JP4543141B2 - Defect inspection equipment - Google Patents

Description

[0001]
BACKGROUND OF THE INVENTION
The present invention relates to a defect inspection apparatus, and more particularly to a defect inspection apparatus capable of detecting the presence of fine defects of 0.1 μm or less and the characteristics of defects.
[0002]
[Prior art]
The surface state of photomask blanks, semiconductor wafers, magnetic disks, optical disks, etc. is extremely important in achieving these essential functions. For example, if there is a defect due to the presence of scratches or foreign matter on the surface of the glass substrate of a photomask used when projecting a pattern on a semiconductor wafer wafer, the defect is projected together with the pattern, and an accurate pattern is formed on the semiconductor wafer. Cannot be projected. In particular, along with the miniaturization of patterns, the defects to be detected are also miniaturized. Therefore, there is a strong demand for the development of a defect inspection apparatus that can accurately detect the presence of fine defects present on the surface of semiconductor wafers, photomasks, etc. ing.
[0003]
As a defect detection device for photomask glass substrates, the glass substrate is irradiated with laser light, scattered light from foreign matter on the substrate surface is received by a photodetector, and the presence of defects based on the output signal from the photodetector A laser scattering method for detecting is known. In this laser scattered light type defect detection apparatus, the presence of a defect is determined by detecting scattered light from a defect portion due to scratches or foreign matter.
[0004]
[Problems to be solved by the invention]
The above-described laser scattered light type defect detection apparatus has a high detection sensitivity and can detect the presence of defects relatively accurately when foreign matter is present on the substrate surface.
However, since the size of foreign matter that can be detected is limited due to the use of scattering, sensitivity to minute foreign matter is extremely low, and defects due to the presence of a thin film have extremely low detection sensitivity. In addition, there is a drawback that the detection sensitivity of a concave portion such as a scratch is also extremely low.
[0005]
Further, the scattered light type defect detection apparatus has a defect that the use range of the defect detection information is limited because it is inaccurate to determine the defect characteristics, for example, whether it is a concave defect or a convex defect. . That is, if the characteristics of the detected defect can be determined, an appropriate process to be performed thereafter can be determined, and the surface state detection device can be used more effectively.
[0006]
Accordingly, an object of the present invention is to provide a defect inspection apparatus that can accurately detect even a fine defect of, for example, 0.1 μm or less.
[0007]
Furthermore, an object of the present invention is to provide a defect inspection apparatus that can accurately detect both relatively large defects and minute defects.
[0008]
It is another object of the present invention to provide a defect inspection method and apparatus capable of determining not only the detection of a defect but also the characteristics of the detected defect, particularly whether the shape of the defect is concave or convex.
[0009]
Furthermore, the objective of this invention is providing the defect inspection apparatus which can detect the state of a sample surface at high speed.
[0010]
Furthermore, an object of the present invention is to provide a defect inspection method capable of selecting an optimum process to be executed according to the characteristics of a detected defect.
[0011]
  The present inventionAs a reference exampleThe defect inspection apparatus includes a light source device that generates a light beam,
  A beam deflecting device for deflecting the light beam in a first direction;
  An objective lens that focuses an incident light beam into a spot shape and projects it toward a sample to be inspected for defects;
  A sample stage that supports the sample to be inspected for defects;
  A light beam that is disposed between the beam deflecting device and the objective lens, converts the light beams emitted from the beam deflecting device into first and second sub-beams having mutual coherence, and reflects each other on the sample surface to be inspected. An interference optical system for generating an interference beam including phase information including height information about the sample surface by combining related sub-beams;
  A photodetector for receiving an interference beam emitted from the interference optical system;
  And a signal processing circuit for generating a defect detection signal indicating the presence of a defect on the sample surface based on an output signal from the photodetector.
[0012]
In the present invention, the light beam emitted from the light source device is converted into two sub-beams having mutual coherence by the interference optical system. Then, the sample surface to be inspected for defects is scanned by these sub beams. When there is a defect on the sample surface, one of the two sub-beams scans the defective part, and the other sub-beam scans a normal part adjacent to the defect. A phase difference corresponding to the magnitude occurs. The amplitude I of the interference beam obtained by combining the sub-beams reflected on the sample surface by the interference optical system is expressed by the following equation.
I = A₁ ²+ A₂ ²+ 2A₁ A₂ cosθ
Where A₁ And A₂ Represents the amplitude of the sub-beam, and θ represents the phase difference between the sub-beams. As is clear from the above equation, when the phase difference between the two sub beams changes from 0 to π (corresponding to a quarter wavelength), the amplitude of the interference beam becomes (A₁ + A₂ ) To 0. Accordingly, when the wavelength of the light beam is 488 nm, for example, the amplitude of the interference beam, that is, the luminance information changes from the normal value to almost zero as the defect height changes from 0 nm to 122 nm. A defect detection apparatus having sensitivity can be realized. Moreover, if the sample surface is scanned two-dimensionally with a plurality of light beams, defect detection can be performed at high speed.
[0013]
The signal processing circuit for determining the defect detection can detect the defect by comparing the output signals from each light receiving element of the linear image sensor, or the output signal from each light receiving element passes through the high-pass filter. Then, the defect can be detected by comparing with a reference value using a comparator, and various defect detection methods can be used.
[0014]
In a preferred embodiment of the defect detection apparatus according to the present invention, when m is a positive integer, a phase difference of approximately (2m + 1) π / 2 is introduced between the first sub-beam and the second sub-beam. It is characterized by being. If a phase difference of (2m + 1) π / 2 is introduced between the two sub beams, the light quantity of the interference beam changes greatly in response to a slight phase change, so that even a minute defect can be detected accurately.
