatability (between shots, which means one exposure), and a chip having no repeatability is erased by using coordinate matrix data so as not to be inspected. Here, there are parameters required for each erasing.

Spatial filter pitch is required for erasing by a spatial filter, inter-chip pitch is required for erasing by repeating chips, and pitch position information is required for erasing a chip having no repetition. Therefore,
It is desirable that the detection optical system of the present invention can simultaneously detect at least two chips. That is, the visual field size of the detection optical system must be at least 2 chips or more. Of course, it is only desirable to have this field of view size, and if you want to know the positional relationship of multiple detection optical systems accurately and carry out this comparison process between multiple detection optical systems, Need not have more than 2 chips. However, in view of the required accuracy of the optical system and the complexity of the circuit system for data processing, it is desirable that the field of view has a size of 2 chips or more.

Although two or more chips are described here, when two or more chips are written on the reticle used as a mask when the pattern is transferred onto the wafer by the stepper, these chips are written. In many cases, patterns called test element groups (TEG) are written in between, and in order to erase these patterns as well, when erasing using the above repeating pitch, instead of using the pitch between chips, Shot (1
It is necessary to use the pitch between the pattern printed on one exposure and the pattern on the reticle. Of course, this method is not always necessary, and there is no problem even if the TEG pattern formed in these one shots is erased in the subsequent processing.

These pieces of information are measured in advance, the parameters corresponding to the substrate are selected, and fed back to the foreign matter defect inspection apparatus of the present invention. Therefore, it is necessary to identify the substrate when using this method. For the purpose of this identification, a number or a symbol corresponding to the substrate is written on the substrate. Read this symbol prior to inspection, know the product version number, lot number, product type on the board from the number, know the process from the data of the place where the foreign substance inspection device of the present invention is installed, pitch of spatial filter, threshold Set the value of.

In implementing the foreign matter defect method of the present invention, it is not always necessary to acquire the parameters as described above and send them to the apparatus of the present invention as described above. Rather, as described below, it may be desirable for the device to be uniquely acquired by the present invention. With the above method, it is necessary to know the value of the parameter to be input in advance, but if it is acquired independently, such a trouble is not required. Of course, it is not necessary to read the number printed on the board.

In the present invention, as described above, in order to distinguish a foreign substance or defect attached on a substrate having a complicated background pattern from a background pattern and extract and detect the foreign substance or defect, three-step pattern removal is performed. It has a function. With this pattern removing function, a portion that is actually determined to be a pattern is discarded as an inspection target.
Specifically, repeats with a pitch of several hundreds of microns or less are erased with a spatial filter, repeats with a pitch of several hundreds of microns or more are erased using the repeatability between chips, and chips that do not have repeatability should not be inspected. The data is deleted.

The reason why the area in which the pattern is formed is excluded from the inspection object is as follows.
Even if a pattern is formed, adjacent chips are formed with patterns having the same shape and emitting the same amount of light in the same emission direction. Therefore, if the detected light intensities of the light from these two patterns are compared, it should be possible to inspect for foreign matters or defects even in a region where a pattern having a shape that cannot be erased by the spatial filter is formed. However, in the case of detecting scattered light, the intensity of the detected light is likely to be unstable in these patterns, and if the above-described pattern removal by comparison is performed, a false alarm (pattern information that is not a foreign matter will be detected as a foreign matter). ) Will increase. Therefore, it may be more effective to remove the area where the pattern is formed from the inspection object. That is, in consideration of stability, it should be decided whether to remove the area where the pattern is formed from the inspection object or to perform the foreign matter inspection by comparing the detected light intensities of the light from the adjacent chip patterns. .

(Parameter Acquisition Method) A specific parameter acquisition method will be described below with reference to FIG. At the next point where the wafer is transferred to a position where the detection optical system can capture the repeated pattern of the wafer, the spatial filter control system 21
2 changes the pitch of the spatial filter from the maximum position to the minimum position. At this time, the all-pixel adding circuit 245 adds the signals taken in by the one-dimensional detector 107 to all the values of each pixel, and the pitch at the position where the added value becomes the minimum with respect to the change in pitch is the pitch calculating circuit. 246 selected. This value is sent to the spatial filter drive mechanism 106, and the spatial filter is set to a predetermined pitch.

Further, in selecting the pitch of the spatial filter, it can be calculated by performing the frequency analysis shown in FIG. 5 without changing the spatial filter in this way. The FFT circuit 242 frequency-analyzes the signal detected by the detector at the next point where the wafer is transferred to a position where the detection optical system can capture the repeated pattern of the wafer, and the spatial filter pitch calculation means 244 is obtained from the result of this frequency analysis.
The spatial filter pitch is selected so that the spatial frequency has a peak in the frequency domain.

In this frequency analysis, the fast Fourier transform is the most desirable from the viewpoint of processing speed and the like, but it is not necessarily the fast Fourier transform, and other methods such as Hadamard transform, frequency analysis by integration, and autocorrelation function calculation are available. There is no problem. In this frequency analysis method, not only the pitch of the spatial filter but also the inter-chip pitch (operator pitch) for removing the component that cannot be removed by the spatial filter is simultaneously processed by the operator pitch calculating means 241. It It is desirable to use the largest pitch between chips calculated by frequency analysis and smaller than 1/2 of the visual field of the detection optical system. This is because the maximum one often corresponds to the pitch between shots described below.

After the parameter values for the inspection are set as described above, the inspection is carried out.

The method described above has a problem that the foreign matter inspection cannot be performed on the first portion of the wafer being transferred. On the other hand, there is an effect that the inspection device can be made independent from other signal transmission systems.

(Telecentric Optical System) In the present invention,
As explained above, repeats with a pitch of several hundreds of microns or less are erased with a spatial filter, repeats with a pitch of several hundreds of microns or more are erased using the repeatability between chips, and chips that do not have repeatability are not inspected. It is configured to erase data.

Here, in order to erase the detection signal of the pattern of the chip by utilizing the repeatability between the chips, the detection signal between the chips is compared, and when the difference is larger than a certain value, it is detected as a foreign substance. I am taking it. That is, it is premised that scattered light or diffracted light from the patterns of adjacent chips does not have the same intensity. Therefore, it is necessary to stably detect the light from the corresponding positions of the adjacent chips. However, since the diffracted light from the pattern has directivity, when the field of view is wide and the angle of viewing the lens greatly differs from each position within the field of view, this directivity causes the light intensity to vary depending on the position within the field of view.

Here, the telecentric optical system is a technology developed by making the principal rays from each point on the object parallel to each other so that the magnification of image formation does not change even when the focal position shifts. is there. By using this telecentric optical system in the present invention, the intensity of the detection light from each point of the object is stabilized by taking measures against the change in the detection light intensity due to the directivity of the scattered light or the diffracted light from the foreign matter. Can be kept constant.

According to the present invention, it is possible to illuminate all points of the object from the same direction and detect all points from the same direction, so that even if the diffracted light or scattered light from the pattern has directivity, This is because if the shapes are the same, the intensity of the detection light will be the same.

The reason why the magnification is changed is to change the pixel size. If you increase the pixel size,
Since the area detected as one signal becomes large, the inspection speed can be increased as a result, but the resolution of the detection system decreases, so that it becomes difficult to detect small foreign matters or defects.
On the contrary, if the pixel size is reduced, the resolution will be higher and it will be possible to inspect smaller defects or foreign particles.
The inspection time is long. Of course, in this case, the resolution of the optical system also needs to be increased.

The lens replacement mechanism will be described with reference to FIG. In the present embodiment, as described above, one-to-one
A telecentric optical system with an image forming magnification of is used. In order to obtain the effect of the present invention sufficiently, telecentricity is important, and it is not necessary that the magnification is 1: 1 as shown in FIG. Therefore, it is possible to use an optical system having another magnification, and in realizing the optical system having another magnification, one of the two Fourier transform lenses 108 and 111 with a spatial filter interposed as shown in FIG. 11 (b). ,
The magnification can be changed by exchanging (specifically, the lens on the object side is optimal) with the Fourier transform lens 161. With such a configuration, it is not necessary to replace the image side lens and the detector, and as a result, it is possible to inexpensively supply an optical system having a different magnification.

