JP2005055447A - Method and apparatus for inspecting defects - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To detect defects in a minute wiring pattern formed on a sample with high sensitivity, by improving the contrast of an optical image of the sample formed by the interference between the 0-th and higher-order diffracted lights of reflected light, arising from illuminating light reflected by the sample surface, in a defects detection optical system which detects defects in a pattern formed on the sample surface. <P>SOLUTION: A high resolution optical system is constructed by an illumination optical system which illuminates the sample with a polarized light, a polarization optical component which transmits higher-order diffracted lights, after being subjected to polarization rotation by the sample more efficiently than the 0-th order light, and a detection optical system which forms an image of the sample onto a photoelectric conversion device with the light which is transmitted through or is reflected by the polarization optical component. <P>COPYRIGHT: (C)2005,JPO&NCIPI

Description

  The present invention relates to a high-resolution optical system used for inspection and observation of fine pattern defects and foreign matters in a semiconductor manufacturing process and a flat panel display manufacturing process, and a defect inspection apparatus using the same.

  As a conventional technique, Patent Document 1 describes a method and an apparatus for visualizing a fine line width structure using an optical microscope. As a feature of this technique, the direction of the polarization axis projected on the sample of the polarized light linearly polarized using the polarizing plate is set to an angle of about 45 ° with respect to the linear pattern on the sample. ing. The optical (delay) axis of the quarter-wave plate placed between the polarizer and the sample is oriented in the direction of the optimum angle (typically 25 °) with respect to the main linear shape of the sample. The quarter wave plate converts linearly polarized light into elliptically polarized light and illuminates the sample. This elliptically polarized illumination light undergoes a phase difference when reflected by the sample. These lights are again transmitted through the quarter-wave plate and polarized in such a direction as to pass through the polarizing plate provided on the detection optical path. The optical system is configured such that light transmitted through the polarizing plate on the detection optical path forms an image of the sample on the light receiving surface of the photoelectric conversion element. It describes that the amount of phase difference due to the sample is predicted in advance, the illumination light is set to elliptically polarized light, and the optical component is set to be circularly polarized after being reflected by the sample.

JP 7-128595 A None.

In the method of imaging a fine line width structure using an optical microscope as described in the prior art described above, a polarizer is arranged in the illumination optical path to transmit linearly polarized light, and this linearly polarized light is transmitted. The optical system is configured to illuminate the sample by converting it into elliptically polarized light using a quarter-wave plate, but this optical system predicts in advance the amount of phase difference caused by reflection from the sample and transmits it through the detection system analyzer. The illumination light is set to elliptically polarized light so that the amount of light to be emitted is maximized. For this reason, the ratio of the amplitudes of the 0th-order light and the higher-order diffracted light that contribute to the formation of the optical image is the same as in the case of the randomly polarized illumination (the amplitude of the 0th-order light is larger than that of the higher-order diffracted light). For this reason, the optical image of the sample formed by the interference between the 0th-order light and the higher-order diffracted light has a relatively low frequency component because the amplitude of the higher-order diffracted light is smaller than that of the 0th-order light. The resolution does not improve much.
In addition, when detecting an image using an image sensor or the like, it is necessary to adjust the amount of illumination light so that the amount of light received by the sensor does not saturate. For example, a periodic fine pattern has a large diffraction angle. The difference in brightness is small and may be detected dark, making it difficult to observe the pattern image. There is a problem that the detection sensitivity is low in a defect inspection apparatus using image processing in an image that has a small difference in brightness of a pattern and is detected dark.

  An object of the present invention is to control the polarization of illumination light and the polarization of zero-order light and high-order diffracted light that form an optical image, and detect minute defects by detecting a pattern targeted for observation or inspection with high resolution. Another object of the present invention is to provide a defect inspection method and apparatus including the optical system.

In order to achieve the above object, in the present invention, an illumination optical system for illuminating a sample with polarization, a polarization optical component that transmits high-order diffracted light that has undergone polarization rotation in the sample with high efficiency to zero-order light, and A high-resolution optical system is configured using a detection optical system that forms an image of a sample on a photoelectric conversion element with light transmitted or reflected by a polarizing optical component (for example, a vibration direction orthogonal to a linear pattern) And a polarizer having a transmission axis of about 45 ° with respect to the pattern is disposed).
The polarization illumination of this illumination optical system can also be performed using light transmitted or reflected by the polarization beam splitter. In this case, a λ / 2 plate and a λ / 4 plate are arranged between the polarizing beam splitter and the sample, and the polarization direction is adjusted by rotating the λ / 2 plate according to the direction of the pattern, etc., and λ / 4 The ellipticity of polarized light can be adjusted by rotating the plate.
There are many combinations of the vibration direction of the illumination light and the vibration directions of the 0th-order light and the higher-order diffracted light that form an optical image. These setting methods are important for practical use. In the present invention, the optimal polarization state can be selected in a short time by listing the images detected by changing the polarization state, or by performing a preliminary defect inspection and selecting a polarization state that increases the defect detection sensitivity. So that it can be calculated accurately.