[0015]
A preferred embodiment of the defect inspection according to the invention is characterized in that the first and second sub-beams are separated from each other in the beam deflection direction of the beam deflection device on the sample. If the distance between the sub-beams is set larger than the defect to be detected, the second sub-beam scans a normal part while the first sub-beam scans the defect site. As a result, a phase difference corresponding to the height of the defect occurs between the two sub beams. This phase difference acts to delay the phase for the second sub-beam on the rear side when acting in the direction of phase advance with respect to the first sub-beam on the leading side. Therefore, the luminance generated when each sub beam scans the defective portion is different from each other.
Therefore, it is possible to determine whether the defect is a convex defect or a concave defect by comparing the output intensities from the light receiving elements that occur when scanning the defective portion. As a result, by scanning in the interference inspection mode, it is possible to determine whether the defect is a concave defect or a convex defect in one scan, and by using this inspection result, a process to be performed next, for example, a sample It is possible to immediately determine whether a cleaning process or a polishing process should be performed. That is, most of the convex defects can be removed by the cleaning process due to the adhesion of foreign matter, and many of the concave defects are caused by the scratches, and the defects can be removed by the polishing process. Therefore, the detected defect is convex. Since it can be determined that the shape is concave, the next processing to be performed can be immediately executed.
[0016]
  The defect inspection apparatus according to the present invention includes a confocal inspection mode for scanning a sample surface with a plurality of light beams and detecting defects present on the sample surface based on luminance information of reflected light from the sample, and a height of the sample surface. A defect inspection apparatus capable of switching between interference inspection modes for detecting information as phase information,
  A light source device that emits a plurality of light beams aligned along a first direction;
  A beam deflecting device for deflecting the plurality of light beams in a second direction orthogonal to the first direction;
  An objective lens that focuses a plurality of incident light beams into a spot shape and projects it onto the sample surface;
  A sample stage that supports the sample to be inspected;
  In the interference inspection mode,In the optical path between the beam deflector and the objective lensInsertThe incident light beam is converted into first and second sub-beams having coherence with each other, and the sub-beams reflected on the sample surface and related to each other are synthesized to obtain phase information on the height of the sample surface. An interference optical system for generating an interference beam including:
  A plurality of photodetectors arranged in a direction corresponding to the first direction and receiving reflected light or interference beams from the sample surface;linearAn image sensor;
  AbovelinearA signal processing circuit that generates a defect detection signal based on an output signal from an image sensor, the signal processing circuit,
  In the confocal inspection mode,linearDefect detection means for detecting a defect using a luminance signal output from the light detection element of the image sensor;
  In the interference inspection mode,linearAnd a means for discriminating whether the defect is a convex defect or a concave defect based on a change in the amplitude of the interference beam output from the light detection element of the image sensor. When defects are detected using an interference beam, there is an advantage that minute defects can be detected accurately, but a phase wrapping problem arises. That is, when the size of the defect in the optical axis direction corresponds to a phase difference of π or more, the defect height information cannot be accurately determined. On the other hand, in the confocal inspection mode, even a defect having a size of 2π or more can be accurately detected. Therefore, according to the present invention, the confocal inspection mode and the interference inspection mode can be selectively switched. Accordingly, defect detection information is first stored by performing defect inspection in the confocal inspection mode. Next, the inspection mode is switched, defect inspection is performed in the interference inspection mode, and defect inspection is performed for minute defects. If comprised in this way, the defect of both a comparatively large defect and a comparatively small defect below 2 (pi) can be test | inspected correctly.
[0017]
  The present inventionAs a reference exampleThe defect inspection apparatus includes a light source device that generates a light beam,
  A beam deflecting device for scanning the light beam;
  A sample stage that supports the sample to be inspected for defects;
  An interference optical system that converts the light beam into two sub-beams spaced apart from each other and generates an interference beam including information on the sample surface as phase information by combining the sub-beams reflected by the sample surface to be inspected;
  Means for introducing a relative phase difference varying between at least 0 and 2π between the two sub-beams;
  An objective lens that focuses the sub-beam into a fine spot and projects it onto the sample surface to be inspected;
  A photodetector that receives the interference beam and converts it into an electrical signal;
  And a signal processing circuit for detecting height information of defects existing on the sample surface based on an output signal from the photodetector. In this example, by using the phase shift method, not only the height value of the defect can be obtained, but also whether the shape of the defect is convex or concave can be accurately determined.
[0018]
In a preferred embodiment of the defect inspection apparatus according to the present invention, the interference optical system is constituted by a prism element that converts an incident light beam into two sub beams and emits it, and controls the amount of insertion of the prism element into the optical path. An actuator is connected, and a relative phase difference is introduced between the sub beams according to the amount of insertion of the prism element into the optical path.
Here, a prism element such as a Nomarski prism, a Wollaston prism, or a Rochon prism can be used as a prism element that converts a light beam into two sub beams.
[0019]
DETAILED DESCRIPTION OF THE INVENTION
In describing a defect inspection apparatus according to the present invention, a confocal inspection is performed in which a sample surface is scanned with a plurality of (m) light beams focused in a minute spot shape, and defects existing on the sample surface are detected from luminance information of the reflected light. Between the mode and the interference inspection mode in which the sample surface is scanned with two focused light beams having coherence and the reflected light from the sample surface is synthesized to detect the height information of the sample surface as phase information. A defect inspection apparatus that can be switched will be described. This switchable defect inspection apparatus is beneficial to inspect defects in the confocal inspection mode when the purpose is to detect relatively large defects, and minute defects of about several tens of nanometers are inspected in the interference inspection mode. It is beneficial to do. Therefore, the mode can be switched according to the size of the defect to be detected. For example, the inspection can be performed first in the confocal inspection mode, and the detected minute defect can be inspected in the interference mode.
[0020]
First, the configuration of the optical system when performing defect inspection in the confocal inspection mode will be described. FIG. 1 is a diagram showing a configuration of an example of an optical system in a confocal inspection mode of a defect inspection apparatus according to the present invention. For example, an Ar laser that emits laser light having a wavelength of 488 nm is used as the light source 1. As other light sources, for example, a fourth harmonic YAG laser that generates laser light with a wavelength of 266 nm or an Ar laser that emits light with a wavelength of 264 nm can be used, or a semiconductor laser with a relatively long coherent length can also be used. It can also be used.
