JPH1090192A - Optical inspection of specimen using multi-channel response from the specimen - Google Patents

Abstract

PROBLEM TO BE SOLVED: To obtain an optical inspection whose inspection accuracy is enhanced by a method wherein a first response and a second response in a first pattern are detected, their results are compared with two responses from corresponding inspection points in a second pattern so as to be processed and a first difference signal and a second difference signal are data-processed together. SOLUTION: In a sensor 16 and a sensor 16', a bright-field signal and a dark-field signal are detected separately, the signals are input separately into a buffer 18 and a buffer 18' and the signals are input further to a defect detector 22 via a delay circuit 20 and a delay circuit 20'. The defect detector 22 performs an operation required to detect the position of a defect on a wafer 14. Then, a highly consistent defect list which combines the bright-field signal with the dark-field signal comprehensively is output. Its data is received by a post processor 24, and whether a captured defect is a pattern defect 44 or a dust particle 46 is determined.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an automatic optical inspection of a sample (for example, a semiconductor wafer). In particular, a single search is performed during the inspection to determine the presence of defects on or within the sample, and thereby at least two independent inspection responses (eg, brightfield and darkfield) (Reflection from the sample). This inspection response can be collectively subjected to data processing in a unified manner to determine the presence of a defect.

[0002]

2. Description of the Related Art Conventionally, it is difficult to optically inspect a wafer.
It is said that there were two technologies. In general, bright-field illumination,
Darkfield illumination, and three methods using spatial filters.

A technique that has already proved its effect in inspecting pattern defects on a wafer is a bright field illumination method using broadband illumination. This method can minimize contrast fluctuations and narrow band-specific coherent noise. The most successful example is the KLA Model 2130 sold by KLA Instruments. The device can operate in either a die-to-die comparison mode (die-to-die comparison) or a cell-to-cell comparison mode for repeated cells (cell-to-cell comparison). However, bright field illumination may not provide sufficient sensitivity to small dust particles.

[0004]

In the image processing in the bright-field observation, a light beam to be focused on an opening for condensing is caused by a small dust particle being diffused to reduce the energy of reflected light. This is the principle of detecting dust particles. According to this principle, when dust is small compared to a point spread (optical point spread function) function value unique to each lens or a pixel size for digitization, the generated energy loss is small. The presence of garbage does not affect observations. Therefore, it is difficult to detect dust in bright-field observation. If the generated energy loss is small, the location dependency of the reflectance of the surrounding bright area acts like noise, and the energy loss information due to dust is hidden. Naturally, if the location dependence of the peripheral reflectance acts like noise, this is not true but false.
sance defect). In bright-field observation, dust is often hidden by false defects.

[0005] Even if there is a region having a low reflectance around smaller dust particles, it is often seen on a wafer, and a reticle, a mask, and a liquid crystal panel have essentially low reflectance around the dust. Even if the reflected energy is small, it is even more difficult to detect if the energy is further reduced by dust.
Wafers, reticles, masks, liquid crystal panels, and the like, as well as other samples, currently employ dark field illumination in commercial inspection and measurement instruments for the detection of small dust particles. In dark-field illumination, a flat specular surface returns very little signal to the detector, so the diffuse image is dark,
Hence the term dark field is used. On the other hand, any foreign matter protruding from the surface can return a larger amount of light energy to the detector than the surroundings. It is understood above that in a dark field, an image usually looks dark except for dusty particles protruding or regions of a circuit pattern having irregularities.

The dust particle detector used in the dark field was designed and manufactured based on a simple hypothesis that dust particles scatter more light than circuit patterns. While this device works well for samples without a pattern called a blank, it can only detect large dust particles protruding high above the pattern when the circuit pattern is in the background. As a result, of course, the detection sensitivity remains low, which is not satisfactory at the most advanced advanced VLSI integrated circuit manufacturing sites.

[0007] In connection with dark field, there are measuring instruments that specialize in some characteristic aspects. A measuring instrument of Company A is based on the principle of distinguishing it from a circuit pattern by paying attention to the polarization characteristics of reflected light scattered by dust particles. That is, the principle is that when incident light collides with dust particles, the reflected light exhibits a polarization state. However, when the pattern size becomes smaller with respect to the wavelength of the incident light, the pattern size gives a polarization state to the reflected light similarly to dust. This only increases false defects. Therefore, this method can be used only when the pattern size is large.

In addition, a mechanism for adjusting the angle of the incident light, such as company A and company B, is incorporated so that the light scattered from the pattern can be adjusted to 0,45.
And a device in which the component at 90 degrees is minimized. This is generally an improved darkfield instrument, but the scattered light from the corners of the pattern is still strong, giving a large noise component. In addition, in order to reduce the "effect of pattern corners" and prevent false defects, it is necessary to intentionally reduce the detection sensitivity of a region having a high-density circuit pattern.

Another method currently used to enhance the detection sensitivity of dust particles is to use a spatial filter. In plane wave illumination, the light intensity distribution at the rear focal plane of the lens can be represented by Fourier transform of the object placed there. In the case of a repetitive pattern, the Fourier transform becomes an array of light points. If a filter is installed on the back focal plane of the lens to block the arrangement of the light spots, an appropriate spatial filter effect is given to the repetitive pattern, and the repetitive component disappears. Therefore, only non-repetitive components, that is, only dust and defect images pass through. Company C, Company D and Company E
The company therefore primarily employs spatial filter technology.

On the other hand, the measuring apparatus based on the spatial filter technique has a serious problem that it can inspect only a blank or an area having a repeated pattern as described above. That is a fundamental limitation of this technology.

The dark field spatial filter technology employed by a model of Company A uses a die-to-die image comparison and subtraction method for wafer inspection. With this technique, non-repeated pattern areas on the wafer can be inspected by die-to-die comparison. However, even with the die-to-die comparison method, it is still necessary to use a spatial filter in the repetitive pattern area to obtain high sensitivity. In high density memory cells, the signal from the normal internal circuit pattern in such a case is significantly stronger than the peripheral circuitry and exceeds the proper dynamic response limit of the sensing sensor. This leads to a saturation phenomenon in the sensor, so that no small dust particles are detected. Otherwise, the signals from the small dust particles in the peripheral circuits will produce results that are too weak. The problem is that in each case small dust particles are not detected.

Further, the dark field type spatial filter type die-to-die defect inspection apparatus of Company A has two major disadvantages. First
This means that this device detects only dust particles and cannot detect any pattern defects. Second, in the case of an image that has passed through a spatial filter, the screen is dark in a normal state without a circuit pattern, so that accurate position alignment for die-to-die cannot be performed. Of course, in order to use the comparison and subtraction algorithm, accurate position matching is essential. Company A tried to solve this problem by introducing an expensive high-precision stage mechanism. However, the introduction of such an expensive high-precision stage mechanism has not been a satisfactory solution. It can be said that this point is caused by the synergistic effect of the residual error of the stage and the fact that there is no fluctuation of the pattern alignment on the wafer due to the mechanical error at the time of the stepper projection operation. Therefore, the results are not satisfactory, and the size of this dirt particle cannot be reduced below about 0.5 micron in a situation where the detection of small dust particles is not noticeable. Such a limitation arises because the position cannot be accurately adjusted for the die-to-die inspection.

There are some patent applications other than Company A, which are described in US Pat. No. 5,276,498 and US Pat.
No. 806,774 and U.S. Pat. No. 5,177,559. However, apart from these listed prior applications, there is no precedent that has shown interest in the combination of brightfield and darkfield techniques due to a lack of understanding of the advantages of the technique.