  According to the present invention, a high-resolution image is detected by controlling the angle formed by the direction of vibration of the electric field vector on the polarization plane of the illumination light and the direction of the pattern according to the pattern formed on the sample, thereby detecting pixels. We have realized an optical system that can detect minute defects of about half the size. Furthermore, the vibration direction of the electric field vector on the polarization plane and the illumination wavelength can be set with high efficiency, and a high-resolution optical system and a high-sensitivity defect inspection apparatus using the same can be realized.

An embodiment of a high resolution optical system of the present invention will be schematically described with reference to FIG.
Light 180 ′ emitted from the light source 8 is collected through the lens 9 and passes through the polarizing plate 14. The light transmitted through the polarizing plate 14 becomes linearly polarized light, and the light transmitted through the half mirror 15 becomes illumination light 180. The light transmitted through the half mirror 15 illuminates the sample 1 through the objective lens 20. Of the light 189 ′ reflected and diffracted by the sample 1, the light 189 in the NA (Numerical Aperture) of the objective lens 20 is again captured by the objective lens 20, reflected by the half mirror 15, and guided to the detection optical path. This detection light 189 transmits light in the vibration direction of the electric field vector on the polarization plane corresponding to the transmission axis of the polarizing plate 22 (hereinafter referred to as the polarization vibration direction), and passes on the image sensor 70 via the imaging lens 30. An image of the sample 1 is formed.
Here, the polarizing plate 14 has a role of determining the polarization vibration direction with respect to the main pattern direction 1a formed on the sample 1. For example, the polarization direction is orthogonal to the pattern direction 1a to be detected. Position so that it has the polarized light. This vibration direction is the vibration direction in the exit pupil 19 of the objective lens 20.
Further, the polarizing plate 22 transmits light in a specific vibration direction out of the detection light 189, and is set to, for example, 45 ° with respect to the pattern direction 1a. The polarization of the illumination light on the exit pupil 19, the 0th-order light reflected from the sample 1 and the higher-order diffracted light will be described with reference to FIG. The polarization of the illumination light on the sample 1 and the polarization of the 0th-order light and high-order diffracted light reflected from the sample 1 will be described with reference to FIG.
In FIG. 1B, the polarization of one point of the illumination light 180 at the exit pupil is 300. In the polarized light 300, the oscillation direction of the polarized light at the pupil position and the direction 301 in which the sample 1 is illuminated coincide (or are orthogonal), so that the polarized light 310 ′ of the specularly reflected light (0th order light) and the polarized light of the higher order diffracted light. 311 is the same as the vibration direction of the illumination light (in practice, since the phase difference is received when the sample 1 is reflected and diffracted, linearly polarized light becomes elliptically polarized light, but is omitted here).
However, when the vibration direction 300 ′ of the illumination light and the illumination direction 302 of the sample 1 do not match or do not intersect with each other, the vibration direction 310 ′ of the zero-order light is the same as the vibration direction 300 ′ of the illumination light. The vibration direction 320 of the folded light receives polarization rotation and is different from the vibration direction 300 ′ of the illumination light (note that the actual vibration direction has a distribution, and the vibration direction indicates the vibration direction with the largest amplitude).
For this reason, as shown in FIG. 1 (d), the vibration direction of the zero-order light and the vibration direction of the high-order diffracted light are different. Therefore, by matching the transmission axis of the polarizing plate 22 with the vibration direction of the high-order diffracted light. It is possible to increase the ratio of the amplitude of the high-order diffracted light at. As a result, the amplitudes of the 0th-order light and the higher-order diffracted light can be made substantially equal, and the resolution of the optical image formed by the interference between the 0th-order light and the higher-order diffracted light can be increased.
FIG. 2 shows line-and-space optical image detection results obtained by conventional epi-illumination and polarized illumination / detection according to the present invention. The transmission axes of the polarizing plates 14 and 22 when the image of the present invention is acquired are the conditions shown in FIG. The conventional illumination and the illumination σ of the present invention are both 1, and the diffraction image of the conventional illumination has 0th order light distributed throughout, and ± 1st order diffracted light is darker than the 0th order light and has a small amplitude. On the other hand, in the diffraction image of the present invention, the amplitude of ± 1st order diffracted light is larger than that of 0th order light and is detected brightly. The ± 1st order diffracted image includes 0th order light, and more precisely, the −1st order diffracted image is the sum of the distribution of the 0th order light and the −1st order diffracted light corresponding to the + 1st order diffracted light. The same applies to the + 1st order diffraction image.
In addition, since the corresponding higher-order diffracted light is outside the NA of the objective lens 20, the 0th-order light reduction unit is not detected in the diffraction image at the exit pupil 19 of the objective lens 20 (the diffraction angle of this ± first-order diffracted light is Determined by wavelength and pattern width). Therefore, since the light of the 0th-order light reducing portion is a low-frequency component that does not contribute to the formation of the image of the sample 1, high resolution due to the reduction can be explained. For this reason, in conventional illumination, the line and space contrast is low (C = 0.028) and the line and space separation is difficult, but in the present invention, the contrast is improved to 0.178 and the line and space separation is improved. It is clear.
As a result, for example, when a detection optical system having a detection size of one pixel of 0.2 μm is used, the conventional illumination can obtain a resolution of only about 0.15 μm. The resolution could be increased to about 0.1 μm, half the size. Thereby, according to the present invention, it has become clear that high resolution can be achieved.