[0021]
The light beam generated from the light source 1 is incident on the diffraction grating 2 and converted into a plurality of light beams aligned in the first direction. In order to clarify the drawing, it is assumed that the diffracted light is diffracted in the paper. These light beams enter the polarization beam splitter 5 through the first and second relay lenses 3 and 4, pass through the polarization beam splitter 5, and enter the galvanometer mirror 6. The galvanometer mirror 6 periodically deflects a plurality of incident light beams at a predetermined frequency in a second direction (direction orthogonal to the paper surface) orthogonal to the first direction. The plurality of light beams reflected by the galvano mirror 6 enter the objective lens 10 through the third and fourth relay lenses 7 and 8 and the quarter wavelength plate 9. These light beams are focused into a minute spot by the objective lens 10 and enter the sample 11 to be inspected for defects. As a result, a plurality of light spots arranged at a predetermined pitch along the direction corresponding to the first direction are formed on the surface of the sample 11. The objective lens 10 is supported so as to be movable along the optical axis direction. Therefore, the relative distance between the sample surface and the focal point of the objective lens is changed by moving the objective lens 10 along the optical axis direction. Can do.
[0022]
The sample 11 may be a semiconductor wafer blank, a patterned semiconductor wafer, a photomask substrate, a patterned photomask optical disk substrate, a magneto-optical disk substrate, or the like. The sample 11 is placed on the sample stage 12. The sample stage 12 is movable in X (direction corresponding to the first direction) and Y (direction corresponding to the second direction orthogonal to the first direction) orthogonal to the optical axis. It is possible to move not only in the X and Y directions but also in the Z direction (optical axis direction), and the relative distance between the focal point of the objective lens and the sample can be changed by movement in the Z direction. With this configuration, the entire sample surface is scanned two-dimensionally with a light beam focused in a fine spot shape. The interval between the light spot rows formed on the sample surface can be set to a desired pitch by configuring any relay lens with a zoom lens and adjusting the magnification. Accordingly, when the sample 11 is a patterned wafer having a pattern formed at a predetermined pitch, for example, the magnification of the zoom lens is adjusted so that the pattern pitch matches the pitch of the light beam. It can be scanned with a light beam.
[0023]
Reflected light from the sample surface is collected by the objective lens 10, enters the galvano mirror 6 again through the quarter-wave plate 9 and the relay lenses 8 and 7, is descanned by the galvano mirror, and enters the beam splitter 5. To do. Since the incident light beam has passed through the quarter-wave plate twice, its polarization plane is rotated by 90 °.
As a result, the incident light beam is reflected by the polarization plane of the beam splitter and enters the linear image sensor 14 via the relay lens 13. The linear image sensor 14 has a plurality of light receiving elements arranged in a direction corresponding to the first direction, and each reflected beam is incident on a corresponding light receiving element. Since the incident light beam is descanned by the galvanometer mirror 6, it remains stationary on the linear image sensor 14 even if it is scanned by the galvanometer mirror.
[0024]
Since the frame defining the light receiving surface of each light receiving element of the linear image sensor 14 performs the same function as a pinhole, a pinhole having a minute opening is arranged on the front surface of each light receiving element of the linear image sensor. Is equivalent to Accordingly, the reflected light from the minute light spot that scans the sample surface is incident on the light receiving element defined by the pinhole, and as a result, the optical system shown in FIG. 1 constitutes a confocal optical system. Since the confocal optical system has a high resolution, by using the optical system of FIG. 1, it is possible to accurately detect a micro defect having a size of about 50 nm, for example.
[0025]
A CCD drive circuit 15 is connected to the linear image sensor 14 to read out the electric charge accumulated in each light receiving element for each line based on the supplied synchronization signal. Each read signal is supplied to the signal processing circuit 17 as a video signal through an amplifier 16 (only 16a to 16e are shown in the drawing). The signal processing circuit 17 processes the output signal from the linear image sensor 14 to detect a defect, and generates a defect detection signal when the presence of a defect is detected. In addition, each light receiving element of the linear image sensor can be constituted by a photodiode, and an output signal from each photodiode can be output continuously in time.
[0026]
The central processing unit 18 supplies a synchronization signal to the stage driving circuit 19, the mirror driving circuit 20 connected to the galvanometer mirror, and the CCD driving circuit 15, and the central processing unit 18 drives the entire apparatus.
[0027]
Next, defect detection in the signal processing circuit 17 will be described. 2A to 2C are circuit diagrams showing the configuration of a defect detection processing circuit. In the processing circuit shown in FIG. 2A, an output signal from each light receiving element is converted into an electric signal by the photoelectric conversion circuit 20, and is passed through a high-pass filter 21 to extract a high-frequency component. The extracted high frequency component is compared with a reference value in the comparator 22. The output signal from the light receiving element is a constant intensity output signal when scanning a normal surface part without defects, but when the defective part is scanned with a light beam, the amount of light incident on the light receiving element is significantly reduced. To do. Accordingly, it is possible to determine the presence of a defect by passing the photoelectric output from the light receiving element through a high-pass filter and comparing the output with a reference value by a comparator. Then, the address of the detected defect is stored in the coordinate memory. As a result, the existence of a defect and its address can be obtained.
[0028]
In the defect detection circuit shown in FIG. 2B, the presence of a defect is determined by comparing output signals between light receiving elements. For example, if the output signals of adjacent light beams or every other light beam or a plurality of light beams are compared, and there is a defect in the position scanned by one of the light beams, the output signal from the differential amplifier 24 Changes significantly. The presence of a defect can be detected by comparing this signal change with the reference value by the comparator 22.