As described above, the telecentric optical system,
Alternatively, an optical system in which the chief ray emitted from each point of the object passes through the center of the pupil (the surface where the spatial filter is arranged) of the detection optical system can be expected to have a great effect when used in a foreign matter defect inspection device using a spatial filter. However, not only when the spatial filter is used, but also when the defect foreign matter inspection not using the spatial filter is applied, it is possible to stably detect the high detection intensity. This is particularly effective when using an optical system with a large field of view.

(Basic Concept of PRES Filter) As described above, the PRES filter according to the present invention can exert the maximum effect when used in combination with the telecentric type detection lens and the illumination system coherent only in one axis. When the foreign object defect inspection apparatus using the spatial filter which is the original purpose is realized, it is not always necessary to use them together. The reason why the illumination is made coherent only on one side is that coherence is necessary when using the spatial filter and one side is sufficient.

Furthermore, since one side is not coherent, the illumination light flux can be narrowed down on the object, and the illumination intensity can be increased. Conversely, if sufficient illumination light intensity can be obtained, the illumination may be coherent on both sides in the x direction and the y direction. That is, the essence of the present invention is that even if a spatial filter is compressed into one dimension and the rotational direction of the wafer is adjusted, the spatial filtering can be performed.

What is more important here is that in the case of oblique illumination, in order to have one spatial filter parameter, the linear spatial filter should be parallel to the incident surface of the illumination, and only one side should be coherent. It's not about using lighting. That is, it is essential to match the incident direction of the illumination, the longitudinal direction of the linear spatial filter, and the repeating direction of the pattern on the wafer. Also, if there is no need to use one parameter, it is not necessary to make the linear spatial filter parallel to the incident surface of the illumination, and it is possible to use the linear spatial filter and adjust the pitch and phase or the pitch and rotational direction. A spatial filter that can handle all patterns can be configured. Further, for illumination from above, the direction of incidence of the illumination and the direction of the linear spatial filter are always the same, so there is the effect that only the repeating direction of the spatial filter and the pattern on the wafer need to be aligned.

However, in any of the above cases, if a linear beam is used for illumination or a one-dimensional sensor is used as the detector, this direction also needs to be adjusted. However, the matching in this case is to obtain the uniformity of illumination, or to eliminate the repetition of a large period by using the inter-chip repetition, and to erase the pattern information of a smaller period in the spatial filter. Above is not necessary.

As the telecentric detection optical system of the optical system, both telecentric optical systems are shown here, but it is not always necessary to be both telecentric, and at least the object side may be telecentric. Even if it is not telecentric, the chief ray of the illumination system at each point on the object,
That is, it suffices that the 0th-order diffracted light from each point on the object passes through the center of the pupil plane (the surface on which the spatial filter is installed) of the detection optical system.

Even with such a configuration, the light diffracted in the same direction with respect to the 0th-order diffracted light from the pattern of each point can be detected by the lens, so that the fluctuation of the pattern output due to the distribution of the diffracted light from the pattern can be avoided. . However, this method has a somewhat lower performance than the above-mentioned telecentric optical system because the incident directions of illumination to the pattern are different. However, this method may be sufficient for some subjects.

Further, the principal ray of the illumination system at each point on the object, that is, the 0th-order diffracted light from each point on the object passes through the center of the pupil plane (the surface on which the spatial filter is installed) of the detection optical system. Even if it does not do so, that is, even if a normal high-field lens is used, it is possible to realize the foreign matter defect inspection apparatus using the spatial filter, which is the original purpose of the present invention.

(Method of Measuring θ in Advance) In the above inspection apparatus, it is necessary to adjust the inspection apparatus to the angle of the wafer or the substrate. Specifically, it is necessary to set the optical axes of the detector and the illumination so as to be perpendicular or parallel to the repeating direction of the pattern formed on the substrate. In order to achieve this, the angle when the substrate is transported is detected with high accuracy by the angle detection mechanism, and the detection optical system is rotated based on the normal line of the surface of the substrate as a result, and the direction of the pattern is detected. Match the directions of the vessels.

A concrete structure is shown in FIG. FIG.
(a) is a diagram showing the arrangement of the spatial filter unit 108 and the rotation direction detectors 123 and 124 formed on the Fourier transform plane, as viewed from the substrate side of the detection optical system. The linear spatial filter 141 and the pupil size are shown in FIG. Aperture to limit the size
42 is also shown at the same time. The detection optical system is made to have a slightly large numerical aperture, and the diffraction pattern from the substrate whose parameters have been compressed protrudes to the outside of the diaphragm, and the detectors 123, 1
Detected at 24. Therefore, if the detector detects the 0th-order diffracted light in the diffraction pattern and detects the fluctuation of the peak position, the rotation direction of the detection optical system head 101 with respect to the substrate is measured.

Specifically, the distance between the two detectors is Lp,
The distances between the peak positions of the diffracted light detected from the center of the detector are hp1 and hp2. Since the diffracted light when the rotation position is displaced has a shape as shown in FIG. 12B on the Fourier transform surface, the rotation angle θp is generally expressed by the following expression 1. FIG. 12B is a view of a spherical surface including a Fourier transform surface viewed from the direction of the Fourier transform surface. Circle 3 shows the light rays on the spherical surface and the substrate surface, and circle 4 shows the light rays on the spherical surface and the pupil surface 142. Point 5 indicates the intersection of the 0th-order diffracted light of the illumination light, that is, the reflected light and the spherical surface.

[0096]

[Equation 1]

Strictly speaking, since the diffracted light pitch Ldp is an unknown number, the distance between the peak positions of the diffracted light is measured again at a position rotated by a known minute rotation angle θk, and hp11, hp21 As a simultaneous equation, Lpd and θp can be calculated. Further, as another method, there is also a method in which θ is rotated so that both hp1 and hp2 become 0.

Here, when the direction of the detection optical system is detected and rotated in the cheap direction, the purpose is to erase the pattern formed on the substrate by the spatial filter.
It is not always necessary to rotate the entire optical system, and the spatial filter may be rotated. When rotating the optical system, even if several units are rotated simultaneously,
There is no problem in rotating each unit.

What is important in this configuration is that the detection optical system is movable for rotation alignment with the substrate without rotating the stage supporting the wafer or the substrate. Here, the foreign matter detection device according to the present invention can be realized because it has a structure capable of detection even if it is not completely perpendicular to the flow direction of the wafer. Further, the rotation of the substrate corresponds to the rotation of the optical system, and the operation of the optical system with respect to the substrate is performed by using the substrate transport system, so that the two degrees of freedom are independently given to the two mechanisms. It is important that each mechanism is simplified.

Further, the angle detection mechanism may calculate the direction of the pattern formed on the substrate from the image captured by the detector 107, without using the method shown in FIG. In this case, it is difficult to measure the real time, but the detector 1
There is an effect that mechanisms such as 23 and 124 are unnecessary.

Furthermore, as shown in FIG. 13, instead of being mounted in the detection head 101, the configuration may be such that the measurement is performed in advance on the substrate being transported.
In addition, the detection result is rotated by the rotation mechanism 181 to rotate the detection head 101 or the detection head array 180 to match the rotation deviation. In such a configuration, since the measurement is completed in advance, there is an effect that the entire substrate can be inspected. Further, the rotation detection head 162 at this time is
As described above, the diffracted light may be detected, the image of the substrate may be formed and processed, or the edge of the substrate such as the orientation flat of the wafer may be detected by another sensor.

In the above, the angle detection and the substrate transfer are shown in the case of linear transfer in the embodiment of the present invention, but the present invention is not limited to this, and the rotation mechanism,
It is also applicable to a rotating arm or the like. However, in the case of such a rear transfer arm, it is desirable that the longitudinal direction of the detection area of the detection optical system is set so as to pass through the central axis of the rotary arm. With such a configuration, the present invention can be applied to a rotary type transport mechanism without any awareness.

(Rotation Control of Optical System) In the present invention, since the spatial filter is used, it is necessary to align the substrate and the optical system in the rotational direction. It is desirable that this rotation adjustment be performed in advance with respect to the setting of other parameters. The semi-invention has a wafer rotation detection optical system as shown in FIG. 3 and FIG. 12 or FIG. This optical system detects the rotation deviation of the wafer at the first portion of the substrate being conveyed, sends the detection result to the optical system rotation mechanism, and rotates the optical system based on this information. The rotating mechanism is shown in FIG.