[0014] Among the inspection devices that enable both bright field and dark field observation, all models on the market use a single light source common to both bright and dark fields. In addition, there is no technical method for determining the defect by integrating the inspection response signals from the bright and dark fields and the visual field. Microscopes with both brightfield and darkfield illumination on the market have only a single light source that provides both illumination simultaneously, so it is not possible to separate the signals of the brightfield and darkfield observations, respectively. It is possible. The only commercially available Zeiss microscope that can simultaneously separate at least the brightfield and darkfield light sources, however, there is only a single detector and a method to separate brightfield and darkfield illumination. No idea. There is only one dome-shaped light that can illuminate the whole sky.

It is considerably beneficial to have two separate illumination devices, brightfield and darkfield. Both advantages are directly linked to the enhancement of the inspection equipment. Accordingly, it is an object of the present invention to provide a method and apparatus for testing as discussed below. The present invention detects and correlates information from brightfield and darkfield separately and produces good results which are completely unexpected from known examples.

[0016]

SUMMARY OF THE INVENTION The present invention refers to at least one known inspection response of a second pattern of the same design to search for a defect present at an inspection point of a first pattern on a sample. It describes the details of the defect inspection method and its apparatus characterized by the following. Testing involves performing at least one search and generating at least two test responses accordingly. It is important to use the same observation points on the first and second patterns on the sample.
The two responses (response signals from the dark field and the bright field are representative) are separately detected by a photoelectric method, compared separately, and individually subjected to a difference signal (a difference between the first and second patterns). ) Is formed.

That is, the first and second responses in the first pattern are detected, and the results are compared with the two responses from the same inspection point of the corresponding second pattern. , Resulting in a first and second differential signal of the response. The individually generated difference signals are subjected to data processing to determine a list of first pattern defects in a centralized manner. Specifically, the first and second difference signals are collectively subjected to data processing to determine a unified first pattern defect list.

Data processing is then performed on the first pattern defect list. Then, the false defects found on the known and harmless sample surface are extracted and eliminated. On the other hand, such known harmless false defects are provided to the user for reference. In addition, various test searches are added to increase the test response so that more than one optical response is obtained and processed from the sample. Thereby, the inspection accuracy is further improved. In addition, in the case of a transmissive sample, a photoelectric detector can be installed at the back of the sample to collect the inspection response depending on the transmitted light, and the accuracy of the pattern defect list can be further improved. It is also possible to search for a defect buried in.

[0019]

According to historical fact, in the majority of defect inspection systems, the illumination means uses either brightfield or darkfield, and both illuminations are used. It can be said that there was no example of using the means. In a typical prior example viewed from the historical background of such a defect inspection apparatus, only one of a bright field or a dark field illumination system is used at the time of observation as shown in FIG. In the prior art FIG.
Illuminated by an illumination system having a suitable brightfield 10 or darkfield light source 12.

Here, PLLAD (Phase Locked Loop A)
nalog to Digital). This is Phase
A known method for A / D conversion of analog data into digital data by the Locked Loop method.
Called LAD. Also, TDI (time delay integration)
Indicates a method of delaying and integrating signals. TD during operation
An image of the wafer 14 is captured by a sensor 16 in which I and PLLAD are combined as shown in FIG. 1, and a signal carrying the image is input to an input buffer 18 (for example, RAM). The data is transferred to the defect detector 22 via the input buffer 18.

Inspection data obtained from a sample wafer die at defect detector 22 is compared with like data from another similar die. Alternatively, the comparison partner may be from another reference wafer as long as the data is of the same type from another similar die. This comparison operation is controlled by the delay circuit 20. Providing the correct delay timing by the delay circuit 20 allows for a die-to-die (die-to-die) or cell-to-cell (cell-to-cell) comparison operation. Defect detector 2
2 is input to the post-processor 24. Here, an operation of determining the size and position of the defect is performed. Thus, it is possible to output a defect list with a threshold value (for example, both KLA product models 2111 and 2131 are designed to operate on the above principle using bright field light). .

It is considered that the apparatus shown in FIG. 1 is modified so that the observation results of bright field and dark field can be obtained independently and separately. This is a completely new approach and has never been done as prior art, so that such systems themselves or their information cannot be obtained from others. The simplest and most obvious way is to carry out such a test in a serial, serial and non-interfering manner in realizing such a function. In the inspection 1, the light source 10 is used to make a bright field. In another inspection, the light source 12 is switched to a dark field.

According to the known art, the first inspection is performed with bright-field illumination, and in the next inspection, the wafer 14 is inspected under the dark field by the dark-field light source 12. The image is sensed and its data is stored in buffer 1.
8 and a delay circuit 20 to reach a defect detector 22 and a post-processing processor 24. As in the case of the bright field image, it is possible to output the defect list 28 with a threshold value for the dark field image in this way. At this time, the post-processor 24 independently outputs the threshold value of the defect found in the dark field. That is, the conclusion here is that it is possible to output a defect list with a threshold even in dark field.

Consider an inspection point on the wafer 14, that is, an image generation point. This image generation point is the bright field defect list 2
In the case of coincidence with the defect occurrence point of No. 6, the data value naturally exceeds the threshold value of a bright field defect. At this time, it can be said that the previous image generation point on the wafer 14 contains a defect. If the same operation is performed on the dark field separately from the above, if a certain image generation point on the wafer 14 coincides with the defect generation point in the dark field defect list, the data value is naturally a dark field defect. Threshold is exceeded.
At this time, the previous image generation point on the wafer 14 contains a defect. Therefore, the image generation point on the wafer 14 may have only one of the bright-field type defect and the dark-field type defect, or may have both types of defects at all. Both of the latter are possible. Thus, the post-processor 24 can output two independent and uncorrelated defect lists. One is based on the image from the bright field light source 10 and the other is based on the image from the dark field light source 12.

FIGS. 2A and 2B are both publicly known,
It is an example which shows the process which determines whether inspection data is a defect. That is, the decision boundary line (34) in FIG.
And 40) need to be established in brightfield and darkfield, respectively. Below the boundaries, reference numerals 32 and 38
As shown in FIG. 5, the region which does not become a defect on the wafer 14 is shown. On the other hand, above the boundary line, as shown by reference numerals 30 and 36, a defective region is shown. In developing the discussion of the present invention below, it should be pointed out that there is a problem in that the boundary line of a defect is not actually represented by a line segment but is treated as a line segment in the known technology. . Referring now to FIG. 3, this figure forms a graph with the number of data step differences in the bright field on the horizontal axis and the number of data step differences in the dark field on the vertical axis. Here, the number of data step differences means that the signal strength is expressed in 256 steps in ascending order. Next, as described above, a comparison operation of signal strength is performed, and numerical values indicate what differences exist in the signal strength stages of the data. Brightfield and darkfield threshold lines are drawn along each axis.

Although there is no recording performed, it is possible to use both the bright field and the dark field by setting using a combination of known techniques.
The use of this improves the accuracy in determining defects.
At this time, only when the data is plotted in an area (area indicated by reference numeral 38 in FIG. 3) in which the data both exceed the threshold values of the bright and dark visual fields, it is only possible to determine this as a defect and recognize it. I can do it. As an example of an apparatus which can use both the bright and dark fields simultaneously, there is an illumination from the whole sky, that is, a dome type as shown in FIG. The literature which introduced this is Yasuhiko Har
a, Satoru Fushimi, Yoshimasa Ooshima and Hitooshi
Kubota, "Automating Inspection of Aluminum Circuit
Pattern of LSI Wafers ", Electronics and Communic
ations in Japan, Part 2, Vol. 70, No. 3, 1987.