[0029]
In the defect detection circuit shown in FIG. 2C, an output signal from each light receiving element is photoelectrically converted and then passed through a delay circuit 26 to form a signal delayed by one line, which is delayed by the differential amplifier 24 with the delayed signal. Compare with the signal that is not. In this case, if there is a defect in one of the scanning portions, a large signal component change occurs, and the presence of the defect is detected by comparing the signal change with a reference value by a comparator. Then, the address of the detected defect is stored in the coordinate memory.
[0030]
Next, the defect detection determination in the signal processing circuit for a photomask or wafer with a pattern will be described. In the case of a photomask or wafer on which a pattern has already been formed, a plurality of dies or chips having the same shape and the same pattern are formed on one mask or wafer. Therefore, when performing defect detection determination for a photomask with such a pattern, the detection signal or video signal for one die or chip is stored in the memory, and the defect detection signal is compared with the information stored in the memory. Defect detection determination can be performed by die comparison. As described above, the defect detection of the patterned photomask can be accurately performed at high speed by utilizing the defect inspection by the multi-beam and the die comparison.
[0031]
Next, the property or state detection of the detected defect will be described. Since the confocal optical system has a high resolving power in the height direction of the sample, when the focal point of the objective lens is focused on the sample surface, a large amount of light is incident on the light receiving element and the focal point is from the sample surface. If they are shifted, the amount of light incident on the light receiving element is significantly reduced. Therefore, beam scanning is performed while changing the relative distance between the focal point of the objective lens and the sample surface, and the relationship between the relative displacement between the focal point of the objective lens and the sample surface and the luminance that is the output of the light receiving element is obtained. By determining, it is possible to determine defect height information, that is, whether the defect is a convex defect or a concave defect, and the size of the defect can be measured.
[0032]
Using the detected defect address, the sample stage is moved to locate the detected defect at the beam scanning position. Next, beam scanning is performed multiple times while changing the relative distance between the focal point of the objective lens and the sample surface. At this time, the relative displacement amount between the focal point and the sample can be changed by moving the objective lens along the optical axis direction, or the relative displacement amount is also controlled by moving the sample stage in the Z direction. be able to. Then, the luminance is stored for each beam scan, and the relationship between the relative displacement amount and the luminance is obtained. FIG. 3 is a schematic diagram for defining the relationship between the relative displacement amount and the luminance. In FIG. 3, the alternate long and short dash line a indicates the relationship between the relative displacement amount and the luminance on the normal sample surface, the solid line b indicates the relationship in the convex defect portion, and the broken line indicates the relationship in the concave defect (for example, a scratch). Show. When the sample surface is normal, a luminance peak occurs at the reference focal position. On the other hand, in a portion where a convex defect (for example, the presence of a foreign substance) exists, a luminance peak occurs above the reference focal position (in the direction approaching the light source from the sample surface), and the concave defect In this case, the peak occurs downward (in the direction away from the light source from the sample surface). Therefore, the size in the height direction of the defect and whether it is convex or concave can be determined from the peak generation position in the relationship between the relative displacement amount between the focal point and the sample and the luminance. Further, since the light beam scans the sample surface two-dimensionally, the size of the defect in the XY plane can be obtained from the relationship between the relative displacement amount and the brightness between the focal point and the sample as two-dimensional information on the sample surface. .
[0033]
Based on the defect detection information detected in this way, the optimum process to be performed next can be determined. That is, when many of the detected defects are convex defects, the convex defects are mostly due to the adhesion of foreign matter. Therefore, by supplying the sample after the inspection process to the cleaning process and cleaning the sample surface Defects can be removed. On the other hand, when the detected defect is a concave defect, since this concave defect is mostly due to scratches on the sample surface, the sample that has finished the inspection process is supplied to the polishing process step, and the sample surface is polished. Thus, defects can be removed.
If comprised in this way, a defect inspection and the following process process can be performed automatically continuously.
[0034]
When the sample to be defect-inspected is a photomask having a plurality of dies each having the same pattern, the output information from the linear image sensor for one die is stored in the memory and stored in this memory. It is also beneficial to perform defect inspection using a die comparison method that detects defects by comparing the detected information with detected output information.
[0035]
Next, the interference inspection mode will be described. FIG. 4 is a diagram showing a configuration of an example of an optical system in the interference inspection mode. This optical system is the same as that shown in FIG. 1 except that an interference optical system is detachably disposed in the optical path between the objective lens 10 and the quarter-wave plate 9 via mode switching means (not shown). It is the same as the optical system shown in FIG. Therefore, only the configuration related to the interference optical system will be described.
[0036]
First, the interference optical system is inserted into the optical path using the mode switching means. The interference optical system includes a Nomarski prism 30 disposed at a converging point of the fourth relay lens 8, and the first and second sub-beams (the first and second sub-beams) that interfere with each other of the light beams incident by the Nomarski prism. (Not shown). These sub beams are separated by a predetermined distance in a second direction (a direction perpendicular to the paper surface) which is the deflection direction of the galvanometer mirror 6. Accordingly, the first sub beam precedes in the beam scanning direction, and the second sub beam follows the first sub beam separated by a predetermined distance. Therefore, the sample surface is first scanned with the first sub-beam, and then the same position is scanned with the second sub-beam. The separation distance between the sub-beams is set in consideration of the size of the defect to be detected, and can be set to 2 μm, for example, by being separated by a distance longer than the size of the defect to be detected. Note that the distance between the sub beams can be set to a desired distance by changing the design value of the Nomarski prism 30. An actuator 31 and a position detector 32 are connected to the Nomarski prism 30 and a drive circuit 33 is connected to the actuator 31. The actuator 31 introduces a phase difference between the two sub beams by moving the Nomarski prism 30 at a constant speed in a direction orthogonal to the optical path by a control signal from the signal processing circuit. Although the Nomarski prism 30 is shown to move left and right in the drawing, the Nomarski prism 30 is actually moved in a direction perpendicular to the paper surface to introduce a phase difference of at least 2π between the two sub beams.