Further, the rotation deviation can be corrected by electrical processing instead of rotating the optical system based on the result of the rotation deviation detection. By shifting the rectangular operator shown in FIG. 4 in the θ direction according to the detected rotation deviation of the wafer, it is possible to obtain an effect as if the rotation deviation of the wafer was mechanically corrected. This method has an effect that the time required for correction can be shortened because it is not necessary to move the optical system. Also, by using this circuit,
The operator 219, 231 is always moved in the θ direction without measuring the rotation deviation θ, and the inspection is continued under the condition that the detected foreign matter is the least (this condition corresponds to the condition where there is no rotation deviation between the substrate and the detection head). There is also a method. It goes without saying that this method has a high-speed signal processing system. Alternatively, even without using a high-speed signal processing system, the above conditions are set in advance by the method described above, and then the inspection is performed, so that the inspection is performed as if the rotational deviation of the wafer was mechanically corrected. You can

Of course, the method disclosed herein is not always necessary, and when the substrate is transported after mechanically adjusting the rotational deviation of the substrate within a certain allowable range with respect to the transport direction, the following method is used. No detection control system is required.

(Log Scale Threshold) FIG. 14 shows a comparison test in which an optical processing method such as a spatial filter is used as a pre-process, and a comparison test is performed only by an electric signal without using such a process. The state of the detection signal at the time is schematically shown. The spatial filter method can remove only the pattern information without losing the defect information in the pattern part, but the chip comparison method detects electrical information by superimposing the foreign particle and defect information and the pattern information. Since the signal is used as a signal, the particle defect signal can be detected only within the dynamic range at the time of photoelectric conversion. That is, when the pattern signal is extremely large and the foreign matter defect signal is extremely small, the foreign matter defect signal is missed in the pattern signal, and it is difficult to detect the foreign matter defect signal separately from the pattern signal.

In FIG. 14, the horizontal axis shows the detection position and the vertical axis shows the detection signal strength. The left side shows the signal 18 containing the foreign substance or defect information 4, and the right side shows the signal 19 not containing the defect information to be compared. Here, when the pixel size detected as one signal is detected as 13, the detection light corresponding to the areas 16 and 17 hatched is detected. In this case, since the information 4 is small with respect to the total area, the two detection signals 16 and 17 cannot be stably compared.
Specifically, it is buried in noise.

In this case, even if the light intensity of the illumination is increased, a detector having a large dynamic range is required to detect the information 4. Here, the pixel size is 13
From 14 to 14, the detection signals are 7 and 8, and the information 4 can be detected by comparison. Reducing the pixel size produces such an effect. On the contrary, if the detection signals 18 and 19 can be stably eliminated (by an essential comparison without changing to electric signals or the like), specifically, the detection signals can be detected at, for example, 10 positions or more. If possible, the detection signals become 5 and 6, which are at a level at which comparison inspection can be performed.

In this case, since the pixel size remains the same as that of 13, the high-speed detection by a large pixel becomes possible. If the light intensity of illumination is increased, the information 4 can be detected even by a detector having a small dynamic range. The method using the spatial filter of the present invention uses the pixel 13 as described above,
It is to detect the minute information 4.

As described above, even if the pattern signal in the portion between adjacent chips to be compared can be made extremely stable by devising the optical system or the like, the light intensity is 1: 100 or 1: 1.
The limit is detection with a dynamic range of about 000. Therefore, if the foreign matter signal is small or the pattern signal is large so that a dynamic range larger than this is required, the foreign matter signal is included in either signal by comparing the signal strengths between adjacent chips. It cannot be determined whether or not it is included. The presence / absence of foreign matter can be inspected by comparison only when the ratio of the foreign matter or defect signal to the pattern signal is sufficiently large. If this ratio is small, foreign objects will be overlooked or false alarms will increase if the threshold value is reduced to inspect foreign particles.

Therefore, it is difficult to inspect foreign matters existing on these patterns without false alarms, and there is no choice but to eliminate false alarms or reduce the foreign matter detection sensitivity to detect only large foreign matters. The purpose of eliminating the false alarm described above is to use the embodiment in which the region in which the pattern is formed is excluded from the inspection object in the present invention.

Further, in order to reduce the foreign matter detection sensitivity so that only large foreign matter can be detected, a log scale comparison inspection as described below is required. Certainly, even if adjacent chips are formed with patterns having the same shape and emitting the same amount of light in the same emission direction, the detection lights from these two patterns are not completely the same. Therefore, the detected light intensities of the light from these two patterns are likely to vary and comparison is difficult. So, when comparing
When comparing a and p, conventionally, it was determined that the two signals differ when the expression 2 is satisfied and that a foreign substance is present.

[0113]

[Equation 2]

In this method, however, when there is a large variation in the absolute level of the signal, if there is a variation that varies with the ratio to the absolute amount, the possibility of so-called false alarm in which it is determined that there is a foreign matter even though there is no foreign matter. Therefore, when the ratio of the two signals satisfies the expression 3, it is determined that the foreign matter.

[0115]

(Equation 3)

However, in practice, the division of two signals increases the scale of the arithmetic circuit. Therefore, in practice, the threshold value is set logarithmically, and when Expression 4 is satisfied, it is determined that a foreign substance exists.

[0117]

(Equation 4)

As described above, by using the equation (4), if the quantization threshold value is set using the logarithmic axis, the operation which originally needs to be divided can be represented by subtraction. The circuit configuration shown in FIG. 4 realizes the logic described above. The logarithmic processing described above is shown in FIG.
The threshold value of the octalization processing system 215 shown in (4) may be set logarithmically.

Further, FIG. 4 shows the subtraction processing using the above-described logarithm octalization processing, but this method is not always necessary. It does not matter even if multi-valued other than octal-ized is used. In this case, if the ternarization is used, all the foreign matters on the pattern are discarded without being detected, and if the large multi-value quantization is used and the optical system is stable, detection of smaller foreign matters on the pattern is performed. Will be possible.

Here, the above-mentioned δ is set in the threshold value setting circuit 221, and the above-mentioned equations 2, 3, and 4 are compared in the foreign matter comparison circuit 220. Further, the threshold setting circuit 222 basically sets the minimum threshold value of the foreign matter to be detected, which corresponds to the case where a pattern larger than the foreign matter threshold value is detected in the operators 219 and 231. It is stored as a large foreign substance. More importantly, in the present invention, at least one threshold value smaller than the threshold value used for so-called foreign matter detection is set. It is needless to say that this is necessary when using the above equation 4.

FIG. 30 shows the state of quantization by the above logarithmic threshold. The horizontal axis represents the detection position and the vertical axis represents the detection signal. Logarithmic threshold values 50, 51, 52, 53, 5
4 is set, and the signals in the part separated by the pitch p are compared. In this case, not only the pattern signals 55 and 58 which are quantized to the same value, but 1
Pattern signal 5659 quantized to different values,
And 57 and 60 are determined to be patterns and are not false information.

That is, by increasing the ratio of the logarithmic threshold value at the time of quantization, the allowable range is increased and the false alarm can be reduced, but only larger foreign matter can be detected on the pattern. Further, both the foreign matter signals 61 and 62 can be detected. Further, by widening the operator 231 in the plane direction, the quantization error in the plane direction can be allowed.

As described above, there is an essential difference between the pattern erasing by the spatial filter and the pattern erasing by the chip comparison. In other words, the spatial filter method can detect defects in the pattern area by emphasizing it, but the chip comparison method requires a large dynamic range because the defect information in the pattern area is photoelectrically converted as it is and detected before comparison. That is the point. The method using the spatial filter is just like emphasizing only defects by cell comparison using interference.

(Quantization Error in Plane Direction and Quantization Error in Depth Direction) Here, the method using repetition between chips used is basically a comparison inspection, but a short wavelength, point light source is used. The following configuration is used in order to stably realize such a comparison inspection by detecting scattered light using the laser light source.
An operator that realizes a pattern removal method using repetition between chips is formed by a plurality of pixels in the x direction and the y direction in the plane direction. For comparison of whether the signals should be judged to be at the same level or should be judged to have different signal levels due to the presence of a defect or a foreign substance on one side, use Equations 2, 3 and 4. ing. The foreign matter and the pattern can be stably distinguished from each other by the sampling enlargement processing in the plane direction and the light intensity direction of the comparative numerical values in these comparisons.