With reference to FIG. 1, a dome type illumination system will be considered. Here, it is assumed that both illuminations are simultaneously performed on the wafer 14 using the reference numerals 10 and 12 in FIG. Reference numeral 16 denotes a sensor, which shares a single sensor between light and dark and both visual fields. It is not only the sensor that is shared by the light and dark and the two fields of view, but also from the process equipment group 18 to 24 (processor channel),
All are the same. As a result, a single output is realized, and the result is shown in FIG. It should be noted here that FIG. 2 (c) does not mean that the separated inspection was performed twice in the bright field and the dark field. Here, it is pointed out once again that the problem is that the threshold value is determined linearly and is different from reality.

FIG. 4 is a known diagram showing a dome-shaped lighting device of the first type.
This is a second modification of the defect search apparatus. Here, the defect search can be performed simultaneously in the bright field and the dark field. In FIG. 4, reference numerals 18 to 2
4 sets of 2 processor channels,
The first is dedicated to brightfield and the second is dedicated to darkfield. Now, consider the light source in this case. The light sources may be single or they may be paired instead. A pair of light sources refers to a case where one light source is set for each of the light and dark and each visual field, and a total of two light sources are set. There is also a case where two light sources are used at the same time as a case where a pair of light sources are used sequentially and a dome type lighting. In each case, the bright field and dark field illumination light are given to the wafer 14.

The difference between FIG. 4 and FIG. 1 is whether only one or two processor channels are provided. Thus, even if there is a difference between sequential and simultaneous, a bright and dark field inspection can be performed. The number of processor channels held is not an essential difference. Therefore, the bright field defect list 26 and the dark field defect list 28 can be obtained independently of each other in the simultaneous inspection or the separate inspections. If the channels of each processor are operated in synchronization with each other, the speed of the slowest channel is matched, so that the sweep for capturing the image on the sample needs to be performed only once at the synchronization speed. As a result, defect list data for both bright and dark fields can be obtained at a time.

Therefore, the system shown in FIG. 4 is faster and more effective than that of FIG. 1 (since only one sweep is required). In the following, the data of the defect list for both bright and dark fields will be referred to as a map. Here, the data processing can be performed in parallel with the sample positioning operation, which is very convenient. However, both the known system of FIG. 1 and the known system of FIG. 4 independently determine the threshold value of the defect (lists 26 and 28).
It is different from reality because it employs the simple segment method described above. This has to be called a malfunction. That is,
A discussion of the problems of the simple line segment method is shown in FIG.
(A), (b) and FIG. 3 all apply.

The details of the present invention will be described below with reference to FIG. This is a block diagram of the present invention. The sensors 16 and 16 'are provided, the brightness of the image of the wafer 14 and the respective field data are each separately captured, except that the left side of the diagram of FIG. Same as the part. Here, naturally, the sensors 16 and 16 'separately detect the signals of the bright field and the dark field, and then these signals are buffered 18 and 18'.
, And the signals are sent first through separate delay circuits. The similarity between the present invention and the known example of FIG.

A single defect detector, designated 22 in FIG. 6, receives the bright and dark, both field of view signals which are separately input to buffers 18 and 18 'and are passed through separate delay circuits. I do. The defect detector 22 performs an operation necessary for finding the position of a defect on the wafer 14. The defect detector 22 outputs a highly consistent defect list that comprehensively combines light and dark and both visual fields. This data is received by the post-processor 42, and the captured defect is detected as a pattern defect 44.
Or dust particles 46 are determined.

In FIG. 6, the post-processor 42 is a VME (Virtua) based on a Motorola 68040 CPU having high performance and versatility.
l The bus process board of Machine Environment can be used. Alternatively, an example of KLA Model 2131 known as a high performance post-processor board may be used.
As is well known, a semiconductor wafer may have a pattern on its surface that changes depending on its location in terms of its contrast. It may have a grainy pattern called grain. Further, there is a case where the grains are formed as a lump. In addition, manufacturing processes frequently fluctuate. Irregular shaped patterns (marks) due to chemical causes also adhere. Since these hardly cause problems of yield and reliability, it is assumed that there is no need to recognize them as defects. The patterns, each associated with surface fluctuations, emit signal values in the typical range of images in bright and dark fields when viewed during inspection. In addition, it is necessary to consider noise. If the signal value is small, the noise disturbs the signal, especially the brightness, and the value of the number of steps in each field of view, which has a bad influence. Thus, for a typical example, the relationship between light and dark, the number of steps in each field of view, and the system noise was plotted. This is shown in FIG.

In FIG. 7, the system noise 54, the surface contrast variation 56, and the grain 58 are all close to the origin, and their values are small in both bright and dark fields. On the other hand, the process variation is as high as 75% in the bright field and about 50% of the midpoint in the dark field. Also, when the grains are generated in a lump, both the light and dark and the both visual fields have large values. Ideally, the best system would be able to eliminate these predictable and harmless fluctuations without recognizing them as flaws,
This means that only the signal indicated by 48 in FIG. 7 is recognized as a defect.

FIG. 6 is a partial block diagram of the circuit shown in FIG. Here, the defect detector 22 is shown in detail. Although this defect detector 22 is shown briefly in FIG. 6, when receiving an input signal, the signal passes through input buffers 18 and 18 'and delay circuits 20 and 20'. The defect detector also comprises a group of filters, 90 and 92 and 90 'and 92'. Filter groups 90 and 92
And 90 'and 92' process the image data.
At this time, 3 × 3 or 5 × 5 pixels are digitally filtered. This is also KLA Instruments
There is an example in which the product Model 2131 of the company is applied. The image signal is pre-processed by filters 90 and 92 and 90 'and 92' before being input to subtractors 94 and 94 '. here,
A comparison operation (subtraction) is performed for each image pixel in order to compare with the corresponding delay data. In a typical example, in the case of a die-to-die comparison (die-to-die), the delay time difference is one.
It corresponds to the width of the die. In the case of cell-to-cell comparison, the delay time difference corresponds to the width of one cell, and the same amount of delay is applied to light and dark and each field of view. Thus, the outputs from subtractors 94 and 94 'are, of course, the light and dark,
It is the defect information itself in both fields of view. That information can then be provided to the two-dimensional histogram circuit 96 and the post-processor 42 alternately. On the other hand, the signal components directly input to the post-processor 42 are given the coordinate values of the data in FIG. Then, the two-dimensional histogram circuit 96 forms a two-dimensional histogram diagram of the defect separately in both the bright and dark fields. The two-dimensional histogram signal obtained by such processing is supplied to a decision algorithm 98 for determining whether or not the signal is a defect. And
Known and frequent harmless wafer surface and other fluctuation dependent signals (these are known and frequent problems such as system noise, grains, contrast fluctuations, process fluctuations, grain clumps, etc. as described above) , Which are not defects, and are collectively referred to as “false defects” hereinafter.

The foreseeable and harmless false defects obtained by known prior art methods should not be confused with true defects. FIG. 8 shows what to do for that. That is, as shown in FIG.
Although the threshold value determined by the straight line indicated by 0 does not apply to an actual defect, it has to be adopted because there is no other method in the known example. As can be clearly understood from the above, the region 38 is a proposal as a defect region in which the data of both bright and dark fields are comprehensively summarized, but it is considerably small and out of target. This is not suitable for real use.
FIG. 7 is referred to again. In the present invention, the defect detector 22 can be programmed in various ways since the data of the two fields of view are acquired and processed independently and simultaneously. Therefore,
A complex non-linear threshold function can be applied to that step difference in both light and dark, both fields of view data to eliminate false defects and only black portions excluding white regions in FIG. It can be separated as a signal.

