JPS63200042A - Method and apparatus for inspecting flaw of pattern - Google Patents

Abstract

PURPOSE:To detect a flaw smaller than the resolving power of an optical system without transferring data to a hologram or a filter, by respectively irradiating a specimen to be inspected and a standard specimen with two beams adjusted so as to be different by pi in a phase. CONSTITUTION:A mask 30 becoming a specimen to be inspected and a mask 31 becoming a standard specimen are respectively placed on the specimen-to-be-inspected fixing jig 22 and standard specimen fixing jig 23 on an XYZ table 19. The beam emitted from a coherent beam source 5 is separated into two beams by a half mirror 7 and one of two beams adjusted so as to be different by pi in a phase by a phase plate 9 is allowed to irradiate the mask 30 and the other beam is allowed to irradiate the mask 31. The beams allowed to irradiate two masks 30, 31 are reflected from the masks while possess the data of the patterns on the masks to interfere with each other on a half mirror 24. At this time, since the phase data of the beams diffracted by the patterns of the masks 30, 31 become equal, when one of the phases is delayed by pi, luminous intensity becomes weak by the interference. Contrarily, the beams diffracted from a flaw doe not interfere with each other to be emitted to a detector and, therefore, only the flaw can be detected in a brightly glowing state.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to a pattern defect inspection method and apparatus, particularly for patterns such as reticles and masks used in LSI manufacturing.

[Conventional technology]

For example, JP-A-59-58586, JP-A-59-58586,
Publication No. 87345 discloses a method of capturing images with a pattern defect inspection device and detecting defects through software image processing, but the disclosed method is a method of determining defects based on differences in the shape of the images. Therefore, if the defect is smaller than the resolution of the imaging optical system, the defect cannot be detected.

On the other hand, a method of extracting a defect using light diffraction or interference development can detect a defect even if the defect is smaller than the resolution of the imaging optical system. Such a method is described in Holographic Photomask Defect Inspection System: LH Lin et al.: SPI Volume 538 Optical Microlithography IV (1985) Volume 110
~116 pages (A holographic photo
mask defect inspection systems
tem L, H, Lin D, L, Cavan R,
B, Howe and R,E.

Grave, 5PIE Vo 12.5380ptic
al microlithogaphyrV (198
5) disclosed in ppHo-116). In the disclosed method, a filter and a hologram are formed on a photographic plate by transferring a Fourier transform image of the mask to be inspected, and when the hologram is reproduced, the Fourier transform image is used to reproduce the mask pattern. It detects only defects.

This method has the advantage that precise positioning is not required because the Fourier transform image and hologram are created with the photographic plate fixed.

In addition, JP-A-60-154634 and JP-A-60
Japanese Patent No. 154635 discloses an apparatus for inspecting pattern defects on rewafers by optical filtering.

[Problem that the invention seeks to solve]

In the above-described pattern defect inspection method using a hologram, the phase information of light is transferred to the hologram and the filter on the Fourier transform surface, so missing information is inevitable. Therefore, even if it is possible to detect defect information smaller than the resolution of the optical system due to diffraction or interference phenomena, there is a possibility that this defect information will be lost at the stage of being transferred onto a hologram and reproduced. For example, the above-mentioned document shows a detection example of a 0.65 μm defect, which is larger than the resolution of the optical system.

That is, in order to further reduce the detectable defects, it is necessary to create a highly accurate hologram and filter, but there are problems such as the setting of the exposure amount and phenomenon time, and the accuracy of the photographic plate. Furthermore, two photographic plates are required for each mask inspection.

The present invention makes it possible to provide a pattern defect inspection method and apparatus that can detect defects without transferring information to a hologram, filter, etc., and can detect defects smaller than the resolution of an optical system at high speed. The purpose is to

[Means for solving problems]