[0037]
In the interference inspection mode, the sample surface is two-dimensionally scanned with interference sub-beams (2 m in number corresponding to a plurality of light beams generated from the light source device). At this time, the phase difference between the sub beams generated from the Nomarski prism 30 is set to zero or 2nπ (n is a positive integer). Each sub beam is focused into a minute spot by the objective lens 10 and projected onto the surface of the sample 11. Therefore, two m light spot rows are formed on the sample surface in the beam scanning direction. The reflected light from each minute spot is collected by the objective lens, and is synthesized for each related sub beam by the Nomarski prism 30 to form one interference beam. Each interference beam travels along the optical path, enters each corresponding light receiving element of the linear image sensor 14, is converted into an electric signal, and is supplied to the signal processing circuit 17.
[0038]
When a defect exists on the sample surface, first, the first sub-beam scans over the defect, and the second sub-beam scans the normal part. Accordingly, when one sub-beam scans the defective portion and the other sub-beam scans the normal portion, a phase difference is generated between the sub-beams depending on the height of the defect in the optical axis direction. When synthesized, the outgoing interference beam includes information corresponding to the phase difference generated between the two sub-beams, that is, information in the height direction of the defect, and the amplitude I of the outgoing interference beam is given by the following equation.
I = A₁ ²+ A₂ ²+ 2A₁ A₂ cos (± 2d / λ)
Here, λ represents the wavelength of the illumination beam to be used, d represents the height of the defect in the optical axis direction, and A₁ And A₂ Indicates the amplitude of the sub-beam. As is clear from this equation, when a phase difference of 0 to π occurs between the sub beams due to the defect, the amplitude I of the interference beam emitted from the Nomarski prism 30 is from 0 to A.² (A₁ = A₂ Will change). Accordingly, when the wavelength of the light beam is 488 nm, if there is a concave or convex defect having a height of 122 nm corresponding to a quarter wavelength, the amplitude of the interference beam becomes almost zero, and the light beam having a wavelength of 266 nm. When there is a defect having a height of about 66.5 nm, the amplitude of the interference beam becomes almost zero. Therefore, the interference inspection mode has extremely high sensitivity for detecting a minute defect, and is useful for detecting a minute defect smaller than a defect that can be detected by the confocal inspection mode.
[0039]
In detecting the defect, the entire surface of the sample is scanned using a multi-beam, and each interference beam emitted from the Nomarski prism 30 is received by each light receiving element of the linear image sensor 14. Then, the presence of a defect and its address are detected using the defect detection method shown in FIG.
[0040]
Next, detection of defect height information in the interference inspection mode will be described. First, the presence of a defect and its address are detected by the confocal inspection mode or the interference inspection mode. Next, the sample stage is moved so that the site where the defect exists is scanned. Next, when the entire inspection is performed in the confocal inspection mode, the mode is switched, and the interference optical system is inserted into the optical path. Then, similarly to the confocal inspection mode, beam scanning is performed a plurality of times for the defective portion. In this beam scanning, by changing the amount of insertion of the Nomarski prism 30 into the optical path, the phase difference introduced between the two sub-beams is changed at least from 0 to 2π, and is scanned a plurality of times and detected by the light receiving element. The amplitude of the interference beam is stored in the memory for each scan.
[0041]
Here, the relationship between the output signal intensity I of the light receiving element and the introduced phase difference Δθ introduced by the Nomarski prism 30 can be expressed by the following equation.
I = A₁ ²+ A₂ ²+ 2A₁ A₂ cos (Δθ ± 2d / λ)
Here, ± 2d / λ is defined by whether the defect is concave or convex, and + indicates a concave defect, and − indicates a convex defect. As an example, a case where a defect having a depth of λ / 4 exists will be described. FIG. 5 schematically shows the relationship between the amplitude of the output signal from the light receiving element and the scan position. In FIG. 5, the solid line a indicates the output signal of the light receiving element when the introduction phase difference Δθ is 0 or 2nπ, and the alternate long and short dash line b indicates the output signal of the light receiving element when the introduction phase difference Δθ is (2m + 1) π / 2. The broken line c indicates the output signal of the light receiving element when the introduction phase difference Δθ is π, and the two-dot chain line d indicates the output signal of the light receiving element when the introduction phase difference Δθ is (2m + 1) π / 2. First, a case where no phase difference is introduced between the Nomarski prism and the sub beams will be described. Immediately after the start of scanning, both sub-beams scan the normal part, so that the output signal intensity of the light receiving element becomes the maximum value. Next, the head side sub is positioned on the defect and the rear side sub beam is positioned on the normal part, and a phase difference corresponding to the height of the defect is generated between the sub beams. As scanning progresses further, the leading sub-beam passes through the defective part and both sub-beams scan the normal part. During this period, no phase difference occurs between the sub-beams, and the signal intensity of the interference beam is the original maximum. Return to. As the scanning further proceeds, the rear sub-beam scans the defective portion, the leading sub-beam is positioned in a normal portion, and a phase difference corresponding to the height of the defect occurs between the sub-beams. In this way, the output intensity of the light receiving element decreases by two due to the presence of one defect. Next, when a phase difference is introduced between the sub beams by the Nomarski prism, the output intensity of the light receiving element is inverted according to the introduced phase difference, and the state shifts to the states of curves b, c, and d.
[0042]
FIG. 6 is a graph showing the relationship between the phase difference introduced by the Nomarski prism 30 in the normal part and the defective part and the output signal intensity of the light receiving element, that is, the amplitude of the interference beam. In FIG. 6, a solid line a indicates the relationship between the introduced phase difference of the normal part and the output intensity of the light receiving element, and a broken line b indicates the difference between the introduced phase difference of the concave defect part having a depth of λ / 8 and the output intensity of the light receiving element. An alternate long and short dash line c indicates the relationship between the phase difference introduced by the convex defect portion having a height of λ / 8 and the output intensity of the light receiving element. The output intensity of the light receiving element periodically changes according to the amount of phase difference introduced. Then, when the amplitude of the interference beam of the normal part having no defect is used as a reference, the amplitude change of the defective part has a phase difference caused by the defect of λ / 8 in addition to the introduced phase difference. As a result, the peak of the periodic amplitude change is displaced left and right on the upper side of the drawing depending on whether it is concave or convex. Therefore, by measuring the phase displacement amount at the peak position with reference to the normal part, it is possible to determine the height value of the defect and whether it is a concave defect or a convex defect.