(Filter Size) The design concept of the spatial filter will be described below. In the foreign matter and defect inspection apparatus of the present invention, the pattern information having a large pitch is erased by repeating the chips by the operator. Therefore, the "spatial frequency that can be eliminated by the operator" must be larger than the "spatial frequency that can be eliminated by the operator." Hereinafter, the reason will be described.

Consider a case where patterns of pattern pitches L1 and L2 (L1 <L2) are formed on a substrate as shown in FIG. The illumination light is θ
It is diffracted in the directions of 1 and θ2, and a diffraction pattern of pitches pf1 and pf2 (pf1> pf2) is created on the spatial filter. Therefore, the following formula is established.

[0127]

(Equation 5)

[0128]

(Equation 6)

Here, Df is the diameter on the pupil plane of the optical system,
NA is the numerical aperture of the optical system.

Among the diffraction patterns, the diffraction pattern of the pitch pf1 is shielded by the spatial filter, but the pitch p
Since the pitch of the diffraction pattern of f2 is too small, it cannot be shielded by the spatial filter. That is, the pattern on the substrate with the pattern pitch L1 is erased by the spatial filter and does not form an image on the detector, but the pattern with the pattern pitch L2 is not shielded by the spatial filter and is imaged on the detector and is not erased. (Hereinafter, "the pattern is erased" means that the light emitted from the pattern, that is, the light having the pattern information is blocked by the spatial filter so that the light does not reach the detector. Therefore, the following equation 7 is established from the equations 5 and 6.

[0131]

(Equation 7)

Here, the spatial filter is required to have a certain size or more due to limitations such as the numerical aperture of precision illumination in manufacturing. Therefore, if pf1 is kept larger than the expression 7, the pitch of the pattern that can be erased by the spatial filter becomes small.

Here, a plurality of chips are formed on the substrate at a pitch Lt. The information of the pattern having the pattern pitch L2 is erased by using this repetition between chips. Specifically, it is determined that the signal repeatedly detected at the chip pitch Lt is a signal from the pattern having the pattern pitch L2, and the detection signal itself is deleted. (in this case,
“Erasing pattern information” literally means “deleting or discarding a signal”. ) Here, in order to avoid a quantization error in the plane direction, which will be described later, the operator described above uses nx pixels in the x direction, y
When the pattern has a size of ny pixels in the direction and a pattern exists in this operator, the determination pixel is determined to be a pattern. Therefore, depending on the size of this operator, the rate α of the detection area is expressed by the following expression 8.

[0134]

(Equation 8)

On the contrary, from the equation 8,

[0136]

[Equation 9]

Therefore, if the detection area ratio is kept large, the pitch of the erasable pattern becomes large.

Here, in order to erase the patterns of all the pitches on the substrate, the pattern pitch L1 that can be erased by the above spatial filter should be larger than the pattern pitch L2 that can be erased by the operator.

Here, the pitch variable PRES filter is formed so that the pitch can be continuously changed from the minimum pitch pf1 to the maximum pitch 2 · pf1.

Further, in order to prevent or suppress the interference phenomenon on the detector due to the spatial filter, it is necessary to limit the number of filters, pitch and filter width. That is, the above-mentioned minimum spatial filter pitch is limited not only in production but also in order to suppress the interference phenomenon. At this time, the above relational expression among the number of filters, the field size, and the operator size becomes important.

Whether or not the pattern can be erased by the spatial filter does not depend on the pitch of the pattern to be erased, but on the number of linear spatial filters, as described below. The size of the spatial filter surface is D, the minimum spatial filter pitch is pfs, the number of linear spatial filters is nf, and the maximum pattern pitch that can be shielded by the spatial filter of pfs is Lm. It is assumed that the field of view of the detector is X, the pixel size is w, and the number of pixels is N. Let the operator size be no when the operator compares two chips.

[0142]

(Equation 10)

In order to make the detection area ratio sufficiently large,

[0144]

[Equation 11]

[0145]

(Equation 12)

However, it is necessary. If the pixel size is set near the resolution of the optical system, it is necessary and sufficient from the inspection time and the detection performance.

[0147]

(Equation 13)

From the above equations 10, 11, and 13,

[0149]

[Equation 14]

[0150]

(Equation 15)

Therefore, the detection area ratio is determined only by the operator size and the number of linear filters. If k0 is about 1 and the operator size no is 3, it is sufficient to have 15 or more linear spatial filters to set the detection area ratio to 80% or more.

The above examination results show that regardless of the pitch of the pattern to be detected, if the pixel size is made equal to the detector size and approximately 15 or more linear spatial filters are used, patterns of all pitches are obtained. Means that it can handle

(Preparation Method) A method for preparing the above spatial filter will be described below. The spatial filter of the present invention only needs to have a shape in which a plurality of linear filters are installed on the spring-like support, and it does not necessarily depend on the method described here, but the method described here is the spatial filter of the present invention. Can be manufactured efficiently and inexpensively.

FIG. 19 shows a creating method. According to the present invention, a filter sheet 165 formed by etching linear filters at equal intervals is placed on the coil springs 121, 121, and each linear filter 141 is soldered or set by an adhesive or the like. Here, the coil springs 121 and 122 and the linear filter 141 are preferably made of stainless steel from the viewpoint of strength, but in the case of stainless steel, there is a drawback that solder is hard to deposit. Therefore, it is advisable to plate the surface of stainless steel with chromium, and then to plate gold with gold to prevent oxidation of chromium. However, these platings are not necessary and indispensable, but merely to facilitate soldering. Therefore, it is also possible to use a flux that acts directly on the surface of the stainless steel to facilitate solder loading. In such cases, chrome and gold plating is not needed.

The linear filter in the filter sheet is preferably set to a size with an allowance of 10% to 100% of the image size at the Fourier transform position of the light source of the optical system described above. . Here, it is desirable that the position where the filter sheet 165 comes into contact with the coil spring is as thick as the coil spring. Therefore, it is necessary to change the thickness of the filter portion and the contact portion of the linear filter. In this portion where the thickness changes, the thickness needs to change smoothly in order to avoid stress concentration.

Coil spring 1 in which the filter sheet 165 is installed in advance at equal intervals with the filter sheet 165.
21, 122 and 122, and an appropriate amount of solder paste 166, which is a mixture of solder and flux, is applied to the contacts. Here, the solder paste may be placed on the filter sheet in advance.

Here, in order to prevent the position of the central linear filter from changing even when the interval of the linear filters 141 is changed, that is, in order to always shield the 0th-order diffracted light, When the sheet 165 is placed, the center of the coil spring movable mechanism 108 needs to come to the center of the filter sheet. Therefore, when the movable mechanism is operated in advance and the contact portion of the coil spring with the filter sheet does not come to the center of the movable mechanism, the coil spring is moved around the central axis 23 of the coil spring itself as shown in FIG. It is necessary to rotate so that the center 25 of the coil spring moving mechanism comes to the center 25 of the filter sheet.

The filter sheet 165 on which the solder paste is placed is placed on the above coil spring movable mechanism 108, and the whole is heated to melt the solder paste and then naturally cooled to complete. The reason why the entire coil spring moving mechanism is heated here is to prevent residual stress due to heating. Since the spatial filter mechanism of the present invention uses a coil spring having a small elastic coefficient, the strain becomes large with respect to the residual stress, and due to the position shift of the filter, the light shielding performance, that is, the performance of the spatial filter ring described above is inferior. I will end up. To remove the strain due to such residual stress, it is preferable to heat the entire movable mechanism.

Further, since the present spatial filter blocks diffracted light, it is desirable that the spatial filter be treated in black in order to eliminate irregular reflection and the like when the light is blocked. This black treatment may be applied with a paint or may be accompanied by heat treatment called black dyeing treatment. In addition, this black treatment, even after being soldered, causes the filter sheet 16
5 may be formed on the filter sheet 165 only for the portion other than the joint portion on 5 and the light shielding portion.

Although the method using the solder paste 166 has been described above, it is not always necessary to use the solder paste, and the linear filter in the filter sheet may be soldered one by one. It goes without saying that it is okay.