The structure of the pattern defect inspection method of the present invention, which was adopted to solve the above-mentioned problems, is to place a test sample and a standard sample with no defects on a sample stage, and emit two beams of light from a coherent light source. One of the lights, adjusted so that the phases of the two beams differ by π, is irradiated onto the test sample and the other light is irradiated onto the standard sample. The pattern defect inspection apparatus of the present invention is characterized in that reflected light or transmitted light is guided as the same light flux to form an image of the test sample, and the configuration of the pattern defect inspection apparatus of the present invention includes: a coherent light source; means for separating the light emitted from the coherent light source into two beams;
A light irradiation part having a phase plate that adjusts the phase difference between the light beams, a fixing jig for an I quasi-sample with no defects, and a fixing jig for a test sample provided with a stage that can be moved slightly within its plane. a drive system that drives the XYZ stage and the finely movable stage;
A light beam is irradiated onto the test sample and the standard sample, and a phase shift of the light reflected from or transmitted through each of the test sample and the standard sample is detected, and the phase plate and the fine movement are detected. a stage section having a photodetector that generates a signal for possible adjustment of the stage, which guides the light reflected from each of the test sample and the standard sample as the same light beam, or the light transmitted through each of the test sample and the standard sample; a photodetector having a photodetector that forms an image of the photodetector; and a binarization circuit that passes the detection result of the photodetector of the photodetector through an amplifier and then converts it into a binary value.

The present invention is characterized in that it has a control section including a microcomputer into which the output of the binarization circuit is input and outputs a control signal for the XYZ stage and a display signal to a defect image display device.

That is, the present invention has a configuration that compares two masks instead of using a hologram and a filter, and utilizes light diffraction and interference.The configuration that optically compares two masks is based on the Twyman A mask is placed at the position of the reference light generation mirror of the green interferometer and the position of the test sample,
The diffracted light is focused by a lens and interfered with on a half mirror, and in order to further improve the signal-to-noise ratio, the irradiated light is made into a spot and two masks are moved in synchronization. It becomes.

[Effect]

Since the structure is as described above, the light irradiated onto the two masks is reflected with information about the pattern on the masks,
Interference on half mirror. At this time, the phase information of the light diffracted by the two mask patterns is equal, so
By delaying one phase by π, the light intensity can be weakened by interference. On the other hand, since the light diffracted from the defect is emitted to the detector side without interference, only the defect shines brightly and can be detected.

If the defect in the test sample is smaller than the resolution, the defect information will not interfere, so it can be detected as a bright spot even if the shape is not reproduced.

Furthermore, by making the irradiation light into a spot shape, the intensity of the light irradiated onto the defect becomes relatively large, so that the S/N ratio is improved.

〔Example〕

Examples will be described below.

FIG. 1 is an explanatory diagram of the configuration of a pattern defect inspection apparatus that implements an embodiment of the pattern defect inspection method of the present invention.

This pattern defect inspection apparatus includes a light irradiation section 1. Photodetector 2
．． It consists of a stage section 3 and a control section 4.

The light irradiation unit 1 includes a He-Cd laser 5. Beam expander 6, half mirror 7, mirror 89 phase plate 92 rotation axis 1
0 and a phase plate driving system 11, the photodetecting section 2 includes objective lenses 12 and 13, half mirrors 14 and 15. The stage section 3 is composed of an imaging lens 16, a pinhole 17, and a photodetector 18 made of, for example, a silicon photodiode.
Z stage drive system 20. XY fine movement stage 21, mask fixing jigs 22 and 23. Half mirror 24, imaging lens 25. Photodetector 26. Fine movement stage control system 27. A mask 30 serving as a test sample and a mask 31 serving as a standard sample, which are composed of light absorption plates 28 and 29, are placed on a test sample fixing jig 22 and a standard sample fixing jig 23, respectively. Note that the control unit 4 includes an amplifier 32 and a binarization circuit 33.
, a microcomputer 34 and a CRT 35.

In the light irradiation section 1, the wavelength λ=3 from the He-Cd laser 5
Coherent light of 25 nm is made into a parallel light beam with a diameter d = approximately 1 m through the beam expander 6, and is converted into a parallel light beam with a diameter d = approximately 1 m.
It is divided into two directions by 7. After being reflected by mirror 8, one of the two directions is reflected by half mirror 1 of photodetector 2.
5 and the objective lens 13 to illuminate the mask 31 of the stage section 3, and the other one passes through the phase plate 9.
Through the half mirror 24 of the stage section 3, the light detection section 2
The mask 30 is illuminated through the half mirror 14 and the objective lens 13.

The phase plate 9 of the light irradiation unit 1 has a thickness of about 8 nn and a diameter of about 100 mm.
It is a circular disk having a taper on the order of a wavelength in the diameter direction, and can be rotated by a phase plate drive system 11 around a rotating shaft 10 provided at the center. Although quartz is used for the phase plate 9, other materials may be used as long as they produce a phase difference.