In this way, false defects are safely eliminated. In other words, in the present invention, all 0 to 255
7 can handle both bright and dark signals and signals in both fields of view, and furthermore, false defects 5 and 5 shown in white in FIG.
0, 52, 54 and 58 and 56, all of which can be avoided.

Next, FIG. 9 shows one physical optical embodiment of the present invention. The wafer 14 has the dark field light source 12
(Eg, a laser) and a bright field light source 10 (eg, a lamp based on the mercury arc principle).
At this time, the light beam passes through the two lenses 60 and 62, and further passes through the beam splitter (beam splitter) 66. The light beam carrying the image signal of the wafer 14 according to the thus integrated brightfield and darkfield is a condenser (compression).
Upward through the lens 60, the beam splitter 64
Pass through the beam splitter 66. After passing through the beam splitter 66, the bright-field light travels further upward and passes through the condenser lens 7
2 and enters the bright field sensor 16. On the other hand, the dark field light is reflected by a film (dichroic film) on the beam splitter 66 on which a dichroic film is applied. This is because the wavelengths of the light sources in the dark and light fields of view are different from each other, so that the light sources can be designed to be so. The dark field light passes through a spatial filter 68 and is incident on the relay lens 70, and then reaches the dark field light sensor 16 '.

As used herein, dark field illumination is configured by a combination of a laser and a spatial filter 68. Therefore, in this system, a Fourier transform plane of the image of the wafer 14 can be formed. The purpose of using the spatial filter 68 is to select and remove a normal pattern without a defect. It is pointed out that the search accuracy of the defect is further improved by this. As described above, the present invention uses two separate light sources, a mercury arc light source passing through the beam splitter 64 in the bright field and a laser light source capable of spatial filtering in the dark field.

Here, since the laser light source has a high luminance and a high output, even if the information of the bright field and the information of the dark field are handled separately from each other, the light loss is small, that is, only a few percent. Considering that a narrow-band laser light source is sufficient for dark-field illumination, a helium-neon laser with a wavelength of 633 nanometers can be used. Alternatively, laser diodes with wavelengths between 630 and 830 nanometers can be used. Then, the dichroic coating material (dichroic film) on the beam splitter 66 can reflect only dark field light well. Alternatively, dark field light can be easily separated by using Oriel Model No. 52720 as a laser light interference filter. Combining the narrow-band spectral filter with the latter, Oriel's model number 56460 can be used for mercury filters for bright field light. Alternatively, Oriel has a custom designed narrowband notch filter for laser applications that can be used. As a result, spatial filtering is used only for the dark field optical path. In this way, the quality of the image is kept high, without affecting the bright field optical path. When a narrow band light source is used (above, an example using a narrow band laser light source for dark field illumination), a spatial filter operation is required. It is generally easy to separate the bright-field and dark-field signals from each other using a filter or a beam splitter using the narrow-band characteristic characteristic of laser light. The spatial filter 68 inserts a piece of photographic negative film as shown in FIG. 9 and exposes it. After the exposure, the film is taken out and developed, so that a spatial filter can be manufactured. Of course, after development, the film itself becomes a filter, so it is set in the correct location indicated by 68 in the figure. In addition to this, the spatial filter 68 is a fuse field effect filter SLM (Spatial Light Modulator),
This can be realized using a liquid crystal LCD (Liquid Crystal Display).

It has been stated that the preferred embodiment has the various options described above, in the case of successfully separating only the dark-field signal component from the entire image signal beam. However, it is pointed out by the present invention that the use of the reflection effect of the dichroic film on the beam splitter 66 in combination with the dichroic film in combination with the spatial filter 68 is the most effective overall. I do. The reason can be clearly understood in consideration of the controllability and the wide dynamic range. Since the dark field signal is processed simultaneously with the bright field signal, it is particularly important to consider the controllability and the wide dynamic range.

The method of using the reflection effect of the dichroic film on the beam splitter 66 has been described above. However, with the development of current optical technology, or with more advanced technology, it is quite possible that there are still unknown methods.

FIG. 10 is a diagram showing the principle of the second embodiment of the present invention. In this embodiment, a single laser 7
6 to provide both brightfield and darkfield illumination to the wafer 14. Here, the beam splitter 80 reflects light downward, and the light beam enters the wafer 14 via the condenser lens 78. In the present embodiment, the image processing is performed simultaneously in light and dark and in each field of view. The dark field detector 74 is installed at a small elevation angle with respect to the wafer 14, while the bright field sensor 82 is installed directly above the wafer 14. At this time, the light beam passes through the condenser lens 78 and the beam splitter 80. In this embodiment, in order to optimize the efficiency of the defect search, it is preferable that the output of the bright field detector 82 and the output of the dark field detector 74 are simultaneously processed for the target defect. Using a bright-field image set in a wide band and a dark-field image formed using a spatial filter for die-to-die comparison (die-to-die) is a legitimate method and overcomes all the limitations of previous inspection equipment. I do. Since a bright-field image exists, it can be used to conveniently align the two dies for comparison with high accuracy. If the positions of the sensors in the dark and light fields are aligned (pre-aligned) in advance, the same area can be observed and inspected accurately. Then, the alignment deviation between the two dies to be compared may be measured in the bright field channel, and the alignment is automatically performed in the dark field channel.

Adjustment and calibration are performed at the time of manufacture, and the offset (offset) of the alignment is, as is obvious from the above, the value thereof is fixed between light and dark and between visual fields. The offset amount is fixed and device specific and is also a known value. Thus, a high-speed electronics system for the measurement for offset need only be provided for the bright field, so that it is not needed in the dark field channel. If the alignment information obtained from the bright-field image is used, even in the dark-field channel, a considerably accurate die-to-die alignment is performed for the above reason, so that the residual error of the alignment is sufficiently small and small dust particles on the sample. There is no hindrance to the search for etc.

As described above, in the data processing of information obtained in a dark field using a spatial filter, at present, most of the repetitive pattern on the die and an area having many straight lines are filtered. It is desirable to adopt a method of excluding in the above. By doing so, the dynamic range (operating range) becomes uniform, so that all small dust particles can be captured in one search regardless of whether the pattern is in a high-density state or in a sparse state. In addition, the technique of processing the dark-field image and the bright-field image simultaneously provides more information about the defect. For example, a bright-field image can capture both pattern defects and defects such as small dust particles at the same time, whereas a dark-field image can only detect defects such as small dust particles. It is possible. Therefore, if the difference between these two results is taken, the answer is only the pattern defect. The ability to automatically separate defects such as small dust particles from pattern defects in real time can be said to be a very characteristic ability in the technology of the present invention.
The value of this in wafer inspection is immense. In this special application, for a defect such as a small dust particle, a much higher detection sensitivity is realized in a dark-field image than in a bright-field image. Therefore, matching (coincidence) with the case of the bright-field image can be achieved by reducing the detection sensitivity in the dark-field image. As a result, the defect obtained from both the bright and dark channels can be regarded as a defect such as a small dust particle, and the defect obtained from only the bright-field image can be a pattern defect.