FIG. 3 shows an embodiment of an optical inspection apparatus using the optical system of the present invention. The sample 1 is vacuum-sucked on a chuck 2, and the chuck 2 is mounted on a θ stage 3, a Z stage 4, a Y stage 5, and an X stage 6. The optical system 111 disposed above the sample 1 detects an optical image of the sample 1 in order to inspect the appearance of the pattern formed on the sample 1, and mainly includes the illumination optical system and the sample 1. A detection optical system that captures an image and a focus detection optical system 45 are included. The light source 8 disposed in the illumination optical system is an incoherent light source, for example, a xenon lamp.
The light emitted from the light source 8 passes through the aperture of the aperture stop 11 through the lens 9, and the light that has passed through the wavelength selection filter 12 through the lens reaches the field stop 13. The wavelength selection filter 12 considers the spectral reflectance of the sample 1 and limits the illumination wavelength region in order to detect an image of the sample 1 with high resolution. For example, an interference filter is disposed. The light transmitted through the field stop 13 passes through the polarizing plate 14 and becomes linearly polarized light. The polarized light passes through the half mirror 15 and enters the objective lens 20 to illuminate the sample 1.
The light that illuminates the sample 1 is reflected, scattered, and diffracted on the sample 1, and the light within the NA of the objective lens 20 enters the objective lens 20 again, is reflected by the half mirror 15, and enters the polarizing plate 22. The polarizing plate 22 is positioned in the θ direction so that the high-order diffracted light has a relatively high transmittance with respect to the zero-order light, and takes an image of the sample 1 through the imaging lens 30 and the zoom lens 50. An image of the sample 1 is formed on the light receiving surface of the image sensor 70. The image sensor 70 is a linear sensor, a TDI image sensor, a TV camera, or the like. Further, in order to perform automatic focusing, it is necessary to guide the focus detection light to the focus detection optical system 45, and this light splitting means 25 is, for example, a dichroic mirror.
As the focus detection light, an optical image having the height information of the sample 1 is formed on the sensor 41 by the imaging lens 40, and the sensor output signal is input to the focus detection signal processing circuit 90, and this focus detection signal processing is performed. The circuit 90 detects the amount of deviation between the height of the sample 1 and the focal position of the objective lens 20, and sends data on the amount of focal deviation to the CPU 75. The CPU 75 issues a command for driving the Z stage 4 from the CPU 75 to the stage control unit 80 in accordance with the amount of focus shift, and sends a predetermined pulse from the stage control unit 80 to the Z stage 4 so that the automatic focus function is activated.
Further, the image data of the optical image of the sample 1 detected by the photoelectric conversion element 70 of the detection optical system is input to the image processing circuit 71 to be stored and processed, and the defect determination circuit 72 determines the defective portion, The result is displayed on a display means such as a display, and transmitted to an external storage / control device such as a workstation or a data server via a communication means.
As a specific processing method of a series of image processing from the photoelectric conversion element 70 for determining the defective portion from the detected image data to the defect determination circuit 72, for example, Japanese Patent Laid-Open No. 2-170279 or Japanese Patent Laid-Open No. As described in Japanese Patent No. 33605 and the like, a method of comparing image data corresponding to adjacent chips, a method of comparing image data corresponding to adjacent chips, and image data of adjacent patterns There are a method of comparing, a method of comparing design data and image data, and the like.
The movement of the sample 1 in the XY directions is performed by two-dimensionally controlling the movement of the X stage 6 and the Y stage 5 by the stage controller 80. The θ stage 3 is used when performing θ alignment of the movement direction of the XY stages 6 and 5 and the pattern formed on the sample 1.
In the present invention, an epi-illumination type optical system is an embodiment, but the resolution R of this optical system is generally determined by R = λ / (2NA). However, it has been clarified in FIG. 2 that an optical system using linearly polarized light can obtain a resolution higher than R obtained previously.
FIG. 4 illustrates another embodiment of the present invention. In this embodiment, the same effect as that shown in FIGS. 1 and 2 can be obtained.
The light emitted from the light source 8 passes through the aperture stop 11 through the lens 9, passes through the wavelength selection filter 12 and the field stop 13, and enters the polarization beam splitter 15. The P-polarized light transmitted through the polarization beam splitter 15 is used as illumination light, and the light transmitted through the λ / 2 plate 16 and the λ / 4 plate 17 that gives a phase difference to the P-polarized light illuminates the sample 1 through the objective lens 20. Here, by rotating the λ / 2 plate 16, the vibration direction of polarized light is set to a predetermined angle with respect to the pattern (in the example shown in FIG. 1, it is 90 °, but it is not necessarily 90 °). Absent). The λ / 4 plate is a phase difference plate that converts linearly polarized light into elliptically polarized light. The detection light that illuminates the sample 1 through the objective lens 20 is transmitted again through the λ / 2 plate 16 and the λ / 4 plate 17. This is to reflect the polarized beam splitter 20 and guide it to the detection optical path.
In the present embodiment, the phase difference due to the reflection of the sample 1 is not taken into consideration. However, when the λ / 4 plate 17 is not provided, the light that has been specularly reflected (0th order light) through the sample 1 is transmitted by the polarization beam splitter, and thus the detection optical path. Only the high-order diffracted light is guided to. However, since a slight phase difference is actually received by reflecting the sample 1, leakage light of the 0th order light is detected.
Therefore, when the polarizing beam splitter 15 is used as an alternative to the polarizing plate 14 and the polarizing plate 22 shown in FIG. 1, a λ / 4 plate is used to guide zero-order light equivalent to the amplitude of high-order diffracted light to the detection optical path. It is necessary to set the ellipticity of the polarized light that illuminates the sample 1 using 17 to an appropriate value. When the crystal optical axis of the λ / 4 plate 17 and the vibration direction of the linearly polarized light of the illumination light are 45 °, the illumination light transmitted through the λ / 4 plate 17 is circularly polarized.
In this case, the 0th-order light reflected from the sample 1 and reciprocating along the λ / 4 plate 17 becomes S-polarized light, and is reflected by the polarization beam splitter 15 and guided to the detection optical path. Accordingly, the amount of light guided to the detection optical path increases. This is because the amount of zero-order light, which is a low-frequency component, increases. The contrast of the optical image is comparable to that of a conventional microscope, and the resolution of the optical image is not improved. For this reason, the angle formed by the optical axis direction of the crystal of the λ / 4 plate 17 and the vibration direction of the linearly polarized light varies depending on the pattern formed on the sample 1 or the like. Therefore, it is preferable that the structure be rotatable so that the phase difference by the λ / 4 plate 17 (the ellipticity of polarized light illuminating the sample 1) can be changed.
As a result, a part of the light reflected and diffracted from the sample 1 is transmitted again through the λ / 2 plate 17 and the λ / 4 plate 17, reflected from the polarization beam splitter 15, and the optical image of the sample 1 on the image sensor 70. Form. Here, by adopting a configuration in which the λ / 2 plate 16 and the λ / 4 plate 17 can be rotated by an electric motor, the ellipticity of polarized light illuminating the sample 1 and the direction of the major axis thereof can be controlled.
Further, by creating a database of the ellipticity of polarization and the direction of the major axis as the number of pulses of the motor, automatic setting of the polarization state for realizing high contrast of the optical image is facilitated. Further, when illumination light having a wide wavelength width is used, it is desirable that the polarization beam splitter 15, the λ / 2 plate 16, and the λ / 4 plate 17 have less change in characteristics due to wavelength. This is because, when the phase difference is added to the wavelength using the λ / 2 plate 16 and the λ / 4 plate 17, if the unevenness of the phase difference due to the wavelength is large, a specific wavelength region is guided to the detection system. This is because the colors are emphasized. Therefore, even when broadband wavelength illumination is performed, when a transparent film is formed on the sample 1, uneven brightness due to thin film interference is emphasized.