[0043]
  Next, a method for determining the presence of a defect when scanning the entire sample surface in the interference inspection mode and a method for determining whether the defect is convex or concave will be described. In this example, the phase of the leading sub-beam is changed by the Nomarski prism 30 with respect to the rear sub-beam.π / 2The sample surface is set to advance by (90 °) and the sample surface is scanned two-dimensionally. This phase difference state corresponds to an intermediate position during the transition from the lower peak to the upper peak of the periodic change in the output intensity of the light receiving element with respect to the phase difference of FIG. Therefore, the output intensity of the light receiving element changes greatly according to a slight phase change. As an example, it is assumed that a convex defect corresponding to a phase of 10 ° exists. In this case, the defect acts to further advance the phase by 10 ° with respect to the leading sub-beam. As a result, during the period when the leading sub-beam scans the defective portion, a total phase difference of 100 ° is generated between the leading sub-beam and the rear sub-beam. Therefore, the amplitude of the combined interference beam, that is, the output intensity of the light receiving element is further lower than the output intensity when both sub beams scan the normal part. On the other hand, when the scanning progresses and the rear sub-beam scans the defective portion and the leading sub-beam scans the normal portion, the defect is a rear sub-beam having a phase delay of 90 ° with respect to the leading sub-beam. Acts to advance the phase by 10 °. As a result, the total phase difference between the front sub-beam and the rear sub-beam is 80 °, and the output intensity of the light receiving element is higher than the output intensity when both sub-beams scan the normal part. However, since the optical system of this example is a confocal optical system, the amount of light incident on the light receiving element itself is reduced due to the presence of defects, so the actually detected output intensity is when both sub-beams scan the normal part. The output is slightly lower than the output intensity. Therefore, when scanning a convex defect, the light receiving element is convex if a signal whose output intensity is greatly reduced and a signal that is slightly reduced are output after a time corresponding to the scanning time delay between the sub-beams has elapsed. It can be determined that a defect exists.
[0044]
On the other hand, a case where a concave defect corresponding to a phase of 10 ° is scanned will be described. Since the defect acts to delay the phase by 10 ° with respect to the leading sub-beam, the total phase difference when the leading sub-beam scans the defective portion and the subsequent sub-beam scans the normal portion is 80 °, Both sub-beams have an output intensity slightly lower than the light-receiving element output intensity when scanning the normal part. On the other hand, when the rear sub-beam scans this defect and the leading sub-beam scans the normal part, the defect is further 10 ° relative to the trailing sub-beam having a 90 ° phase lag with respect to the leading sub-beam. Only the phase is delayed, and the total phase difference between both sub-beams is 100 °. As a result, the output intensity of the light receiving element is greatly reduced compared to the output intensity when both sub-beams scan the normal part. Therefore, when a signal with a greatly reduced brightness is generated following a signal with a slightly reduced brightness, it is determined that a concave defect exists.
[0045]
By using the interference inspection mode in this way, it is possible to determine the presence of a defect from the change in the light amount of the interference beam. Furthermore, it is possible to easily determine whether the defect is a concave defect or a convex defect based on the generation order of a signal whose luminance is greatly reduced and a signal that is slightly reduced at a time interval corresponding to the distance between the sub-beams. In this example, the Nomarski prism 30 gives a phase difference so that the leading sub-beam advances by 90 ° relative to the rear sub-beam, but it may be set to give a phase difference of 270 °. That is, a phase difference of (2m + 1) π / 2 can be given between the sub-beams. Further, even if the leading sub-beam is set so that its phase is delayed by 90 ° with respect to the rear sub-beam, the unevenness of the defect can be similarly determined.
[0046]
Next, a method and a specific circuit for implementing the above-described defect inspection method will be described. FIG. 7 is a signal waveform diagram when performing unevenness determination. FIG. 7A shows an output signal waveform of the light receiving element when a convex defect portion is scanned. When both sub-beams scan the normal part, a signal with a high luminance and constant amplitude is output. After that, when the leading sub-beam scans the convex defect part, a pulse with a greatly reduced luminance is generated. Next, after a time delay t corresponding to the distance between the sub-beams, a pulse with a slight decrease in luminance is generated. The presence of a defect is determined by the generation of a pulse with reduced brightness. FIG. 7B shows a signal obtained by delaying the output signal of the light receiving element by the scanning delay time between the sub-beams. Further, FIG. 7C shows a difference signal obtained by subtracting a delay signal from the output signal of the light receiving element. In the difference signal, a pulse having a positive sign is generated after the second pulse, that is, a time delay t of scanning between sub-beams.
[0047]
FIG. 7D shows an output signal waveform of the light receiving element when the concave defect portion is scanned. When both sub-beams scan the normal part, a signal with a high luminance and constant amplitude is output, and when the leading sub-beam scans the convex defect part thereafter, a pulse with a slight decrease in luminance is generated. Next, after the time delay t corresponding to the distance between the sub-beams, a pulse having a greatly reduced luminance is generated. FIG. 7E shows a signal obtained by delaying the output signal of the light receiving element by the scanning delay time between the sub-beams. Further, FIG. 7F shows a difference signal obtained by subtracting a delay signal from the output signal of the light receiving element. In this difference signal, a pulse with a negative sign is generated after the second pulse, that is, the time delay of scanning between sub-beams.
In this way, after the generation of a pulse whose brightness has decreased due to scanning of the defective portion, a convex shape is formed depending on whether a positive sign pulse appears or a negative sign pulse appears at a timing delayed by the scanning delay time between the sub-beams. It is possible to determine whether it is a defect or a concave defect.