The light that has passed through this phase plate 9 is delayed within a range of 2π from the phase O, corresponding to the passing position on the phase plate 9. That is, the phase of the passing light can be controlled. In a general initial state, one phase is delayed by π. Diameter d = approx. I
The luminous flux of the cabinet is transmitted through the masks 30 and 31 by the objective lenses 12 and 13 whose focal length is f, as shown in the formula below.
The light is focused upward as a spot diameter ds.

ds"1.22λ-...(1) For example, when using an objective lens with f=approximately 10 mm, ds
is approximately 4 μm. The spot diameter ds can be adjusted by changing the diameter d of the injection point by the beam expander 6, if necessary.

Also, instead of the He-Cd laser 5, a coherent light source of another wavelength, such as a U e -N e laser, an Ar
A laser, NZ laser, etc. may also be used.

If the pattern dimensions on the masks 30 and 31 are large and the size of the defect to be detected is large, it is better to use a light source with a long wavelength. This is because the longer the wavelength, the easier the alignment accuracy will be at the stage of aligning the masks 30 and 31, which will be described later. Experiments have shown that a wavelength approximately 0.5 to 5 times that of the defect to be detected gives good results.

Here, the optical path length from the half mirror 7 to the mask 30 and the optical path length from the half mirror 7 to the mask 31 are configured to be approximately equal. Luminous flux 3 of I-I e -Cd laser 1
If 6 is an ideal parallel light beam, the two optical path lengths do not need to be equal, and may be within the temporal coherence length of the light source, but in reality.

Because it has a spread of several milliradians, the mask 30
The spot diameters on and 31 will be different. It is necessary to keep this spot diameter within a range of about λ/10,
For this purpose, it is necessary to set these two optical path lengths within a range of approximately 6 μm. The allowable range of this spot diameter is S/N
It can be determined through experiments.

The light irradiated onto the masks 30 and 31
The light that is diffracted by the patterns above 0 and 31 and enters the apertures of objective lenses 12 and 13 exits as nearly parallel light. The light emitted from the mask 31 passes through the half mirror 15.
The light emitted through the mask 30 is directly bent onto the half mirror 14 and interferes with the light. As a result of interference, all nine diffracted lights reflected from the same pattern are half mirror 24.
The light passes through the imaging lens 25 and is detected by the photodetector 26. Further, the reflected light from the defective portion on the mask 30 travels approximately 50% toward the imaging lens 16 and approximately 50% toward the half mirror 24 without interference.

An image is formed on the pinhole 17 by the imaging lens 16.

Photodetection) Detected by pt IB.

Here, too, the difference between the optical path length from the mask 30 to the half mirror 14 and the optical path length from the mask 31 to the half mirror 14 must be in the range of approximately 6 to 10 ILm.

Here, an example is shown in which a silicon photodiode is used as the photodetector [8], but instead of this, the pinhole 17 can be removed and a charge-transfer type one-dimensional surface image sensor, a silicon photodiode, for example, can be used as the photodetector. A one-dimensional array of diodes may also be used.

When these photodetectors are used, a large area can be inspected at one time, thereby further speeding up the inspection time. Further, a photomultiplier tube may be used as the photodetector 18, and in this case, since the photodetector has good sensitivity, the light source can be made small by using a low output He-Cd laser or the like.

Note that during detection by the photodetector 18, it is effective to make the area to be detected small by using the pinhole 17 or the aperture of each detection pixel of the one-dimensional surface image sensor or one-dimensional array. By doing so, light from sources other than the defect, that is, noise light, can be reduced, and the signal of the defect can be relatively increased. Therefore, the size of the minimum defect that can be detected also depends on the size of the pixels of these photodetectors, so it is necessary to optimally design it in consideration of the detection speed.

In the stage section 3, masks 30 and 31 are installed.

The mask is fixed onto the XYZ stage 19 and the XY fine movement stage 21 by mask fixing jigs 22 and 23. XY
The fine movement stage 21 is driven by a fine movement stage drive system 27 based on a signal from the photodetector 26.

In addition, the Z stage of the XYZ stage 19 has a mask 30.
The X and Y stages of the XYZ stage 19 change the position of the irradiated light spot on the mask 30.

The X stage has a uniform acceleration time of approximately 0.03 seconds and a
The maximum speed is approximately 1 m/sec, the amplitude is 2.0 m/sec, and the uniform deceleration time is 0.03 seconds. OO
Performs periodic motion of +m. The Y stage is synchronized with the uniform acceleration and deceleration times of the X stage, and is sent in steps of 41 meters in the - direction. If it is sent 7,500 times during one inspection time, 30 na can be sent in about 12 minutes. With this configuration, an area of 30+m square can be scanned in about 12 minutes.