FIG. 5 shows a defect detection method using illumination light having a wavelength of about 200 to 250 nm or less. Since the resolution is improved by shortening the wavelength, the use of ultraviolet light or DUV (Deep Ultra Violet) light further increases the resolution.
In the optical system of FIG. 5A, the light emitted from the ultraviolet light source 8 passes through the illumination optical system 112 and enters the beam splitter 15. The P-polarized light transmitted therethrough is subjected to a phase difference by the λ / 2 plate 16 and the λ / 4 plate 17, and illuminates the sample 1 through the objective lens 20. The incident light is ultraviolet light for bright field detection and focus detection light for detecting the height of the sample 1. As the focus detection light, light such as visible light that is not absorbed by the planarizing film is effective. This is because the height cannot be detected when light is absorbed by a planarizing film or the like. This focus detection wavelength region is determined by the spectral reflectance (R) of the dichroic mirror 25. As an example of this, in FIG. 5B, light of 650 nm or more is used as a focus detection wavelength region.
An optical path for detecting a bright field image by ultraviolet rays is an optical path reflected from the second dichroic mirror 26, and an image is detected on the surface of the sensor 70a by the imaging lens 30a. This ultraviolet wavelength range is determined by the spectral reflectance of the second dichroic mirror and is shown in FIG.
Further, when the sample 1 is irradiated with ultraviolet rays, fluorescence is generated depending on the material. By detecting this fluorescent image, it is possible to detect defects that cannot be detected with a normal bright-field image. This optical system can be put to practical use by constructing an intermediate image of light transmitted through two dichroic mirrors using an imaging lens and enlarging and projecting this image onto the image sensor 70 using a zoom lens 50. In the ultraviolet objective lens, aberration correction is performed with ultraviolet rays, but it is necessary to select focus detection light with relatively little ultraviolet rays and aberrations.