Another example will be described. This is an example of inspection on a metal wiring layer (layer) on a semiconductor wafer. Combining the results obtained in the dark and bright images and the two fields of view, it is possible to successfully separate false defects created by grains (granular patterns on the surface) of metal wiring from true defects. This need is great and therefore of great importance. Both the light and dark images and the delayed images corresponding to them are received and stored separately. Thereafter, as shown in FIG. 5, the stored data is taken out by aligning the die-to-die, and inputted to the defect detector 22. Performing the detection operation in this manner requires a gigabyte-class dynamic memory (DRAM).
The dynamic memory stores the detected data, and the data is read out from the memory and used as necessary while considering the timing as described with reference to FIG. Although the system operates in this manner, the system shown in FIG. 5 seems to be a little ahead of the times. With today's technology, it is not only preferable to inspect images of both visual fields and light and dark fields in real time in real time in the wafer inspection, the result is obtained quickly, but also the cost is reduced. Either method may be used, but it is, of course, important to obtain bright and dark images of the same observation point on the wafer 14 and images of both fields of view and input them to separate detectors. . It is very important to recognize the fact that the inspection is being performed on exactly the same observation point on the wafer 14, and based on this recognition the relationship between bright and dark, both-view images becomes apparent. (For example, if the dark-field image is very bright, but the bright-field image is fairly dark, understand what this means.) The two image signals in both bright and dark, both fields, are simply If the signal is input to one detector after the addition, the same result as in the present invention is not obtained. The reason will be described below. If the superior results of the present invention are ignored and the latter is adopted, the differences between the two image signals in the bright and dark, both fields of vision will cancel each other out and will simply lose important information and capability in detecting defects. Ends in This is where the true value of this invention lies. In the case of light and dark, each field image can set the illumination light to a completely different light source. In addition, the incident angles of the light rays from the light source are supplied while being changed considerably from one another. Since the obtained images are input to separate detectors, all the actual situations on the wafer 14 can be completely drawn. In order to perform this operation, it is necessary that the two detection sensors are linked to each other and that their optical axes are correctly aligned. However, associating such detection sensors with each other and aligning the optical axes has the opposite effect in reducing the cost of the defect search device and increasing the complexity. On the other hand, there is a merit of the present invention described above, and there is no known example, so that the merit of the present invention is difficult to understand.

Further, in the discussion so far, it has been assumed that light having a single frequency (single wavelength) is used for a light source of a bright field or a dark field in detecting a defect. However, there is no reason why the technique of the present invention must stop there. It is right to think about further extensions, which can increase the number of channels of information. That is, a multi-channel can be constructed using a light source having a plurality of frequencies in a bright field or a dark field. The key to this extension is 2
The technology of the channel, of course that is the basis. When the main point is repeated again, the first point is to use the same observation point on the wafer 14. Next is to perform photoelectric data processing with individual detectors in each field of view, dark and light. Then, the detection results are arranged as shown in FIG. 7 (the case of two channels is discussed here, but the same applies to multiple channels). In the case of multi-channels of two or more channels, FIG. 7 is not two-dimensional but naturally multi-dimensional. It is almost impossible to properly represent three or more dimensions on paper. However, in multidimensional cases, computer-based numerical methods are known to be useful.

FIG. 11 shows a system obtained by developing FIG. This is an example in which a multi-channel bright field and a multi-channel dark field are used. Specifically, each of the two channels has a total of four channels. Therefore, FIG. 11 is based on FIG. 9, but repetitive description of FIG. 9 is omitted here. However, all of the elements described in FIG. 9 function as they are in FIG. In the second dark field channel, laser 12 ′ is oscillating at a different frequency than laser 12 but is additionally introduced to illuminate the same location on sample wafer 14. The details of observing the sample wafer 14 will be continued. In FIG. 11, a second bright field illumination is further added.
0 '. This is naturally set so as to emit light having a different frequency from that of the light source 10. Lens 62 'and beam splitter 6
4 'is needed to direct bright field illumination to the same location on the sample wafer 14. Even in the case of the reflection mode, there is not much difference in the operation content from that of FIG. A beam splitter 66 'using a dichroic film coated with a dichroic film is further added so that the light from the second laser 12' is reflected. This reflected light is applied to the spatial filter 6
The light is incident on the detector 16 "through the 8 'lens 70'. In addition, the beam splitter 73 having a dichroic film coated with a dichroic film is used as a beam splitter 73 of the bright field illumination light source 10 or 10 '. Either one is selected and set so as to be reflected, so that the reflected light is incident on the detector 16 '". At this time, the bright-field illumination light beam that passes through the beam splitter 73 has a different frequency from the reflected light.
The bright field illumination light beam passing through beam splitter 73 can be set to be split as it passes through beam splitters 66, 66 'and 73'.

The embodiment of FIG. 11 is merely a modification of the embodiment shown in FIG. The invention can provide a wider range of examples. When a special defect is detected separately from other defects, it is necessary to combine various elements. Embodiments in which various elements are combined in this way are different from those described here, but the concept can be considered to be the same. When the obtained signal is determined to be "defective" using each information in the multiple channels, the present invention is characterized in that the signal information from both dark and light fields is used simultaneously. All known examples are different, and the information channel is limited to a single information channel. On the other hand, in the case of a multi-channel, as described above, a plurality of wavelengths can be set in both bright and dark fields, and a light source corresponding thereto can be used. Moreover, the techniques used here for searching for defects in a wafer can be further extended, for example, to search for defects on transparent materials. In such applications,
Bright field light and dark field light due to transmitted light are made incident on the detector, while some light is reflected, so bright field light due to reflected light,
Searching for various defects by integrating each electric signal of dark field light,
Further, the position can be recognized. FIG. 12 shows an embodiment used for executing the above, but the actual state is simplified.

FIGS. 12 and 9 are substantially the same, except that the latter optical signal detector portion is reproduced below the sample 14 'to construct the embodiment of FIG. The combined light beam carrying the image information of the transmitted light in each of the bright and dark fields travels further down from the bottom surface of the sample 14 'and passes through the compression lens 60T through the beam splitter 66T.
To reach. At the beam splitter 66T, the bright-field image signal light beam travels further down to the compression lens 72T, whereupon the bright-field light sensor 16T dedicated to transmitted light.
Set to be projected on top. On the other hand, it is possible to set so that only the image of the dark field light among the transmitted light is reflected by the film (dichroic film) coated with the dichroic film on the beam splitter 66T. The reason for this is that the wavelength of the illumination light source in each field of view is generally different between light and dark. The remaining dark-field light passes through a spatial filter 68T and passes through a relay lens 70.
The light passes through T and enters the dark field optical sensor 16T '.

The concept of the present invention has been described above with respect to the specific case of the bright field and dark field illumination sources. That is, it stipulates that the light irradiating the sample uses the illumination light source in each of the bright and dark fields of view, and the resulting signal and its detection are performed independently. In general terms, the present invention can be configured to include all the following elements a), b) and c). a) at least one
There are two image search mechanisms (probes) which are arranged to have at least two independent optical response systems in searching for the same location on the same die of the sample to be examined. If more than one probe is used, all probes are arranged to search the same location on the same die of the sample to be tested. b) In order to perform a search operation through each optical response system, an operation of comparing each image data with the image data at the same position of the other die is performed as described above. Generate. c) Further, the same operation is performed for other optical response systems, and a multiplexing operation is performed to create a first list of pattern defects based on the same principle. We described above how to use each field of view, dark and light, and also discussed extending it to a generalized process. Now, post-processing (post-processing) can be performed on the obtained first list of pattern defects as follows. That is, this post-processing should be performed on known false defects on the surface. False defects are not "true" defects, as mentioned earlier. Therefore, it should be excluded from the final pattern defect list. This post-processing makes it possible.

In the discussions made with respect to the above-cited drawings, one or more probes may be provided to generate one or more sets of optical responses. In FIG. 10, a single probe is used. Laser 76 is used as an illumination source to generate both bright and dark field images of the sample.
The wafer 14 is introduced, and two other independent detectors are provided, one is a detector 74 dedicated to dark field, and the other is a detector 82 dedicated to bright field, so that the information signal system has two channels. I have to. In FIG. 9, there are two probes. The laser 12 is a common illumination light source of two different dark-field systems and irradiates the sample way 14. Separately, the lamp 10 is a light source of a bright field system. There are two mutually independent detectors 16 and 16 'which receive reflected light from the sample in both bright and dark fields to form two channels. FIG. 11 is obtained by expanding the system of FIG. Here, a second dark-field light source and a bright-field light source are added, respectively. The result is four probes. The detector also added 1 to each of the bright and dark visual fields to make the information system a total of 4 channels.