(Image restoration by convolution) FIG. 2
As shown in FIG. 1, in the device according to the present invention, the light rays are diffracted by the spatial filters 141 arranged at equal pitches to form a diffraction pattern on the image plane. Specifically, a diffraction image of a point image is formed on the image plane. In this image, the ± 1st order diffracted light appears around the 0th order diffracted light. When such diffracted light appears, it becomes a signal 26 in which ± 1st-order peaks appear around the original peaks, and not only is foreign matter increased to three and detected, but also in the case of a pattern. Will be judged as a pattern, and there will be a large number of places where erasure or detection sensitivity will be degraded. To avoid this, the width of the linear spatial filter may be narrowed.

Specifically, when the width of the linear spatial filter is 1/2 of the pitch of the spatial filter, the above 1
The diffracted light of the 0th order is half the diffracted light of the 0th order, but the width of the linear spatial filter is 1 with respect to the pitch of the spatial filter.
In the case of / 8, the 1st-order diffracted light is reduced to 1/30 times the 0th-order diffracted light. These results were calculated by Fourier transforming the shape of the spatial filter.
Therefore, it is necessary to select the width of the linear spatial filter with respect to the pitch of the spatial filter, if necessary. Further, especially when it is necessary to know the ratio to be small, it is necessary to change the way of focusing at the position of the Fourier transform of the illumination system so that the diffraction pattern is sufficiently shielded by the spatial filter.

Specifically, it is achieved by increasing the numerical aperture of the component in the y direction of the incident light on the substrate of the illumination. The size of the image formed on the Fourier transform plane of the illumination at this time is calculated by the image formation theory of coherent light.

The influence of the above diffraction can be avoided by an image processing method known as a Wiener filter. Specifically, as shown in FIG. 21, assuming that the shape of the spatial filter is L (u, v), the value of 1 / L (u, v) is calculated in advance and its Fourier transform is calculated. May be convolved with the image detected by the convolution means 251. However, here, since the value of 1 / L (u, v) has a part that diverges to infinity, it is necessary to approximate this value to a necessary and sufficient large sum.

Further, the value of the complex number obtained as a result of the Fourier transform is approximated so that the magnitude is an absolute value of the complex number, with the numerical value of the phase inversion being positive or negative. Also, it should be set so as to have a minimum size that is sufficiently effective when cutting out an image to be convolved.

In the pattern removal by the spatial filter and the repeated chips (rotating the wafer in accordance with the fine pattern), it is better that the pitch of the pattern removed by the spatial filter is smaller, and the pattern removed by the method by the repeated chips. The better the pitch, the better. In particular, in order to reduce the area that is removed as a pattern, it is desirable that the pitch of the pattern removed by the repeated chip method is large.

Therefore, in the inspection, it is preferable to determine the rotation direction of the substrate at the time of carrying the substrate according to the shape of the pattern formed on the substrate so that the area removed as a pattern is reduced. Specifically, as shown in FIG. 22, the pattern pitch L that cannot be erased by the spatial filter.
When there are 11 and L12, it is better to scan in the direction 28 so that the direction of the larger pattern pitch L11 becomes the direction of the sensor. The substrate 1 is rotated 90 degrees in advance so as to be in such a direction, and the inspection is performed.

(Illumination from Directly Above) Here, for example, in order to realize miniaturization and securing of low resistance, the current or future semiconductor devices have a higher aspect ratio.
Therefore, the foreign matter or defect hidden behind this step is
Not detectable. Since the present invention has a deep depth of focus even for such an object, it can be inspected. Further, as shown in FIG. 23, by using a mirror 168 and a half mirror 169 and using an illumination optical system that illuminates from above, a foreign matter or a defect hidden behind can be illuminated and can be detected.

(Application to TFT) A large-sized substrate such as a TFT substrate and having a small thickness has a large strain when supported.
In such a case, as shown in FIG. 24, the substrate 30 is supported at both ends by supporting portions 171 and 172, and the longitudinal direction 31 of the illumination spot and the longitudinal direction of the detector are set in the direction 32 connecting the supporting ends. The substrate is set to be vertical and the substrate is operated parallel to the direction 32. With such a configuration, even if only the peripheral part can be supported and the strain becomes large, the direction of the strain can be made perpendicular to the direction of the illumination and the detector, so there is no strain in the direction of the illumination and the detector. , It is possible to put the inspection object within the depth of focus within the detection visual field.

Further, even when the strain of the entire substrate is large, there is an effect that the strain shape of the substrate can be easily calculated. Therefore, based on parameters such as the thickness, length, width, longitudinal elastic modulus, and lateral elastic modulus of the substrate, calculate the strain of the double-ended free-end support, and adjust it according to this shape so that the focus is adjusted. It is possible to inspect while driving the stage in the focal direction. Such a configuration has an advantage that an automatic focus detection mechanism is unnecessary. With the above configuration, it is possible to inspect foreign matters and defects on a large substrate such as a TFT.

When a plurality of detection head arrays are used as shown in 33 and 34, since the strains at the respective positions are different, the substrate 30 is not adjusted in the z direction, but the detection head arrays 33 and 34 are independent. Need to be adjusted.

Needless to say, such a configuration is not always necessary when the substrate is fixed to a flat support table and inspected in such an inspection of the substrate. In addition, the strain can be reduced by supporting the four corners of the substrate.

(High-precision foreign matter inspection device) As described above, the inspection device using the spatial filter is not only for achieving high speed and small size. If the resolution is improved by exchanging the Fourier transform lens on the object side as shown in FIG.
Specifically, if the resolution is about 1 micron, the minimum is 0.
It can inspect foreign particles or defects of 1 to 0.3 microns at high speed. Further, if the resolution is set to about 3 microns, it is possible to inspect foreign matters or defects with a minimum size of about 0.3 to 0.8 microns at high speed.

With such a configuration, since the pattern signal that becomes noise can be erased well by utilizing the pattern repeatability, a larger pixel than the pattern inspection device for design data comparison, cell comparison, or chip comparison is used. Since small foreign matters or defects can be inspected even if the size is used, high-speed inspection can be realized as a result.

Even in such an inspection, since it is basically an inspection of only repetitive patterns, it has no repeatability.
The pattern part is not subject to inspection. It is necessary to devise a portion other than the inspection target with another inspection device, perform a visual inspection, or the like.

(Use of Detection Head) In the foreign matter defect detection apparatus described above, it is most effective to inspect the substrate during conveyance in one direction, as shown in FIG. For this reason, it is desirable to arrange a plurality of inspection devices (detection heads) in parallel as shown in FIG. However, as shown in FIG. 26, even if only a part 35 of the substrate is inspected using one unit of the detection head 101, the effect is often sufficiently exerted. This is because the situation can be detected more than the occurrence of foreign matter defects in some foreign matter defect inspections. Of course, this need not be one unit, and multiple units can be placed if desired.

Further, it is also possible to use one unit or more detection heads and use a structure for inspecting the entire substrate by xy scanning of the stage as shown in FIGS. 27 (a) and 27 (b).

(Dynamic Inspection Matched to Pattern Shape) FIG. 28 shows an inspection apparatus in the case where the pitch of repeated patterns changes in the substrate. In this embodiment, the repetitive pattern pitch detecting section 174 and the detecting head 101 are provided.
Consists of

When the substrate is transported, first, the pitch of the repeating pattern is calculated by the parameter calculating means 212 using the frequency analysis method described above based on the signal detected by the repeating pattern pitch detecting section 174. . The pitch calculated here is sent to the detection head 101,
The pitch of the spatial filter and the pitch of the operator are changed. The detection head 101 is inspected at the time point when the portion where the pitch detection unit 174 pitch is calculated is conveyed to the inspection position of the detection head 101. The inspection proceeds while the pitch is constantly calculated. Here, when the pattern pitch does not change discontinuously, the inspection can be performed by this method.
When the pitch changes discontinuously, there is a possibility that the setting of the pitch may be too late. In such a case, it is necessary to stop the feeding of the transport system until the setting of the pitch is completed.

(Diffraction Interference Method Using Sine Curve Diffraction Grating) A pattern inspection method combining diffracted light and interference light will be described below with reference to FIG.

This device is basically the same as the inspection device using the spatial filter described above.
02, detection optical system 103 spatial filter unit 10
6, a detection mark 107, and a diffraction grating 175 having a sine curve phase distribution and formed by a transparent substrate at the position of the spatial filter, and a drive mechanism 176 are added. Here, in the method described above, the pattern removal process using repeated chips is realized by the electrical processing system shown in FIG.