FIG. 6 illustrates a method for setting optical parameters on an “inspection recipe screen” 915 in a pattern defect inspection apparatus in a semiconductor manufacturing process.
First, the ID of a wafer to be inspected is registered on the “inspection recipe screen” 915, and the type of the wafer is input. As a result, it becomes clear what defects are located in which process of each wafer. Furthermore, the area to be inspected is designated by, for example, the coordinates of the wafer. Furthermore, the defect inspection acquires an image of the pattern formed on the wafer and detects it using the acquired image. However, as an inspection method, the images of adjacent chips are compared to determine a mismatched portion as a defect. There are a detection method and a method of comparing design data with an image. This method is selected in the “inspection method” on the “inspection recipe screen” 915.
Next, in the “image processing parameter” section on the inspection recipe screen 915, a threshold value of a difference image that determines image alignment and defect detection sensitivity is input.
Furthermore, by selecting “optical parameters” on the inspection recipe screen 915, an “optical parameter setting screen” 916 is displayed on the screen, and various defect detection sensitivities can be set by changing the optical parameters. It has become. In this “optical parameter setting screen” 916, 1. Polarization characteristics 2. Wavelength Illumination σ 4. At least one of the spatial filters can be set.
1. In the polarization characteristics, the polarizations of the illumination optical system and the detection optical system are determined by numbers, for example. If the numbers of the polarization characteristics that increase the defect detection sensitivity are known, the numbers are input. However, in the initial inspection of the wafer, since the polarization characteristic that increases the defect detection sensitivity is not yet determined, a polarization characteristic that is advantageous for improving the defect detection sensitivity is selected based on the shape of the pattern. The relationship between the shape of the pattern and the polarization characteristics can be obtained from the relationship as shown in FIG.
In addition, 2. This wavelength is used to select an illumination wavelength region in which the defect contrast is high. For example, a wavelength region in which the reflectance of the pattern to be inspected is high is selected. Alternatively, a wavelength region in which the difference in brightness between the pattern and the pattern background (region without the pattern) is large is selected. Further, in a wafer that has been subjected to CMP (Chemical Mechanical Polishing) processing, if the film thickness is uneven, the brightness of the detected image is uneven due to the thin film interference of the insulating film. Since the film thickness unevenness is not a defect, the difference in brightness becomes a noise in defect inspection. In order to reduce the difference in brightness, it is effective to widen the illumination wavelength width. However, the extent to which the wavelength width is widened depends on the design film thickness of the insulating film and its error, so it is effective to obtain the illumination wavelength width from the film thickness of the insulating film. Therefore, the thickness of the insulating film is input as a condition for determining the illumination wavelength range.
Next, 3. The illumination σ is used to select the aperture diameter of the aperture stop 11 of the illumination system, and when a hole-like pattern is formed, the defect detection is preferably made relatively smaller than the illumination σ of the line-shaped pattern. Sensitivity is improved.
4. This spatial filter reduces the amplitude transmittance of zero-order light or gives a phase difference to the Fourier transform plane (exit pupil position of the objective lens or a position conjugate with this exit pupil (pupil position of the zoom lens 50)). If the optical filter is arranged according to the pattern shape or density, the resolution of the image can be increased.
By setting the optical parameters as described above, an image advantageous for defect detection can be detected. However, whether the set condition is good or not depends on the shape of the pattern and the structure of the wafer.
As shown in FIG. 7, patterns are regularly arranged on the wafer 1 in units of exposure fields (shots) of the exposure apparatus. In FIG. 7B, which is an enlarged view of 1 ′ which is a part of this pattern, there are patterns 1a and 1a ′ which are assumed to be inspected, a pattern 1c formed in the previous step, and the like. FIG. 7C shows a cross-sectional structure when the wafer 1 is cut along the line 125 in FIG. It is assumed that the inspection target patterns 1a and 1a ′ are formed on the insulating film 1b, and the pattern 1c that is not the inspection target is formed in the insulating film 1b. In this case, since the main direction of the inspection target patterns 1a and 1b is vertical, the defect detection sensitivity can be improved by setting the optical parameters so that the pattern in this direction has high contrast. However, since the optimum value of the optical parameter varies depending on the shape and structure of the pattern, a large amount of time is required for determination.
FIG. 8 shows a flow of highly efficient optical parameter setting.
First, the wafer is loaded, and the inspection target area is moved to the visual field of the optical system. Next, a parameter range for acquiring images preliminarily for optical parameters 1) polarization characteristics 2) illumination σ 3) illumination wavelength 4) spatial filter is input. Next, images are acquired under the conditions of the set range and displayed so that diffraction images and pattern images can be listed, and the sum of secondary differential values indicating the contrast of the pattern and the sharpness of the pattern are displayed.
As a result, information on the optical parameters and the resolution of the optical image can be listed, and the appropriate values of the optical parameters can be easily determined. However, the final optical parameter is determined by whether or not the image is advantageous for defect detection. For this reason, it is necessary to narrow down the range of the optical parameter from the sum of the previous contrast and the secondary differential value, and finally perform preliminary inspection while changing the optical parameter to determine the optical parameter. The contents displayed as the preliminary inspection result include, for example, the amount of difference (average, maximum, deviation) of the difference image in the normal part and the comparison inspection result (the number of detected defects, the signal level of the defect, the maximum amount of mismatch N in the normal part, and the defect) The optical parameters are determined from the displayed results.
With the above optical parameter setting means, it is possible to efficiently set an optical parameter with high defect detection sensitivity.
FIG. 9 shows a list display 920 when the polarization state is used as a parameter as an example of a list display of diffraction images / pattern images. First, the polarization vibration direction may be defined with respect to the direction of the XY stage or the orientation flat of the wafer. For example, the vibration direction of illumination light: 90 ° and the vibration direction of detection light 45 ° are angles formed by the main direction of the pattern to be inspected and the vibration direction of the principal ray on the object.
By displaying in the list in this way, it is possible to predict the ratio of the amplitudes of the high-order diffracted light and the zero-order light from the data of the diffraction image. When the inspection target pattern is in various directions, the pattern direction and contrast can be detected from the pattern image to be inspected. Further, by displaying the light intensity distribution of the image as a line graph or the like, it becomes easy to recognize the brightness level and the like. Furthermore, from the sum of the contrast and the second derivative value, it can be easily read that the oscillation direction of the polarization of the illumination light has a high resolution near θ: 90 ° and 68 °. Therefore, by performing the preliminary inspection to be performed next under these conditions, the optical conditions that enable highly sensitive inspection can be set in a short time.
In the configuration of the optical system shown in FIGS. 4 and 5, the ellipticity of the polarized light that illuminates the sample 1 and the direction of the major axis are the parameters of the polarization state. Accordingly, it is possible to display a list similar to that in FIG. 9 using the ellipticity and the direction of the major axis as parameters, and it is possible to easily specify an approximate polarization state that can realize a high-sensitivity inspection.