On the other hand, FIG. 12 is similar to FIG. 9 except that the number of probes is two, and an illumination system is provided for each of the bright and dark visual fields. In addition, a detector for transmitted light in the bright field and a detector for radiated light (reflected light) in the dark field are provided, for a total of four channels. That is, there are two channels of transmitted light and reflected light in a bright field, and two channels of transmitted light and reflected light in a dark field. In each of the above examples, there is no shift in the frequency and phase of the optical signal. That is, there is no difference in frequency and phase between light emitted from the probe and light incident on the detector. What is happening is the separation of the optical signal beam.

Fluorescence is one of the photoreactions, and is a phenomenon that is widely known in some kinds of materials, and occurs when the material is irradiated with light in a certain wavelength band. When a material emits fluorescent light (secondary light), the frequency (wavelength) of that light
Is smaller (has a longer wavelength) than the primary light, that is, the probe light. In some kinds of materials, it may be effective to observe a shift in frequency due to the generation of fluorescence when searching for defects. The frequency of fluorescence in luminescence is intrinsic to the material and is well studied. The applied dichroic film used in the beam splitter according to the present invention can be made of a material that is sensitive to the frequency of the incident light, and the detector can be so. Beam splitters and detectors that are sensitive to the frequency of the incident light, along with other useful techniques, can be introduced into the system of the present invention.

Similarly, if the optical path from the probe to another inspection area on the sample (there is a case where there is a step on the surface of the silicon wafer or the refractive index of light is different in a distant area on the wafer), reflection occurs. The generated illumination light causes a phase shift (phase difference) with the light source. Since phase difference information is useful in certain types of defects, it is advantageous to introduce a phase difference detector in one channel. The use of an interferometer facilitates phase difference detection. The interferometer can also detect differences in contrast on the sample. There are many types of interferometers. Mach-Zehnder, Mirau and Jamin-Lebedeff,
There are other devices that utilize the principle of twisting a light beam, and all can be used in the present invention. In addition, the observation can be performed in consideration of the gradient value of the phase change. For this, a difference method, that is, a differential or Nomarski-type contrast microscope can be used. Depending on the state of the sample, a change in the polarization state may occur particularly with the phase change of the observation light. This is also a source of information for searching for defects.

As an example, when the sample has a double refraction and it has a position dependence of the observation area, the transmitted light is:
It carries that information. Similarly, the sample may have polarization-sensitive reflection or scattering characteristics. At these times, the reflected probe light has its characteristics.
The variation (shift) of the polarization state in the probe light described above can be detected by ordinary equipment and easily provides a new information channel in the sample inspection process. Give me Whether to irradiate the sample with light from above or below can be determined from design factors in consideration of the incident angle. Confocal illumination can be used as probe light. At this time, the topography (surface shape) information on the sample can be observed, and a new information channel can be formed.

Another technique for searching for a defect using probe light will be described. Other techniques may not yet be discovered. One of them is pulsed irradiation light (selecting the pulse shape of the probe light,
That is, a pulse pattern is created by the on-off operation)
Is a technique for observing time-varying information using. The time-varying dynamic signal described above may be used to organize and edit detection information or multiplex (multiplex) information to further simplify the process of detecting defect information. By observing the time lag (time lag) seen in such a dynamic signal, the topography characteristic of the sample surface can be detected. Some commercialized cameras have multiple sensors housed in the same housing. A camera equipped with RGB (red, green, blue) color elements is an example, and has three independent CCDs.
(Charge-coupled device) in one camera housing.
With such a camera, the alignment of the optical axes of three independent CCD sensors can be performed automatically with a single adjustment. Each sensor is an individual sensor in such a camera and has an individual signal processing mechanism. In each embodiment of the present invention, it is absolutely necessary for the plurality of probe mechanisms to be positioned so as to aim at the same point of the sample. Naturally, it is absolutely necessary for the detection mechanism to perform the same alignment, and the image size at the time of image capture must be the same.

The various modes of operation of the present invention have been described above. The contents are typical setting examples and technical examples based on a model example in performing the continuous work of the inspection, and details thereof. If an expert peruses the present specification with reference to the drawings, variations that apply the present invention will immediately appear. Therefore, the claims as set forth herein should be considered to include all modifications, adaptations and variations. In short, the claims should be interpreted appropriately and broadly based on the true spirit and prospects of the technology based on this invention.

[0059]

As described above, according to the present invention, the first and second responses in the first pattern are detected, and the results are compared with the two responses from the same inspection point of the corresponding second pattern. A comparison process is performed between the two, and as a result, first and second difference signals of the response are formed. These individually created difference signals are subjected to data processing collectively,
A unified first pattern defect list is generated. First
Data processing is then performed on the pattern defect list.
Then, false defects found on the known and harmless sample surface are extracted and eliminated. In addition, various test searches can be added to increase the test response so that more than one optical response is obtained and processed from the sample. Thereby, the inspection accuracy is further improved. In addition, in the case of a transmissive sample, a photoelectric detector can be installed at the back of the sample to collect the inspection response depending on the transmitted light, and the accuracy of the pattern defect list can be further improved. It is also possible to search for a defect buried in.

[Brief description of the drawings]

FIG. 1 is a block diagram of an optical inspection system based on a single-channel response method in a prior inspection technique, which sequentially inspects a wafer for defects in a bright field and a dark field.

FIG. 2A is a graph showing a result of an inspection using a bright field in a prior inspection technique, in which a signal whose threshold value is determined and whose signal intensity exceeds the value is classified as a defect; FIG. 2B is a graph showing a result of an inspection using a bright field in the prior inspection technology, in which a threshold value is determined, and a signal having a signal intensity exceeding the value is classified as a defect. FIG. 2C is a graph showing a result of an inspection using a dome illumination system in the prior art, in which a threshold value is determined and a signal having a signal intensity exceeding the threshold value is classified as a defect.

FIG. 3 shows a result of a wafer inspection in the prior inspection technique, in which the horizontal axis represents the number of steps of signal data in a bright field and the vertical axis represents the number of steps of signal data in a dark field. A graph in which a value is determined and a signal having a signal strength exceeding the value is classified as a defect.

FIG. 4 is a block diagram of an optical inspection system in the prior inspection technique, which performs a defect inspection of a wafer in a bright field and a dark field by two signal processing systems (processor channels) independent of each other. .

FIG. 5 is a block diagram of an automatic optical inspection system according to the present invention, in which wafers are inspected by inputting signals to the same processor channel to integrate a bright field and a dark field, respectively.

FIG. 6 is a diagram showing in detail a defect detector in the automatic optical inspection system shown in FIG.

FIG. 7 is a graph in which the number of steps in a bright field is plotted on the horizontal axis, and the number of steps in the dark field is plotted on the vertical axis, showing the relationship between intrinsic defects, false defects and noise.

8 shows in the prior art by plotting the combination of FIG. 3 and FIG. 7 how the line segments indicating their bright-field and dark-field thresholds are far from actual; Graph.

FIG. 9 is a schematic diagram showing a first embodiment of the present invention, which simply illustrates the principle of using different illumination systems in a bright field and a dark field.

FIG. 10 is a schematic diagram showing a second embodiment of the present invention, which simply illustrates the principle of using a single same illumination system in a bright field and a dark field.