[0048]
FIG. 8 is a circuit diagram showing a configuration of an example of the unevenness determination circuit. The video signal from the light receiving element is supplied to the positive input of the delay circuit 31 and the first comparator 32. The delay circuit 31 delays the video signal by a time corresponding to the time delay t between the sub-beams, and supplies this delayed output to the negative input of the first comparator 31. A difference signal corresponding to FIGS. 7C and 7F is output from the first comparator 32. This difference signal is supplied to the second and third comparators 33 and 34, respectively. These comparators 33 and 34 determine whether the difference signal takes a positive value or a negative value. The outputs of these comparators are supplied to a NAND gate 35, and the output signal is supplied to a one-shot multivibrator 37 via a t / 2 delay circuit 36. The output of the multivibrator is supplied to the first and second AND gates 38 and 39, respectively. Output signals from the second and third comparators 33 and 34 are supplied to the other input terminals of the AND gates 38 and 39, respectively. With this configuration, a defect detection signal indicating the presence of a concave defect is generated from the first AND gate 38, and a defect detection signal indicating the presence of a convex defect is generated from the second AND gate 39.
[0049]
  Next, the defect inspection of the phase shift mask as a sample to be inspected will be described. FIG. 9 is a diagram showing a configuration of an example of a phase shift mask. FIG. 9A is a plan view and FIG. 9B is a plan view of FIG.II-II lineIt is sectional drawing. In the phase shift mask 40, a light-shielding pattern 42 that defines a light transmission opening is formed on a transparent substrate 41, light transmission openings are repeatedly formed at a predetermined period, and projection light 43b passing through adjacent light transmission openings 43a and 43b are transmitted. A phase difference of λ / 2 is given to the projection light passing therethrough. And since the depth of the recessed part 43b which comprises a phase shifter is about 100 nm, it is necessary to test | inspect whether the depth is formed correctly. In this case, by detecting the phase shifter height information as the phase difference between the two sub beams using the interference beam inspection mode according to the present invention, the defect inspection of the phase shifter having a depth of about 100 nm is accurately performed. be able to. Even using the confocal inspection mode, it is possible to inspect the concavo-convex pattern that is inspected on the surface of the transparent substrate and functions as a phase shifter. As shown in FIG. 9A, the first sub-beam b is arranged along the arrangement direction 44a of the light transmission apertures as a pattern._s1And the second sub beam b along the arrangement direction 44b of the adjacent light transmission apertures._s2Scan with. Then, by synthesizing these sub-beams, the phase-shifter height information is included as phase information in the beam-combined interference beam. Therefore, if the beam scan is performed by introducing a phase difference of (2m + 1) π / 2 between the sub beams by the Nomarski prism, the depth constituting the phase shifter is slightly less than the reference value.SlipTherefore, the brightness can be greatly changed, so that a higher defect defect inspection can be performed. That is, for example, when a phase shift pattern is formed by etching a substrate in a Levenson type phase shift mask, a portion where a defect is likely to occur and inspection is required is a substrate etching portion. When this portion is inspected, a defect can be detected by comparing the interference light intensity when the chrome light-shielding pattern and the reflected light from the substrate etching portion overlap and interfere with each other. In this case, if the phase difference due to the Nomarski prism is adjusted so that the phase difference of the reflected light on the photodetector becomes (2m + 1) × π / 2, it interferes with a minute change in phase difference. Since the light intensity changes greatly, it is possible to detect a defect with high sensitivity. The displacement direction of the sub beam may be either the beam scanning direction or the direction orthogonal to the beam scanning direction. In this case, the defect detection process can be performed by storing inspection information of one die in a memory and performing die comparison between the detected defect inspection information and the stored inspection information. Alternatively, it can be performed by comparing the output intensities of interference beams adjacent to each other, or can be performed by cell comparison or array comparison in which the output intensities of interference beams separated by an integral multiple beam pitch are compared. As described above, the interference beam inspection according to the present invention is extremely useful for the defect inspection of the phase shift mask in which the concavo-convex pattern constituting the phase shifter is formed on the surface. There are various types of phase shift masks. For example, the present invention can be applied to a phase shift mask that introduces a phase difference of λ / 3, 2λ / 3, or the like between projection beams. The present invention can also be applied to defect inspection of a chromeless phase shift mask in which no is formed or a halftone phase shift mask in which the light shielding pattern itself gives a phase difference. Further, it is useful for defect inspection of an optical device such as an optical integrated circuit and defect inspection of a substrate of a semiconductor device during a manufacturing process.
[0050]
In the defect inspection of the phase shift mask, the defect inspection is performed on the entire surface of the phase shift mask in the confocal inspection mode or the interference inspection mode, the presence of the defect and its address are detected, the head is then moved to the defect position, Defect height information can be detected by scanning multiple times while changing the relative distance between the focal point and the surface of the phase shift mask, or using the phase shift method described above using an interference beam.
[0051]
The present invention is not limited to the above-described embodiments, and various modifications and changes can be made. For example, although the multi-beam illumination method in which a sample is scanned using a plurality of light beams is used in the above-described embodiments, the present invention can be applied to a case where a defect inspection is performed using a single light beam.
[0052]
Further, in the above-described embodiments, the Nomarski prism is used as the interference optical system. However, various lateral shears such as a prism element that generates two sub-beams having coherence with each other, such as a Wollaston prism and a Lotion prism, are used. A ring interference element can be used.
[0053]
Furthermore, a white light source or LED can also be used as a light source.
[Brief description of the drawings]
FIG. 1 is a diagram showing a configuration of an example of a defect inspection apparatus according to the present invention.
FIG. 2 is a circuit diagram showing a configuration of a defect detection circuit.
FIG. 3 is a graph showing a relationship between a distance between a focal point of an objective lens and a sample surface and luminance.
FIG. 4 is a diagram showing the configuration of a defect inspection apparatus according to the present invention that performs interference mode inspection.