FIG. 10A shows a means for detecting a bright field image and a dark field image using polarized light, and a neutral image of the bright and dark field image. The illumination light 180 is randomly polarized light. The illumination light 180 incident on the polarization beam splitter 15 transmits only the P-polarized component light. The P-polarized light is transmitted through the λ / 2 plate 16 and the λ / 4 plate 17 to change the rotation of the polarization in the vibration direction and the ellipticity.
The illumination light transmitted through the λ / 2 plate 16 and the λ / 4 plate 17 and further transmitted through the objective lens 20 is reflected by the sample 1 to give a phase difference according to the pattern of the sample along with the phase jump. The polarization state changes. Further, the light diffracted at the edge of the pattern also changes its polarization according to the direction of the diffracted light. These reflected lights are captured again by the objective lens 20, pass through the λ / 4 plate 17 and the λ / 2 plate 16, enter the PBS, and the S-polarized component is reflected to form a bright field image on the sensor 70. Image.
Here, in the case of normal bright field detection without using the λ / 2 plate 16 and the λ / 4 plate 17, if the detected image is a grayscale image as shown in FIG. As shown in FIG. 10C, the brightness distribution of the portion -A is a dark image in the dense pattern portion on the left side of the image. In this case, if the amount of light is increased, the density difference in the dense pattern portion can be increased. However, the brightness of the flat portion on the right side of the image is further increased, and in the CCD sensor 70 or the like, the sensor 70 is saturated and blooming occurs. It will occur. For this reason, the amount of light cannot be increased, and it is very difficult to detect a dense pattern with good contrast.
However, if the phase difference amount is adjusted using the λ / 2 plate 16 and the λ / 4 plate 17 as in the present invention, the light of the flat portion reflected by the polarization beam splitter is reduced and the dense pattern portion is reduced. The ratio at which the diffracted light is reflected can be increased. As a result, the brightness of the dense pattern and the flat portion can be detected at the same level. Therefore, an image having a high contrast can always be detected regardless of the pattern density and the like, and pattern defects and the like in the dense pattern portion can be easily detected.