The stage speed shown here is an example and should be designed depending on the mask size, inspection time, and detection pixel size.

Further, an XYz stage 19, an XYR moving stage 21, and holes 37, 38 and 39 are formed, so that the light that is irradiated onto the mask 30, 31 and passes through the portion where no pattern is formed is transmitted. Further, at the backs of the holes 37 and 38, light absorbing plates 28 and 29 made of black rasher paper are provided, respectively.

With this configuration, the reflected light from the pattern-free areas passes through the masks 3o and 31 and is absorbed by the light-absorbing plates 28 and 29, respectively, so that the light reflected from the pattern-free areas is irradiated onto the pattern-free areas on the masks 30 and 31. The influence of reflected light can be reduced, and reflected light from the patterned portion can be effectively detected.

In the control unit 4, the microcomputer 34 controls the XY
The Z stage drive system 20 is controlled, and the XYZ stage 19
is controlled. At the same time, the signal from the photodetector 18 is binarized by an amplifier 32 and a binarization circuit 33, stored in a microcomputer 34, and the position of the defective part is displayed on a CRT 35.

Next, the operation of this embodiment will be explained.

Masks 30 and 31 having the same pattern are placed on mask fixing jigs 22 and 23, respectively, and He-
Light is emitted from the Cd laser 5.

The light irradiated onto the masks 30 and 31 is diffracted by the shape of the pattern, and passes through the objective lenses 12 and 13, respectively.
and reach Half Mirror 14. here,
Considering an arbitrary point 40 on the half mirror 14, the diffracted light from the entire surfaces of the respective masks 30 and 31 reaches this point 40. Since the patterns on masks 30 and 31 are identical, the light that reaches point 40 is determined only by the phase of the light that falls on masks 30 and 31. As mentioned above, the phase of the irradiated light is shifted by π by the phase plate 9, so the amplitude intensity of each light beam is expressed as the first and second terms on the left side of equation (2) below. The intensity decreases due to interference.

-jut -1(ut+x)=O...(2)I
e + Ie Here, ■ is the respective amplitude intensity, and ω is the angular frequency of the light.

As described above, since the amplitude intensity becomes 0 at any point, the detection signal at the photodetector 18 also becomes almost 0. However, in this case, since the light also reaches the half mirror 24 side, the detection signal at the photodetector 26 becomes maximum. Take advantage of these properties.

The positions of the XY fine movement stage 21 and the phase plate 9 are adjusted. That is, the XY fine movement stage 21 is positioned at the position where the signal from the photodetector 26 is detected and the signal becomes maximum. Thereafter, the phase plate 9 is rotated once, and the phase plate 9 is set at a position where the signal from the photodetector 26 becomes maximum. By repeating this operation 2 to 3 times, the accuracy of adjustment can be further improved.

After this adjustment is completed, the XY stage of the xyz stage 19 is driven according to the specifications shown above, and the signal from the photodetector 18 is taken into the control section 4.

Next, signal processing will be described using as an example the mask 30 having a pattern 42 with a defect 41 and the mask 31 having a pattern 43 without a defect shown in FIGS. 2(a) and 2(b).

When the scanning position 44 on the two masks 30 and 31 is scanned, the photodetector 18 detects the horizontal axis. As shown in FIG. 3, in which the stage position and detection symbol are plotted on the and vertical axes, a beak 45 is placed at a position corresponding to the defect 41.
A signal with . When this signal is binarized by providing a threshold value 46, a signal as shown in FIG. 5 is obtained, with the stage position and the binarized signal taken on the horizontal and vertical axes, respectively. Defects can be detected and, as a result,
An image as shown in FIG. 5 is obtained in which only the defective portion 41 is bright.

Here, a case where an X-ray mask is inspected will be explained.

FIG. 6 is a cross-sectional view of the X-ray mask, in which a gold pattern 49 is held by a substrate film 47 and a protective film 48, and these films are stretched over a silicon frame 53.

In the case of such a film, since the substrate film 47 or the protective film 48 has uneven J'S, the phase of the incident light reflected by the gold pattern 49 may change.