As described above, this method has a problem that the foreign matter in the pattern cannot be inspected unless it is extremely large. Further, as shown in FIG. 14, as described above, information in such a large background noise can be detected only by a method capable of uniformly and reliably cutting the offset such as optical processing. Therefore,
It is intended to realize the pattern removal processing using the repetition of the above chips by an optical method, specifically, an interference method.

Here, the illumination shown in FIG. 29 is shown for the substrate as transmitted light, but it does not matter whether it is transmitted light or reflected light. Here, the illumination is performed coherently in at least one axis direction. here,
When there are repeated patterns 37 and 38, the principal ray (0th-order diffracted light) of the light emitted from each proceeds along each optical axis and forms each image on the detection mark 107. However, since the diffraction grating 175 is installed on the Fourier transform surface, the light is diffracted and the diffracted light is emitted in the ± first order directions.

Here, what is important is that since the sine curve (the same as the cosine curve) is used as the diffraction grating 175, the 0th-order diffracted light disappears (the light intensity is 0), and ±
The point is that only the primary light is emitted. Here, the + 1st order light from the pattern 37 and the −1st order light from the pattern 38 can be overlapped on the detector by adjusting the diffraction grating 175 in the optical axis direction by the driving mechanism. Furthermore, if a phase plate 178 that shifts the phase by π is placed on the illumination side or at an appropriate position on the optical axis, the two light beams can be made to interfere on the detection mark 107. As a result, if the phase change of the phase plate 178 can be finely adjusted so that the phase of the light flux in the middle can be corrected, the pattern removal process using the repetition of the chips due to the interference can be realized.

Further, the diffraction grating 175 and the driving mechanism 176.
If the inter-grating distance of the diffraction grating is appropriately changed by changing the wavelength of the ultrasonic wave by using a refractive index variable mechanism by the surface wave of the ultrasonic wave such as SAW, each + 1st order light emitted from the diffraction grating 175. -1st order light is emitted in the same direction,
Can be stacked on the detector. This method has the effect of automatically creating a sine curve on the diffraction grating 175. Further, it is not necessary to move the diffraction grating 175 in the optical axis direction. Of course, these may be used together.

In the above, the sine curve is used as the diffraction grating. However, the diffraction grating is not limited to this, and when comparing patterns 37 and 38 which are separated by a sufficiently large interval, it is not a sine curve and the It is not necessary to have a sine curve because the influence of folding light can be suppressed.

Further, FIG. 29B shows a configuration in which, when the pattern is illuminated, only spots separated by the pitch of the pattern are illuminated with spots. In this configuration, the light source 179 has a scanning device and the half mirror 18
0, via the mirror 181, only spots separated by the pitch of the pattern are illuminated with spots. With such a configuration, an accurate phase shift π can be created and the detection performance can be improved.

(Others) As described above, the above-described inspection apparatus is for inspecting foreign matters or defects on the substrate being transported, so that the substrate does not enter the depth of focus of the detection optical system during transportation. Need to be Therefore, it is better that the depth of focus is deeper so that inspection can be performed even in a transport system that is not so high in accuracy. Therefore, in order to give priority to the depth of focus over the detection resolution, a diaphragm may be installed in the detection optical system to increase the depth of focus.

In the inspection apparatus described above, a linear filter is used, but this filter does not necessarily have to be such a filter, and a filter using a liquid crystal display element or silver chloride may be used. The reversible optical cut filter used may be a complex conjugate type non-linear element.

In the above embodiments, the illumination light is light from information, that is, reflected light. However, when the effect of the present invention is obtained, the present invention is not limited to this, and transmitted light is used. It is a structure with the lighting of, and it does not matter at all.

Further, the foreign matter defect information detected by the inspection apparatus of the present invention not only detects abnormalities by counting the number of foreign matter defects, but also knows the cause of occurrence of foreign matter defects by knowing the generation distribution of foreign matter defects. It can also be a clue to analogy. Further, the arrangement, the number of arrangements, the sensitivity, etc. of the monitor of the present invention may be set based on the result of the highly accurate inspection device.

Although the illumination system of the present invention is constructed independently of the detection system as the illumination system 102, this illumination system can be omitted by using a part of the detection system. Specifically, it can be realized by mounting a semiconductor laser and a cylindrical lens having an appropriate focal length on the Fourier transform surface of the detection optical system. By doing so, the detection head can be made smaller, lighter, and cheaper.

Although the semiconductor laser is used in the above construction, the present invention is not limited to this, and the effect of the present invention can be obtained even with a gas laser, a solid-state laser, white light of a substantially short wavelength, and a point light source.

The method of using the foreign matter detection system will be described below with reference to FIG.

In the manufacture of semiconductors, this series of steps is repeated with the basic steps of coating pattern exposure, development, etching, cleaning, and film formation to form a multilayer structure on the silicon substrate.

Here, in each of the steps, there is a problem that the wafer becomes defective due to the adhesion of foreign matter. Therefore, the state of foreign matter adhesion is monitored during each process, and if any foreign matter is found, countermeasures are taken.

However, in reality, since the foreign matter inspection apparatus is expensive and takes up space, it is practically impossible to inspect all the wafers in all these steps. Therefore, the foreign matter monitor described above is effective. In other words, 100% inspection was realized by simply making a foreign matter inspection device. Here, in order to easily make the foreign matter inspection device, the inspectable area is limited to the place on the product wafer where the pattern is not formed and the place where the repeated pattern is formed. This inspection method focuses on the fact that the foreign object inspection for the purpose of managing foreign objects in a process or an apparatus can achieve the object even if the inspection area is limited in this way.

The concept of this detection method will be specifically described.
In the above-mentioned foreign substance monitor, a portion where a repeating pattern exists is inspected by using a spatial filter, and a non-repeating portion has a practically low sensitivity, or it cannot be inspected by an inhibit process (non-inspection process). Therefore, the ratio of the portion that is actually inspected to the entire area is defined as the inspection area ratio α. This detection area ratio changes depending on the shape of the pattern formed on the wafer.

In foreign matter inspection in which a single-wavelength laser is irradiated from one direction (coherent illumination), the light diffracted or scattered from the foreign matter has directivity. This directivity is determined by the size, shape and refractive index of the foreign matter. In reality, since foreign matters of various shapes and sizes are generated, the detection sensitivity (how many microns of foreign matter is inspected) is often not uniquely determined, but statistically each size is different. It is possible to define the detection rate β, which is the ratio of foreign matter of the foreign matter. FIG. 32 shows a conceptual diagram of the detection rate β. The detection rate is also affected by fluctuations in the detection signal. Therefore, if the ratio of the detectable foreign matter among the foreign matter deposited on the wafer is defined as the supplement rate γ, the following equation is established.

[0200]

(Equation 16)

Here, n is the number of detected foreign particles, and N is the number of foreign particles attached on the wafer.

Therefore, the number N of foreign matter attached on the wafer is

[0203]

[Equation 17]

It is indicated by.

Therefore, the detection area ratio α and the detection ratio β can be measured in advance for each process and each product, and the number of adhered foreign matters can be controlled by the above formula.

However, since this control method is a statistical method, the accuracy may decrease when the number of foreign matters is small or the number of samples is small. Therefore, if the relative change in the number of foreign particles in a specific process is followed, the number of detected foreign particles n
In some cases, the method using is good.

Here, in order to make the detection performance uniform among the processes, the difference in detection sensitivity may be corrected by the foreign matter capturing rate as described above, but the following method may also be used. FIG.
The number of foreign matters on the same wafer is plotted before and after a certain process as shown in FIG. At this time, the number of foreign substances when the process is processed without problems is plotted in advance. The points plotted at this time are examples of good cases. From this first-order correlation in the favorable case, the threshold value for the defective case is determined. The slope of the primary correlation at this time means a difference in detection sensitivity before and after the process. Further, the defect means that foreign matter was attached in the process. Therefore, the number of foreign matters before and after this step is plotted on this FIG. 37, and the case where it is out of the threshold value may be regarded as a defect. Here, in setting this threshold value, the above-mentioned supplement rate may be used.