In the configuration of the optical system shown in FIGS. 3, 4, 5, and 10, linearly polarized illumination light passes through the quarter-wave plate 17 and becomes elliptically polarized light. The elliptically polarized light is illuminated onto the sample 1 through the objective lens 20. The light reflected and diffracted by the sample 1 is captured by the objective lens 20 and is transmitted through the quarter-wave plate 17 again to receive a phase difference. By rotating the quarter wavelength plate 17, the ellipticity of elliptically polarized light that illuminates the sample 1 can be changed.
FIG. 11 shows the relationship between the contrast of the optical image of the sample 1 detected by the image sensor 70 and the detected light amount ratio when the ellipticity is changed between −1.0 and 1.0. Note that the sign of ellipticity distinguishes clockwise and counterclockwise polarized light. The contrast was obtained from a line and space image formed on the surface of the sample 1. Furthermore, when the illuminance of the illumination light is constant, the amount of light reaching the image sensor 70 decreases as the absolute value of the ellipticity decreases. For this reason, the detected light amount ratio when the ellipticity is changed is shown with reference to the light amount detected by the image sensor 70 when the elliptical polarization is -1.0 or 1.0.
As can be seen from FIG. 11, the contrast is maximized when the ellipticity is in the vicinity of −0.18, and the point where the detected light quantity ratio is maximized, that is, the point where the ellipticity is 1 or −1 is one. I have not done it. Therefore, in order to perform a line and space defect inspection, it is advantageous to set the ellipticity in the vicinity of −0.18.

  Patterns having various shapes are formed on the wafer to be inspected. In this case, it is efficient to perform a preliminary inspection in determining the ellipticity of polarized light and the direction of the major axis that improve the inspection sensitivity. For example, as shown in FIG. 12, an “illumination condition setting mode” is provided in the “inspection recipe creation mode” of the inspection target wafer.

In this “illumination condition setting mode”, 1) a “wavelength filter” selection button for selecting the wavelength width of illumination light, 2) an “aperture stop” selection button for changing the illumination incident angle range, and 3) a thin film is laminated. “Focus offset” value setting button for determining which layer to focus on in case 4) A button for setting the polarization state of illumination light (the ellipticity of polarization and its direction) in order to change the “pattern enhancement” Prepare.
These buttons are set to pre-registered conditions by designating numbers (No in the figure). The polarization condition includes a database of polarization conditions that increase the contrast of the detected image in accordance with the shapes and orientations of various inspection target patterns. Therefore, a wafer to be inspected is preliminarily inspected under all or a part of the polarization conditions in the database, and the polarization state from which the inspection result with the highest detection sensitivity is obtained from the preliminary inspection results is obtained. A method of setting the polarization condition at the time of inspection is conceivable. When a preliminary inspection is performed, the inspection is executed under the conditions 1) to 4) set provisionally by selecting a “preliminary inspection button” displayed on the screen.
The polarization condition may be the number of pulses of a motor that electrically rotates the λ / 2 plate 16 and the λ / 4 plate 17 in the optical systems shown in FIGS. 4, 5, and 10. Furthermore, from the results of preliminary inspection under several types of polarization conditions, the following determination standards can be cited as the determination standards for determining the polarization conditions used in this inspection.
1) Polarization condition that detects the most defects 2) Polarization condition that detects the most fatal defects 3) If the coordinates of the defect to be detected are known, this defect is detected the most In accordance with the polarization condition, when a plurality of preliminary inspections are performed, determination is facilitated by displaying a list of judgment criteria as shown in the above 1) to 3). Further, by automatically determining the polarization condition in which any one of the values 1) to 3) is maximized, it is easy for the apparatus user to determine the conditions.
As described above, in determining the polarization condition, it is actually desirable to review the defect portion with an optical microscope or an electron microscope and determine the polarization condition with high inspection sensitivity.

It is a figure explaining the polarization state of the 0th-order light and high-order diffracted light by linearly polarized illumination. It is a figure which shows an example of high resolution using a linearly polarized light. It is a schematic front view showing an example of a schematic configuration of an inspection apparatus using the high resolution optical system of the present invention. It is a schematic front view which shows the other example of schematic structure of the test | inspection apparatus using the high resolution optical system of this invention. FIG. 2A is a schematic front view showing an example of a schematic configuration of an inspection apparatus using a short-wavelength illumination high-resolution optical system according to the present invention, and FIGS. 2B and 2C are diagrams showing spectral reflectances when the optical system is used. It is. It is a front view of the display screen which shows an example of the optical parameter setting means by this invention. It is a top view which shows an example of the test object pattern of this invention. It is a flowchart which shows an example of the optical parameter setting procedure by this invention. It is a figure which shows an example of the list of an exit pupil image, an optical image, and a resolution evaluation value. It is a figure which shows an example of the high contrast image detection by the polarization illumination by this invention. It is a figure which shows an example of the relationship between the ellipticity of polarization | polarized-light, and a pattern detection light quantity ratio. It is a figure which shows an example of the polarization condition setting method by this invention.