FIG. 5 is a graph showing the relationship between the scanning direction position and the output intensity of the light receiving element in the interference inspection mode.
FIG. 6 is a graph showing the relationship between the introduced phase difference and the intensity indicated by the light receiving element.
FIG. 7 is a signal waveform diagram of a video signal, a delay signal, and a difference signal in the interference mode inspection.
FIG. 8 is a circuit diagram showing a configuration of an example of an unevenness determination circuit.
FIG. 9 is a diagram showing a configuration of an example of a phase shift mask.
[Explanation of symbols]
1 Light source
2 Diffraction grating
5 Beam splitter
6 Galvano mirror
10 Objective lens
11 samples
12 Sample stage
14 Linear image sensor
17 Signal processing circuit
30 Nomarski Prism
31 Position detector

Claims (12)

Confocal inspection mode that scans the sample surface with multiple light beams and detects defects on the sample surface based on the luminance information of the reflected light from the sample, and interference that detects the height information of the sample surface as phase information A defect inspection device switchable between inspection modes,
A light source device that emits a plurality of light beams aligned along a first direction;
A beam deflecting device for deflecting the plurality of light beams in a second direction orthogonal to the first direction;
An objective lens that focuses a plurality of incident light beams into a spot shape and projects it onto the sample surface;
A sample stage that supports the sample to be inspected;
In the interference inspection mode, the incident light beam inserted into the optical path between the beam deflecting device and the objective lens is converted into first and second sub beams having coherence with each other, and the sample surface An interference optical system for generating an interference beam including phase information related to the height of the sample surface, by combining sub-beams that are reflected by each other and related to each other,
A linear image sensor having a plurality of light detection elements arranged in a direction corresponding to the first direction and receiving reflected light or interference beams from the sample surface;
A signal processing circuit for generating a defect detection signal based on an output signal from the linear image sensor;
In the confocal inspection mode, defect detection means for detecting a defect using a luminance signal output from the light detection element of the linear image sensor;
And a means for discriminating whether the defect is a convex defect or a concave defect based on a change in the amplitude of the interference beam output from the light detection element of the linear image sensor in the interference inspection mode. apparatus.   2. The defect inspection apparatus according to claim 1, wherein the interference optical system converts each of the incident light beams into first and second sub beams separated by a predetermined distance in the second direction. A defect inspection apparatus that scans the surface of the sample with the first sub-beam and then scans with the second sub-beam. The defect inspection apparatus according to claim 1 or 2, wherein the defect detection means of the signal processing circuit detects a defect by comparing an output signal from each light detection element of the linear image sensor with a reference value, Means for detecting defects by comparing output signals from the light detection elements, or a signal obtained by delaying the output signal from the light detection element by one line, and a signal delayed by one line and a signal not delayed A defect inspection apparatus comprising any one of defect detection means for detecting defects by comparing.   4. The defect inspection apparatus according to claim 1, wherein a photomask or wafer on which a plurality of dies or chips having the same pattern is formed as the sample, and the signal processing circuit is in a confocal inspection mode. A defect inspection apparatus characterized in that defect detection is performed by die-to-die comparison inspection.   5. The defect inspection apparatus according to claim 4, wherein a phase shift mask having a transparent substrate and having a phase shifter formed on the surface thereof is used as the photomask.   6. The defect inspection apparatus according to claim 1, wherein a prism element is used as the interference optical system, and m is an integer, the interval between the first sub-beam and the second sub-beam. (2m + 1) π / 2 phase difference is introduced into the defect inspection apparatus. Confocal inspection mode that scans the sample surface with multiple light beams and detects defects on the sample surface based on the luminance information of the reflected light from the sample, and interference that detects the height information of the sample surface as phase information A defect inspection method switchable between inspection modes,
In the confocal inspection mode, a light source device that emits a plurality of light beams aligned in the first direction is used, the sample surface is scanned with the plurality of light beams emitted from the light source device, and reflected light from the sample a step received by the linear image sensor, which detects the defects and the address present on the sample surface based on the luminance signal outputted from the linear image sensor,
In the interference inspection mode, an interference optical system is inserted in the optical path between the light source device and the sample, a plurality of light beams emitted from the light source device are incident on the interference optical system, and the light beams are made coherent with each other. Converting the first and second sub-beams to each other, and scanning the sample surface with the generated sub-beams;
Synthesizing sub-beams reflected on the sample surface and related to each other to form an interference beam including phase information corresponding to the height of the sample surface;
Receiving the interference beam by the linear image sensor and determining whether the defect is a convex defect or a concave defect based on a change in amplitude of the interference beam output from the linear image sensor. A defect inspection method characterized by that.   8. The defect inspection method according to claim 7, wherein the interference optical system converts a plurality of incident light beams into first and second sub beams separated by a predetermined distance in a direction orthogonal to the first direction. A defect inspection method comprising: scanning a sample surface with a first sub-beam and then scanning with a second sub-beam.   9. The defect inspection method according to claim 7, wherein when a prism element is used as the interference optical system and m is an integer, (2m + 1) π / between the first sub-beam and the second sub-beam. A defect inspection method, wherein a phase difference of 2 is introduced. In the defect inspection method according to claim 8 or 9, linear when the amplitude of the first output signal from the linear image sensor when the sub-beam is scanned over the defect, the second sub-beam is scanned over the defect A defect inspection method for determining whether a defect is convex or concave by comparing the amplitude of an output signal from an image sensor.   11. The defect inspection apparatus according to claim 7, 8, 9, or 10, wherein a photomask or wafer on which a plurality of dies or chips having the same pattern is formed as the sample, and a die pair in the confocal inspection mode. A defect inspection method characterized by performing defect detection by die comparison inspection.   12. The defect inspection method according to claim 11, wherein a defect of the phase shift mask is detected using a phase shift mask having a transparent substrate as a sample and having a phase shifter formed on the surface thereof. Defect inspection method.

Images