[0208] Here, at the start-up stage of the manufacturing line,
In a foreign matter inspection device or pattern inspection device with high detection sensitivity, the state of foreign matter adhesion and the result of investigating the cause of it are stored as a file in correlation with the inspection result of the above foreign matter monitor. Depending on the result of this correlation, change to a highly sensitive foreign matter inspection device,
Move to the foreign object monitor. As a result, it is possible to perform 100% inspection with a foreign substance monitor, and it is possible to deal with dust that occurs at low frequency. In particular, at the stage when mass production has started, preventing dust generation that occurs infrequently is the key to improving yield. At the time of such replacement, the method using the above supplementary rate γ is effective. This is because the result of inspection with a highly sensitive pattern inspection device or the like is as close as possible to the number N of adhered foreign matters.

Further, in the inspection after development, it is possible to detect defects such as defocus during exposure.

In normal semiconductor production, two wafers are used.
Put in 5 cases (1 lot) and transport. At this time, it is often the case that some of the lots are inspected for inspection. In addition to automatic foreign matter appearance inspection equipment, this sampling inspection also includes characteristic inspections such as lifetime and resistivity, and visual inspections using a light projector (which inspects the surface for foreign matter, etc. by irradiating light from the direction of the light). Includes visual appearance inspection with a metallurgical microscope. In this sampling inspection, especially by visual inspection, there are many cases in which scratches or foreign substances are attached due to an unexpected operation failure and the yield is reduced. Therefore, the above-mentioned foreign matter monitor has a great effect in monitoring these operations.

In this visual observation monitoring, the number of foreign matters on the wafer may be inspected before and after the visual observation. As a result, not only the work level of the visual observer is monitored, but also the morale of the visual observer is improved.

In this case, when the number of adhering foreign matters is large, it is conceivable to perform processing such as discarding the wafer midway in consideration of the efficiency of subsequent processing. As a result, it is possible to prevent a substantial decrease in the operating rate of the apparatus due to the processing of wafers with a low yield in the subsequent processing.

Further, in the line for complete single-wafer processing, the effect of 100% inspection becomes even greater. This is because in batch processing, the process conditions and the like are considered to be relatively stable within each batch, but in single-wafer processing, each wafer is transferred into a vacuum and process processing is performed. This is because the conditions are likely to change from wafer to wafer, and inspection of each wafer becomes indispensable.

(Analysis of randomly generated foreign matter) A change point analysis system will be described as one of the methods for investigating the cause of foreign matter. Disturbances usually cause defects. This disturbance causes so-called foreign matter generation. On the other hand, the generation of foreign matter occurs randomly like thermal noise, so it is difficult to identify what triggered it.

Here, if it is random like complete thermal noise, it means that it is irrelevant to disturbance, and there is no related disturbance, that is, the cause (which cannot be determined in principle). Here, what we want to investigate is the disturbance associated with the generation of foreign matter. In this way, when a phenomenon is related to a disturbance, a random phenomenon also shows a probability distribution as the number of samples increases. This mathematical phenomenon can be used.

That is, in order to associate an event with a low occurrence frequency related to something with a cause, the number of samples can be increased by superimposing the occurrence of foreign matter on the basis of the cause, that is, each disturbance as a trigger, and the probability distribution May be visible.

For example, with the timing of vacuum system (or vacuum gauge) exchange, cleaning of the equipment (total cleaning), sputter target exchange, visual observation with a metallurgical microscope, etc. as the starting point, the number of sheets to be processed thereafter is plotted on the horizontal axis. Then, the number of foreign substances attached to each wafer is plotted. At this time, the change in the number of foreign matters
If there is a tendency to synchronize in each cycle, the disturbance is likely to be related to the generation of foreign matter. Here, as a method of reducing the noise characteristic of foreign matter generation, a method of integrating the number of foreign matter is also effective in addition to this superposition. Of course, these methods are effective alone, but if both are used at the same time, the effect becomes even greater.

In this changing point analysis method, the data for each lot may be arranged by using a disturbance as a trigger, but of course, the data for each wafer may be arranged. Furthermore, the data for each lot may be overlaid. Ie 1 in the cassette
The number of foreign matters in each wafer may be plotted by plotting the first sheet on the origin of the horizontal axis and plotting the arrangement order in the cassette on the horizontal axis.

An example will be described with reference to FIG. FIG.
(a) shows the number of foreign matters adhering to the wafer, arranged in the order of wafer processing. Here, the timing of the work related to this device is shown in the lower part of FIG. Here, Fig. (B) shows the changes in the foreign matter when the timing of each work is taken as an opportunity. In each figure, the correlation coefficient between the wafer processing order and the number of foreign particles is calculated. It can be seen that the correlation coefficient is large when work A (for example, replacement of the sputter target) is arranged. This suggests that work A is strongly related to the generation of foreign matter. Therefore, work A should be reviewed and the problem solved. (For example, a defective work method when replacing the sputter target) When the work is performed as a point at a certain point in this way, the transition of the number of foreign matters may be rearranged at that time.

An example in which the parameters related to FIG. 34 are not the work completed at a specific time but the process conditions or the like is shown below. For example, parameters such as the pH of the cleaning liquid, the number of sheets charged during batch processing, the indicated pressure of the vacuum gauge, and the indicated level of electric power applied to the electrodes during sputtering are used as parameters.
In this case, the levels of these parameters and the number of foreign substances are correlated. In this case, there is a strong correlation between the parameter A and the number of foreign particles, and it can be seen that this parameter needs to be controlled with high accuracy for foreign particle management.

Similarly, the correlation may be obtained by using the difference in process etc. as a parameter. For example, the chamber number when performing the same process in a multi-chamber apparatus may be used as a parameter.

In the system for realizing this change point analysis method, the collection of work data such as work contents and changes is the key. Therefore, when some processing, repair, parts replacement, etc. are performed for each device, it is necessary to record it. Also,
It is necessary to collect variable process conditions, etc., which are related to foreign matter (almost all because it is not always clear what is actually related). In reality, this is satisfied by the digitization of work notes. Alternatively, an input system that facilitates recording at the work site may be constructed. In order to facilitate the input work, it is desirable to code some work that is routinely performed by each device, input the code with a bar code reader, and file the time at that time. The work database created in this way is linked to the foreign matter database and the above-described data processing is performed to investigate the cause of the foreign matter.

In addition to the number of foreign matters, the foreign matter data can represent the distribution of foreign matters using the distribution of foreign matter coordinates (x direction, y direction, xy direction) and the average value of the foreign matter coordinates (center of gravity). . Thus, for example, for a foreign substance in the process chamber, it is possible to represent a foreign substance generation mode in which the position in the chamber where the foreign substance is generated. Similarly, the average value and the variance of the detected particle level can also represent the generated particle mode. Such a center of gravity value, an average value, and a variance value are statistically sufficiently meaningful and have the advantage of being simple to calculate.

The foreign matter attached to the wafer is often large among the foreign matter generated in the process. Here, larger foreign matter is more likely to move downward due to forces such as static electricity and plasma potential. Therefore, if the wafer being processed is turned upside down, that is, the wafer surface is placed downward, the foreign matter can be prevented from adhering to the wafer.

In addition, the results of these foreign matter inspections are immediately
It is also good to perform visual confirmation at the visual confirmation station. In addition, by disposing the above-mentioned foreign matter inspection device next to, for example, a metal microscope, an operator must inspect the foreign matter before and after work,
Since the worker can recognize the foreign matter adhered by his / her own work,
The worker's morale can be improved.

FIG. 35 shows a bar code for wafer recognition. Drawing at one or more locations around the wafer
Place barcodes like (b), (c), (d). This arrangement position may be on the front surface (pattern formation surface) of the wafer shown in FIG. 35 (a) or on the back surface of the wafer. FIG.
A wafer edge portion as shown in 6 may be used. When the wafer edge portion is used, the wafer surface can be effectively used.

The bar code may be a one-dimensional bar code as shown in FIGS. 35 (b) and 35 (c) or a two-dimensional pattern as shown in FIG. 35 (d). The method of creating the barcode may be a laser marker or lithography. In any case, it is necessary to roughen the surface in order to create the light and dark. However, since the roughened surface causes dust generation, it is preferable to use a bar code by changing the pitch of thin wires as shown in FIG. 35 (b). This fine line is also created by a laser marker or the like, but optically the fine line may act as a so-called edge. Therefore, it is not necessary for the thin line portion to be provided, and a clean surface may be a concave surface or a convex surface. The cross-sectional structure is shown in the figure. Here, when this bar code is formed on the back surface, a bar code detector for back surface observation is required.

[0228]

Since the present invention is configured as described above, a foreign matter inspection system with high precision is used by dividing the foreign matter inspection system at the time of starting mass production in the semiconductor manufacturing process from the mass production line. As a result, the functions of detecting, analyzing, and evaluating foreign matter required at the time of mass production startup can be maximized, so that feedback to the mass production line can be smoothly advanced, and the mass production startup period can be shortened. It is possible to realize high-frequency sampling close to 100% inspection by using the monitor, and to realize high-quality product and high-yield production.

[Brief description of drawings]

FIG. 1 is a configuration block diagram showing an embodiment of an apparatus according to the present invention.

FIG. 2 is a configuration block diagram showing an example of the detection head of FIG.

3 is a perspective view showing the spatial filter mechanism of FIG. 1. FIG.

4 is a configuration block diagram showing an operator processing unit in FIG. 1. FIG.

5 is a block diagram showing the configuration of the parameter detection system shown in FIG.

6 is a perspective view showing the rotation adjusting mechanism of FIG. 1. FIG.

FIG. 7 is a configuration block diagram showing a conventional method.

FIG. 8 is a block diagram showing the basic concept of the present invention.

FIG. 9 is a configuration block diagram showing a pattern removal method of the present invention.

FIG. 10 is a configuration block diagram showing another embodiment of the parameter detection system.

FIG. 11 is a configuration block diagram of a detection lens.

FIG. 12 is a configuration block diagram showing a rotation deviation detection system.

FIG. 13 is a configuration block diagram showing another embodiment of the rotation deviation detection system.

FIG. 14 is a schematic diagram showing an effect of a comparative inspection by optical filtering.

FIG. 15 is a perspective view showing a state when light is shielded by a spatial filter.

FIG. 16 is a schematic diagram illustrating operator processing.

FIG. 17 is a schematic diagram illustrating the shape of a spatial filter.

FIG. 18 is a schematic diagram illustrating conditions for pattern erasing.

FIG. 19 is a perspective view showing a method of manufacturing the spatial filter mechanism.

FIG. 20 is a perspective view illustrating a method of adjusting the coil spring.

FIG. 21 is a configuration block diagram of a method for removing the influence of diffraction.

FIG. 22 is a schematic diagram illustrating a scanning direction of a sensor.

FIG. 23 is a configuration block diagram of a configuration for realizing vertical illumination.

FIG. 24 is a perspective view showing a method of using the detection head.

FIG. 25 is a perspective view showing a method of using the detection head.

FIG. 26 is a perspective view showing a method of using the detection head.

FIG. 27 is a schematic diagram showing a scanning method of the stage.

FIG. 28 is a configuration block diagram showing a pattern pitch measuring means.

FIG. 29 is a configuration block diagram showing a pattern removal method using interference.

FIG. 30 is a diagram illustrating a signal processing method of the present invention.

FIG. 31 is a diagram illustrating a method of using the foreign matter detection system of the present invention.

FIG. 32 is a conceptual diagram of a detection rate β.

FIG. 33 is a diagram for explaining changes in the number of foreign matters attached to a wafer.

FIG. 34 is a diagram showing a correlation between process conditions and the number of foreign matters.

FIG. 35 is a diagram showing a barcode for recognizing a wafer.

FIG. 36 is a diagram showing the arrangement position of the barcode.

FIG. 37 is a diagram showing a method of correcting detection sensitivity.

[Explanation of symbols]

102 ... Illuminating means, 103 ... Detection optical system, 105 ... Rotation adjusting mechanism, 106 ... Spatial filter unit, 107
... detector, 108 ... rotation detecting means, 101 ... detection head, 212 ... pitch detecting means, 203 ... operator processing system, 206 ... foreign matter data memory, 207 ... large foreign matter data memory, 208 ... pattern memory, 210 ... software processing system , 211 ... Foreign matter memory, 232 ... Coordinate data creating means, 229 ... Microcomputer, 230 ... Display means, spatial filter unit 108, plural linear spatial filters 141, operator processing system 203, 4-pixel adding means 214, 8-value Conversion means 215, a plurality of line memories 216, clipping means 204, buffer memory 21
7, judgment pixel cutout means 218, operator cutout means 219, 231, foreign matter comparison circuit 220, threshold value setting circuit 221, a plurality of noise comparison circuits 222, threshold value setting circuit 223, pitch detection means 212, operator pitch calculation means 241, filter pitch calculation means 24
4, spatial filter control system 243,

─────────────────────────────────────────────────── ─── Continuation of the front page (51) Int.Cl. ⁶ Identification code Internal reference number FI technical display location // G06F 17/60 G06F 15/21 R

Claims (15)

[Claims] 1. A semiconductor device is inspected for foreign matter in a repeating pattern portion and a portion having a low formation pattern density in a chip of a semiconductor substrate during a semiconductor manufacturing process, and based on the inspection result and information on the semiconductor manufacturing process. A method of manufacturing a semiconductor, comprising: determining a state of generation of foreign matter in the manufacturing process and each manufacturing apparatus during the manufacturing process. 2. The semiconductor manufacturing according to claim 1, wherein the number of adhered foreign substances on the substrate is calculated based on at least the number of foreign substances detected in the inspection region of the substrate and the area of the inspection region. Method. 3. A foreign substance adhering to the substrate of the semiconductor is inspected during the semiconductor manufacturing process, and based on the correlation between the transition of the number of foreign substances on each inspected substrate and a parameter relating to the foreign substance generation. A method for manufacturing a semiconductor, comprising estimating a cause of foreign matter generation. 4. In the course of a semiconductor manufacturing process, foreign substances adhering to the substrate of the semiconductor are inspected, and the distributions of the inspected foreign substances are overlapped for each work step, and the transition of the distribution of the superposed foreign substances is checked. A method for manufacturing a semiconductor, comprising estimating the cause of foreign matter generation based on the above. 5. A semiconductor manufacturing method, wherein in a semiconductor manufacturing process, the generation state of foreign particles in a semiconductor manufacturing apparatus is controlled based on information on the distribution of foreign particles on an inspected substrate. 6. The semiconductor manufacturing method according to claim 5, wherein the method is performed based on the center of gravity or dispersion of the coordinates of the foreign matter. 7. The semiconductor manufacturing method according to claim 5, wherein the management of the generation state of the foreign matter is performed based on the average or variance of the foreign matter detection signal levels. 8. The semiconductor manufacturing method according to claim 5, wherein the inspection is performed during a semiconductor manufacturing process. 9. In a semiconductor manufacturing process, when inspecting a foreign substance on a semiconductor substrate, a repeating pattern portion and a portion having a low formation pattern density in a chip in the semiconductor substrate are selectively inspected, and a manufacturing process is based on the result. Also, a semiconductor manufacturing line that manufactures semiconductors by knowing the foreign material generation level of the manufacturing equipment, investigating the cause of foreign material generation, and taking countermeasures against it. 10. A semiconductor manufacturing method, wherein the number of foreign matters adhering to a substrate is estimated from a foreign matter detection rate and a detected foreign matter number obtained by multiplying a foreign matter detection area ratio and a detection rate in each step. A semiconductor manufacturing line according to claim 1. 11. A semiconductor manufacturing line characterized by estimating the cause of foreign matter generation by correlating the transition of the number of foreign matter for each substrate with a parameter expected to be related to the number of foreign matter in semiconductor manufacturing. 12. A semiconductor manufacturing line characterized in that, in semiconductor manufacturing, a cause of overlapping foreign matter generation is estimated at a working time point when the transition of the number of foreign matter for each substrate is expected to be related to the number of foreign matter. . 13. A semiconductor manufacturing line, wherein in semiconductor manufacturing, foreign particles are controlled based on the distribution of the foreign particles on a substrate. 14. The foreign matter is managed according to the center of gravity of the foreign matter coordinates or the origin of dispersion of the foreign matter.
3. The semiconductor manufacturing line described in 3. 15. The semiconductor manufacturing line according to claim 13, wherein the foreign matter is managed based on an average or a dispersion source of the foreign matter detection signal level.