Explanation of symbols

  DESCRIPTION OF SYMBOLS 1 ... Sample, 1b ... Insulating film, 2 ... Chuck, 3 ... (theta) stage, 4 ... Z stage, 5 ... Y stage, 6 ... X stage, 8 ... Light source, 14 ... Polarizer, 22 ... Analyzer, 30 ... Imaging lens , 70 ... Image sensor, 300, 300 '... Direction of vibration of electric field vector in polarization plane of illumination light, 310, 310' ... Direction of vibration of electric field vector in polarization plane of zero-order light (regular reflection light), 320 ... High next time Direction of vibration of the electric field vector in the polarization plane of the folded light, 15: Polarizing beam splitter, 16 ... λ / 2 plate, 17 ... λ / 4 plate, 915 ... Inspection recipe screen, 916 ... Optical parameter setting screen, 920 ... Optical conditions and optics Image list screen, 921... Optical conditions and preliminary inspection result list screen.

Claims (11)

  A method of inspecting a defect of a sample having a pattern formed on a surface, wherein the sample is irradiated with polarized light through an objective lens, and reflected light from the surface of the sample by the irradiation is compared with the objective lens. A defect inspection characterized in that an image of the sample surface is obtained by detection through a phase difference plate and an analyzer, and the defect of the sample is detected by comparing the obtained image with a previously stored image. Method.   The retardation plate attenuates the amplitude of the zero-order diffracted light from the pattern formed on the surface out of the light reflected from the sample surface by the irradiation at a large ratio with respect to the amplitude of the high-order diffracted light. The defect inspection method according to claim 1.   A method for inspecting a defect of a sample having a pattern formed on a surface, wherein the sample is irradiated with polarized light in a state where an optical system is set to a first polarization condition, and reflected by the irradiation on the surface of the sample. A first image is obtained by detecting the detected light, the first image is displayed on a monitor screen, and the optical system is made to have a second polarization based on the first image displayed on the monitor screen. The sample is irradiated with polarized light in a state where the optical system is set to the second polarization state, and the second image is detected by detecting light reflected by the surface of the sample by the irradiation. A defect inspection method comprising: obtaining a defect of the sample by comparing the obtained second image with a previously stored image.   The first polarization condition and the second polarization condition each include at least one of a polarization characteristic, a wavelength of polarized light, a numerical aperture of an aperture stop of an illumination system, and a shape of a spatial filter. The defect inspection method according to claim 3.   In the second polarization state, the amplitude of the 0th-order diffracted light from the pattern formed on the surface of the light reflected from the sample surface by the irradiation is attenuated at a large ratio with respect to the amplitude of the high-order diffracted light. The defect inspection method according to claim 3, wherein:   A method for inspecting a defect of a sample having a pattern formed on a surface, wherein the sample is irradiated with polarized light through an objective lens, reflected by the surface of the sample by the irradiation, and passed through the objective lens Defect inspection characterized in that light is detected by changing the state of polarization to obtain an image of the sample surface, and the obtained image is compared with a previously stored image to detect defects in the sample. Method.   The state of polarization of light that has passed through the objective lens is changed by transmitting light that has passed through the objective lens through a half-wave plate and a quarter-wave plate. 6. The defect inspection method according to 6.   An apparatus for inspecting a defect of a sample having a pattern formed on a surface, the illumination unit irradiating the sample with polarized light via an objective lens, and the illumination unit irradiating the sample with the polarized light Polarized image detecting means for detecting the reflected light reflected by the surface of the sample through the objective lens and a polarization unit that changes the polarization state of the reflected light to obtain an image of the sample, and the polarized image A defect detection means for detecting a defect of the sample by comparing an image obtained by the detection means with an image stored in advance, and a display means for displaying a result detected by the defect detection means Defect inspection equipment.   The defect inspection apparatus according to claim 8, wherein the polarizing unit includes a retardation plate and an analyzer.   The retardation plate has a half-wave plate and a quarter-wave plate, and the amplitude of the zero-order diffracted light from the pattern formed on the surface out of the light reflected from the sample surface is a high-order diffracted light. The defect inspection apparatus according to claim 8, wherein the defect is attenuated at a large rate with respect to the amplitude of the defect.   9. The defect inspection apparatus according to claim 8, further comprising optical parameter setting means for setting optical parameters such as polarization characteristics, wavelength, illumination σ, and spatial filter of the illumination means and the polarization image detection means.

Images