FIG. 11 is a third embodiment of the present invention,
9 is partially similar to FIG. 9 with two illumination systems and two detector subsystems in the bright field, and two in the dark field.
FIG. 1 is a schematic diagram of a system having one illumination system and two detector subsystems.

FIG. 12 is a fourth embodiment of the present invention,
FIG. 4 is a schematic diagram of a system in which light is transmitted through a transparent sample and different illumination systems are used in a bright field and a dark field.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 10 ... Bright field illuminator 12 ... Dark field illuminator 14 ... Sample wafer 16 ... Sensor 18 ... Input buffer 20 ... Delay circuit 22 ... Defect detector 24 ... Post-processing processor 26 ... Bright field defect list 28 ... Dark field defect list 30 ... Possible values of the signal 32... Possible values of the signal 34.. 44: pattern defect 46: dust particle defect 48: defect 50: grain lump 52: false defect due to process variation 54: system noise 56: false defect due to surface contrast variation 60: lens 62: lens 64: beam splitter 66 ... Beam splitter 68 ... Spatial filter 70 ... Relay lens 72 ... Condenser lens 73 ... Beam splitter 74 ... Night vision Image detector 76 ... laser 78 ... condenser lens 80 ... beam splitter 82 ... bright field image detector 90 ... filter 92 ... filter 94 ... subtractor 96 ... histogram circuit 98 ... defect determination algorithm

 ──────────────────────────────────────────────────続 き Continued on the front page (72) Inventor Bin-Min Benjamin Tsai, United States, 95070, California, Saratoga, Scotland Drive 19801 (72) Inventor Russell M. Pont United States, 95054, California, Santa Clara, number 313, Park View Drive 600

Claims (30)

[Claims] 1. A method of searching for a defect present in a first pattern on a sample by referring to reflection images of a second pattern formed by the same design on the same sample, which are known in dark and light and in both visual fields. A) illuminating the same point of the first pattern with a bright field light source and a dark field light source, respectively; b) detecting a dark field reflection image obtained by reflection from the first pattern; c) detecting a bright-field reflection image obtained by reflection from the first pattern; and d) converting the dark-field reflection image obtained in the step b) to the second field.
E) generating a difference signal between the two images at the same point of the pattern by comparing it with the dark-field reflection image; and e) applying the bright-field reflection image obtained in step c) to the second image.
Forming a difference signal between the two images by comparing with the bright-field reflection image of the pattern; and f) data processing the difference signal between the reflection images in the dark and light fields obtained in the steps d) and e). And integrally determining one defect list of the first pattern. 2. The automatic defect inspection method further comprises: g) adding post-processing to the data of the first pattern defect list obtained in the step f), finding harmless false defects in the list, and detecting the false defect data. 2. The automatic defect inspection method according to claim 1, further comprising the step of: 3. The automatic defect inspection method according to claim 1, wherein the illumination light sources in the step a) are constituted by independent dark-field illumination light sources and bright-field illumination light sources, respectively. 4. The automatic defect inspection method according to claim 3, wherein the light of the light source of each of the dark and light and the visual fields formed independently has a different frequency. 5. The dark-field light source is arranged to provide illumination of a narrow-band light, and the step b) includes the steps of: h) imperfections due to the light beam of the dark-field image passing through a spatial filter. 5. The method of claim 4, further comprising the step of enhancing search capabilities. 6. The sample is optically transparent and refers to a known dark, bright, bi-field transmission image of the second pattern, the method comprising: i) a dark field transmission image of the first pattern. J) detecting a bright-field transmission image of the first pattern; k) converting the dark-field transmission image obtained in step i) at the same point of the second pattern into a dark-field transmission image And 1) comparing the bright-field transmission image obtained in step j) with the bright-field transmission image at the same point of the second pattern to form a difference signal between the two images. Forming a difference signal. The step f) includes the steps d) and e in addition to the difference signals regarding the dark-field and bright-field transmission images obtained in the steps k) and l). ) For both dark-field and bright-field reflection images obtained in each case. 2. The automatic defect inspection method according to claim 1, further comprising a step of performing data processing on the differential signal to be determined and integrally determining a defect list of the first pattern. 7. The automatic defect inspection method includes the step of: m) performing post-processing on the data of the defect list of the first pattern obtained in step f), and searching for a false defect on the surface whose property is known. 7. The automatic defect inspection method according to claim 6, further comprising: 8. The method of claim 6, wherein step a) provides independent dark field and bright field illumination sources. 9. The automatic defect inspection method according to claim 8, wherein the independent dark and light and the illumination of each visual field are constituted by light sources having different frequencies. 10. The dark-field illumination light source is arranged to emit narrow-band light, and the step b) includes the steps of: n) searching for a defect by a light beam of a dark-field reflected image passing through a spatial filter; Enhancing the ability to search for defects, and o) enhancing the ability to search for defects by passing the light beam of the dark-field transmitted image through a spatial filter. Item 9. The automatic defect inspection method according to Item 9. 11. The second pattern on the sample is set as a reference image using a known reflection image based on a dark-field light source that emits a plurality of lights having different frequencies, and the step a) is different from each other. Illuminating the same observation point of the sample by setting a dark field based on a frequency multi-channel light source; step b) illuminating the same observation point of its first pattern on each of said multi-channels Step d) comprises comparing the multiple reflection images obtained in step b) with a reference image of a known second pattern, and detecting the two reflection images in the dark field. 2. The automatic defect inspection method according to claim 1, wherein a difference signal of the reflection image is generated. 12. The second pattern on the sample,
A known reflection image based on a bright field light source that emits a plurality of lights each having a different frequency is set as a reference image, and the step a) includes setting a bright field based on a multi-channel light source having different frequencies. Illuminating the same observation point of the sample, and said step c) separately detecting a reflection image obtained by irradiating the same observation point of the first pattern in each of said multiple channels; e) comparing the multiple reflection image obtained in step c) with the reference image of the known second pattern described above to generate a difference signal between the two reflection images in the bright field. 2. The automatic defect inspection method according to claim 1, wherein: 13. Automatically searching for defects present in a first pattern on a sample by referring to reflection images of a second pattern made by the same design on the same sample, which are known in dark and light and in both fields of view. A defect detection device, comprising: a dark field and a bright field illumination system for irradiating the same point of the first pattern; a dark field image detector for detecting a reflection image of the first pattern in the dark field; A bright-field image detector for detecting a reflection image of the first pattern in the visual field; and a dark-field difference signal which is combined with the dark-field image detector and compared with the reflection image of the second pattern in the dark field. A bright-field comparator coupled to the bright-field image detector and forming a bright-field difference signal by comparing the reflected light image of the second pattern in the bright field; , Combined with its respective comparator in each field of view, its dark and light, both differential signals And a data processor for unitarily forming a defect list of the first pattern by data processing. 14. The automatic defect inspection apparatus, coupled to the data processor, by post-processing the defect list of the first pattern to select only known and harmless false defects therein. 14. The automatic defect inspection apparatus according to claim 13, further comprising a post-processing processor for performing and eliminating the processing. 15. The automatic defect inspection system according to claim 13, wherein said dark field and bright field illumination system includes a dark field illumination subsystem and a bright field illumination subsystem. 16. The automatic defect inspection apparatus according to claim 15, wherein the dark-field and bright-field illumination subsystems emit light having different frequencies. 17. The dark-field and bright-field illumination subsystem provides narrow-band illumination light, and the dark-field image detector includes a spatial filter and passes dark-field reflected light to enhance defect detection capability. 17. The automatic defect inspection apparatus according to claim 16, wherein the apparatus is strengthened. 18. The automatic defect inspection apparatus according to claim 18, wherein the sample is transparent, and refers to a known transmission image in each field of view of the second pattern. A dark-field transmitted light detector for detecting a transmitted-field image, a bright-field transmitted light detector for detecting a bright-field transmitted image, and a first pattern and a A dark-field transmitted light signal comparator for comparing the dark-field transmitted light signals of the same observation point of the two patterns with each other to generate a difference signal; A bright-field transmitted light signal comparator for comparing the bright-field transmitted light signals of the two patterns with respect to the same observation point with each other to generate a difference signal; Combined with optical signal comparator, its dark and light, both Forming a defect list centrally first pattern divided signal and the data process automatically defect inspection apparatus according to claim 13, wherein. 19. The automatic defect inspection apparatus may further include a post-processing processor coupled to the data processor and receiving a first defect list to select and eliminate only data of known harmless false defects. 19. The automatic defect inspection apparatus according to claim 18, wherein: 20. The automatic defect inspection system of claim 18, wherein the darkfield and brightfield illumination systems include a darkfield illumination subsystem and a brightfield illumination subsystem. 21. The automatic defect inspection apparatus according to claim 20, wherein the dark-field and bright-field illumination subsystems emit light having different frequencies. 22. The dark-field illumination subsystem provides a narrow-band illumination light, and the dark-field reflection image detector has a first spatial filter to pass the dark-field reflection light to increase detection capability. 22. The automatic defect inspection apparatus according to claim 21, wherein the dark field transmitted image detector has a second spatial filter, and transmits the dark field transmitted light to enhance the detection capability. 23. The second pattern is arranged so as to refer to a plurality of known dark-field reflection images with respect to the same observation point using illumination systems having different frequencies. The frequency is set to irradiate the same observation point of the sample with a different illumination system, the dark-field image detector is configured with a plurality of separate detectors,
The dark field comparator is configured to respectively detect a first pattern dark field reflection image based on the multiple light sources, wherein the dark field comparator is coupled to each of the dark field image detectors and includes a plurality of the first pattern dark field reflection images. 14. The automatic defect inspection apparatus according to claim 13, wherein a dark field reflection image of each of the second pattern and the second pattern is compared to generate a difference signal therebetween. 24. For the same observation point with illumination systems having different frequencies, the second pattern is arranged so as to refer to a plurality of known bright-field reflection images, and the dark-field and bright-field illumination systems The bright field image detector is set to irradiate the same observation point of the sample with illumination systems having different frequencies, and the bright field image detector is configured by a plurality of separate detectors,
A bright field reflection image of the first pattern based on the multiple light sources, wherein the bright field comparator is coupled to each of the bright field image detectors and obtains the resulting plurality of bright field reflection images; 14. The automatic defect inspection apparatus according to claim 13, wherein each of the bright field reflection images of the second patterns is compared to generate a difference signal therebetween. 25. The dark-field and bright-field illumination system receives a single laser illumination light source, and illumination light traveling in one direction from the laser light source,
A beam splitter installed to reflect it, and a condenser lens for guiding the light beam to an observation point on the sample, wherein the dark-field image detector has a low elevation angle to receive the reflected light from the sample. The bright-field photodetector is installed facing the observation point so as to receive a bright-field image that is incident via the condenser lens and the beam splitter after being reflected from the sample. 14. The automatic defect inspection apparatus according to claim 13, wherein: 26. A narrow-band laser light source comprising: a dark-field and a bright-field illumination system comprising a narrow-band laser having a constant frequency; A mercury arc lamp; a first condenser lens for passing light rays from the mercury arc lamp; a first beam splitter installed to reflect light rays passing through the first condenser lens in one direction; A dark field illumination and a second condenser lens for directing a bright field light beam from the first beam splitter to a common observation point, wherein the dark field image detector transmits reflected light from the sample to the second condenser lens and the first condenser lens. A second beam splitter, arranged to be received via a beam splitter, which allows all other light to pass through, especially by applying a dichroic dichroic film In order to reflect only the dark-field illumination light emitted from the laser, the arrangement angle of the second beam is adjusted with the positional relationship between the second condenser lens and the first beam splitter. A splitter; a third lens having a function of focusing light of a dark-field image reflected from the second beam splitter; a dark-field detector for detecting the dark-field reflected image; Condenser lens, rays coaxial with the first and second beam splitters, but behind the second beam splitter and carrying the remaining bright-field reflection image, i.e. rays passing through the second beam splitter , A fourth lens disposed to receive and narrow the light field, and a bright field photodetector disposed behind the fourth lens for receiving the bright field reflected image and optically coupled thereto. Claim 1 Automatically defect inspection apparatus according. 27. The sample is transparent, and when transmitting light to the sample, the automatic defect inspection apparatus is installed at a low elevation angle with respect to the sample surface in order to receive a dark-field transmission image from the sample. 26. The automatic defect inspection apparatus according to claim 25, further comprising: a dark-field transmitted light detector; and a bright-field transmitted light detector disposed on a rear surface of the sample to receive a bright-field transmitted image. 28. The sample, wherein the sample is transparent and arranged to transmit light, the automatic defect inspection device comprises a fifth condenser lens arranged behind the sample to enlarge transmitted light. A dark field transmitted light image detector, and a bright field transmitted light image detector, wherein the dark field transmitted light image detector is a third beam splitter installed behind a fifth condenser lens. The reflecting surface is constituted by a dichroic dichroic film which receives the transmitted light from the sample and is set to selectively reflect only the dark-field transmitted light, while being arranged so as to allow all other light beams to pass. The second beam splitter is configured to reflect the dark-field transmitted light in a different direction from the optical path defined by the fifth condenser lens, and the third beam splitter; Reflection from 3-beam splitter A sixth lens set so as to narrow the dark-field transmitted light to focus on the dark-field transmitted light, and a dark-field transmitted light detector provided to receive the dark-field transmitted image; The seventh lens is a seventh lens, which is located behind the third beam splitter and is aligned with the fifth condenser lens so that all the remaining light beams, ie, the third beam splitter, A seventh lens disposed to narrow down the received bright-field transmitted light, and a bright-field transmitted light detector set immediately behind the seventh lens and receiving a bright-field transmitted image from a sample. The automatic defect inspection apparatus according to claim 26, wherein: 29. A method of comparing a second pattern on the same sample made with the same design in searching for a defect existing in the first pattern on the sample, wherein the second pattern comprises: An automatic defect inspection method having known first and second responses in at least one search comprising: a) performing an inspection search on at least two corresponding identical points of a first pattern of a sample and generating at least two responses. Obtaining; b) processing said first response of said first pattern with a detector; c) processing said second response of said first pattern with a detector; d) said step b) comparing the first response with the first response at the same point in the second pattern to generate a difference signal between them; e) the second response and the second pulse from step c). F) comparing the second response at the same point in the first response to generate a difference signal between them, and f) data processing two of the difference signal in the first response and the difference signal of the second response. And determining the first pattern defect list in a centralized manner. 30. For the purpose of searching for a defect present in a first pattern on a sample, the same observation points of a second pattern on the same sample made by the same design are compared, and the second pattern is determined at least once. First and second known in the search for
At least one image search mechanism for observing and searching for an observation point of a first pattern on the sample to obtain at least two responses, and detecting the first response. A first response detector, a second response detector for detecting a second response, and a first response detector, wherein the output of the first response detector is compared with the output of the first response in the second pattern. A first comparator for generating a first differential signal of the response, and a second response detector for comparing the output of the second response detector with the output of the second response in the second pattern. A second comparator for generating a second differential signal in response, the second comparator being coupled to the first and second comparators for processing the first and second differential signals to generate one defect list of the first pattern And a processor for centrally determining Automatically defect inspection apparatus.