゛In the case of such an object to be inspected, it is sufficient to adjust the phase plate 9 so that the position (11) of the emitted light from each mask 30 and 31 is shifted by exactly π, that is, the phase of the incident light is During the inspection, the phase plate drive system 1 is adjusted independently so that the signal from the photodetector 26 is maximized.
1, it is necessary to control at high speed. In JL terms, in this example, the adjustment is completed within 1 μscc.

As described in 1- above, the pattern defect inspection apparatus of the embodiment is effective for defect inspection of masks having the same pattern. Furthermore, according to this method, not only defects but also foreign substances on the mask can be inspected.

FIG. 7 is an explanatory diagram of the configuration of a pattern defect inspection apparatus that implements another embodiment of the pattern defect inspection method of the present invention. This embodiment differs from the embodiment shown in FIG. 1 in that it is a transmission type defect inspection apparatus. This is a point.

In this figure, the same parts and parts that act in the same way as in FIG.

In this apparatus, the light irradiation section 1 and the control section 4 have the same configuration as the reflective defect inspection apparatus shown in FIG. The photodetecting section 2 and the stage section 3 are equipped with masks 30 and 31.
The light transmitted through the hole 38 and the objective lens 50 is guided to the half mirror 14 on the one hand, and through the holes 39 and 37 on the other hand. Although the light is guided to the half mirror 14 through the objective lens 51 and the mirror 52, the structure and operation of other parts are the same as in the embodiment shown in FIG.

This embodiment also achieves the same objectives and effects as the first embodiment.

Note that in the X-ray mask shown in FIG. 6, the surface of the protective film 48 is often not flat, so it is better to inspect it from the J-substrate film 47 side using a reflective defect inspection device.

According to these embodiments, defects can be detected using light diffraction and interference without transferring optical phase information onto holograms or filters, so defects smaller than the resolution of the optical system can be detected without losing their information. 6 [Effects of the Invention] The present invention can detect defects without transferring information to holograms, filters, etc., and can detect defects smaller than the resolution of the optical system at high speed. This makes it possible to provide a pattern defect inspection method and apparatus that can be used to inspect patterns, which has great industrial effects.

[Brief explanation of the drawing]

FIG. 1 is a block diagram of an apparatus for carrying out an embodiment of the pattern defect inspection method of the present invention, FIG. 2 is a plan view showing an example of a pattern in comparison with a standard pattern, and FIG. An explanatory diagram of the detection signal obtained from the pattern, Figure 4 is the third
FIG. 5 is a plan view showing detected defects, FIG. 6 is a cross-sectional view of another example of the pattern, and FIG. 7 is a diagram of the pattern inspection method of the present invention. FIG. 3 is a block diagram of an apparatus implementing another embodiment. DESCRIPTION OF SYMBOLS 1... Light irradiation part, 2... Light detection part, 3... Stage part, 4... Control part, 5... He-Cd laser, 9
...Phase plate, 7.14.15...Half mirror-112
, 13... Objective lens, 18.26... Photodetector,
30.31... (1 other person) House 1 diagram missing''; L diagram (dwelling) (6) [type
Sora female

Claims (1)

[Claims] 1. Place a test sample and a standard sample with no defects on a sample stage, emit light from a coherent light source and separate it into two beams, such that the phases of the two beams differ by π. One of the adjusted lights is irradiated on the test sample and the other one is irradiated on the standard sample, so that the light reflected from each of the test sample and the standard sample, or the light transmitted through each, are the same luminous flux. A pattern defect inspection method characterized by forming an image of the sample to be inspected. 2. A light irradiation unit having a coherent light source, a means for separating the light emitted from the coherent light source into two beams, and a phase plate for adjusting the phase difference between the two beams, and fixing a standard sample free of defects. an XYZ stage having a test sample fixing jig provided with a jig and a stage that can be moved finely within its plane; a drive system that drives the XYZ stage and the stage that can be moved finely; detecting a phase shift of light that is irradiated onto the test sample and the standard sample and reflected from each of the test sample and the standard sample, or of light that is transmitted through each; the phase plate and the finely movable stage; a stage section having a photodetector that generates a signal for adjustment of the test sample, and guides the light reflected from each of the test sample and the standard sample, or the light transmitted through each as the same light flux, and generates an image of the test sample. a photodetector having a photodetector that forms an image; and a photodetector that passes the detection result of the photodetector of the photodetector through an amplifier;
It has a control unit that has a binarization circuit that converts into values, and a microcomputer that receives the output of the binarization circuit and outputs control signals for the XYZ stage and display signals to the defect image display device. A pattern defect inspection device featuring: