JP4853758B2 - Surface inspection apparatus and surface inspection method - Google Patents

Description

  The present invention relates to a surface inspection method and apparatus for detecting defects such as unevenness and scratches on a substrate surface in the manufacturing process of semiconductor elements and the like.

2. Description of the Related Art Conventionally, a device using diffraction is known as an inspection device for a defect of a repeated pattern formed on the surface of a wafer in a manufacturing process of a semiconductor circuit element or the like. In an apparatus using diffraction, it is necessary to adjust the tilt angle of the stage according to the pattern pitch. In order to cope with finer patterns, it is necessary to shorten the wavelength of illumination light.
Japanese Patent Laid-Open No. 10-232122

  However, in order to cope with repetitive pitch miniaturization (that is, line and space miniaturization of wiring patterns, etc.), when trying to shorten the wavelength of illumination light, the types of light sources are limited, which is expensive and large. It becomes a light source. Further, the material of the optical element constituting the illumination system and the light receiving system is also limited to an expensive material, which is not preferable.

  An object of the present invention is to provide a surface inspection apparatus and a surface inspection method that can reliably cope with repeated pitch miniaturization without shortening the wavelength of illumination light.

In order to solve the above problems, a surface inspection apparatus according to the present invention includes a light source unit that emits a linearly polarized divergent light beam for illuminating a substrate to be inspected on which a repeated pattern is formed, and a space between the light source and the surface to be tested. An optical member for guiding the principal ray of the linearly polarized divergent light beam to a position shifted from the optical axis, and guiding the linearly polarized divergent light beam to the substrate to be inspected, and the inspection A light receiving means for receiving linearly polarized light whose polarization direction is orthogonal to the linearly polarized light out of the light flux from the substrate, and the optical member for the principal ray of the divergent light flux in an optical path between the light source means and the light receiving means. And a polarization correction member comprising at least one parallel plate that is arranged to be inclined in a direction opposite to the inclination direction of the optical axis and corrects the disturbance of the polarization plane caused by the optical member, and the light receiving means To the light received by And it performs inspection of the surface of the inspection substrate Zui.

The surface inspection apparatus according to claim 2 includes: a light source unit that emits a linearly polarized light beam for illuminating a substrate to be inspected on which a repeated pattern is formed; and a position at which the light beam from the substrate to be inspected is incident. An optical system in which the principal ray of the light beam from the substrate is incident at a position shifted from the optical axis, and the light beam from the substrate to be inspected is emitted as a convergent light beam having a predetermined emission angle. A light receiving means for receiving linearly polarized light orthogonal to the predetermined linearly polarized light among the convergent light flux from the member and the optical member, and a main part of the convergent light flux in the optical path between the light source means and the light receiving means. A polarization correction member comprising at least one parallel flat plate arranged to be tilted in the direction opposite to the tilt direction of the optical member with respect to the light beam and correcting the disturbance of the polarization plane caused by the optical member. And said And performs inspection of the surface of the inspection substrate based on light received by the optical means.

The surface inspection apparatus according to claim 3 is disposed between a light source unit that emits a linearly polarized divergent light beam for illuminating a substrate to be inspected on which a repetitive pattern is formed, and between the light source and the test surface. The linearly polarized divergent light beam is incident on a position where the principal ray of the light beam is shifted from the optical axis, and the linearly polarized divergent light beam is guided to the substrate to be inspected, and the substrate to be inspected. The light beam from the substrate to be inspected is incident at a position where the principal ray of the light beam is shifted with respect to the optical axis, and the convergent light beam is emitted at a predetermined emission angle to the predetermined surface. A second optical member that forms an image; an extraction unit that extracts linearly polarized light orthogonal to the linearly polarized light out of a convergent light beam from the second optical member; and the second optical member and the extracting unit. Light reception for receiving an image of the inspection substrate formed through In the optical path between the light source means and the light receiving means, the direction of inclination of the optical member with respect to the principal ray of the divergent light beam or the direction opposite to the direction of inclination of the optical member with respect to the principal ray of the convergent light beam And a polarization correction member comprising at least one parallel plate that corrects disturbance of the polarization plane of the light beam caused by the first and second optical members.

According to a fourth aspect of the present invention, in the surface inspection apparatus according to the first or second aspect, the optical material includes a concave mirror .

According to a fifth aspect of the present invention, in the surface inspection apparatus according to any one of the first to fourth aspects, at least one of a tilt direction and a tilt angle of the polarization correction member is held to be adjustable. It further has a holding means.

According to a sixth aspect of the present invention, in the surface inspection apparatus according to any one of the first to fifth aspects, the polarization correction member is arranged to be inclined with respect to the surface of the optical member. It is a parallel plate.

The invention according to claim 7 is the surface inspection apparatus according to any one of claims 1 to 5 , wherein the polarization correction member is inclined with respect to a plane perpendicular to the optical axis of the optical member. These are parallel flat plates of two birefringent crystals that are arranged so that their crystal axes are orthogonal to each other.

According to an eighth aspect of the present invention, in the surface inspection apparatus according to any one of the first to fifth aspects, the polarization correction member is inclined with respect to a plane perpendicular to the optical axis of the optical member. Thus, two wedge-shaped birefringent crystals that are arranged so that their crystal axes are orthogonal to each other and are bonded to form a parallel plate.

The polarized illuminating device according to claim 9, wherein the principal ray of the linearly polarized divergent beam is deviated from the optical axis between the light source means for emitting the linearly polarized divergent beam and the light source and the test surface. An optical member that guides the linearly polarized divergent beam to the substrate to be inspected and an optical path between the light source means and the substrate to be inspected with respect to the principal ray of the divergent beam. Having a polarization correction member comprising at least one parallel plate arranged to be tilted in the direction opposite to the tilt direction of the optical member and correcting disturbance of the polarization plane of the light beam caused by the optical member It is.

The light receiving device according to claim 10 is arranged such that a principal ray of a light beam from the substrate to be inspected is incident at a position shifted from an optical axis at a position where the light beam from the substrate to be inspected is incident, An optical member for injecting a light beam having a predetermined polarization component from the substrate to be inspected and emitting the converged light beam with a predetermined emission angle; and a light receiving means for receiving linearly polarized light out of the light beam from the optical member; In the optical path between the substrate to be inspected and the light receiving means, the optical member is disposed in an inclined direction opposite to the inclination direction of the optical member with respect to the principal ray of the convergent light beam, and is generated due to the optical member. It has a polarization correction member composed of at least one parallel plate that corrects the disturbance of the polarization plane of the light beam.

  According to the present invention, it is possible to reliably cope with repeated pitch miniaturization without shortening the wavelength of illumination light.

Hereinafter, the principle of the surface inspection apparatus using polarized light according to the present invention will be described in detail with reference to the drawings.
As shown in FIG. 1, the surface inspection apparatus 30 of the present invention includes a stage 11 that supports a semiconductor wafer 20 that is a substrate to be tested, an alignment system 12, an illumination system 13, a light receiving system 14, and an image processing apparatus 15. It consists of and. The surface inspection apparatus 30 is an apparatus that automatically inspects the surface of the semiconductor wafer 20 in the manufacturing process of the semiconductor circuit element. After exposure / development of the uppermost resist film, the semiconductor wafer 20 is carried from a wafer cassette (not shown) or a developing device by a conveyance system (not shown) and is attracted to the stage 11.

  As shown in FIG. 2, a plurality of chip regions 21 are arranged in the XY direction on the surface of the semiconductor wafer 20, and a repeated pattern 22 is formed in each chip region 21. As shown in FIG. 3, the repetitive pattern 22 is a resist pattern (for example, a wiring pattern) in which a plurality of line portions 2A are arranged at a constant pitch P along the short direction (X direction). A space 2B is formed between the adjacent line portions 2A. The arrangement direction (X direction) of the line portions 2A is referred to as “repeating direction of the repeating pattern 22”.

Here, the design value of the line width D _A of the line portion 2A of the repeating pattern 22 and 1/2 of the pitch P. When the repeated pattern 22 is formed according to the design value, the line width D _{A of the} line portion 2A
The line width D _B of the space portion 2B is equal, the volume ratio of the line portion 2A and the space portion 2B is substantially 1: 1. In contrast, when the exposure focus at the time of forming the repeating pattern 22 deviates from a proper value, the pitch P does not change, the line width D _A of the line portion 2A becomes different from the designed value, the space portion 2B It becomes different even with the line width D _B, the volume ratio of the line portion 2A and the space portion 2B is substantially 1: deviates from 1.

  The surface inspection apparatus 30 of the present invention performs defect inspection of the repetitive pattern 22 by using the change in the volume ratio between the line portion 2A and the space portion 2B in the repetitive pattern 22 as described above. In order to simplify the explanation, the ideal volume ratio (design value) is 1: 1. The change in the volume ratio is caused by deviation from the appropriate value of the exposure focus, and appears for each shot area of the semiconductor wafer 20. The volume ratio can also be referred to as the area ratio of the cross-sectional shape.

  Further, it is assumed that the pitch P of the repeating pattern 22 is sufficiently small as compared with the wavelength of illumination light (described later) for the repeating pattern 22. For this reason, diffracted light is not generated from the repetitive pattern 22, and defect inspection of the repetitive pattern 22 cannot be performed by diffracted light. The principle of defect inspection according to the present invention will be described below together with the configuration of the surface inspection apparatus 30 (FIG. 1).

The stage 11 of the surface inspection apparatus 30 places the semiconductor wafer 20 on the upper surface, and fixes and holds it, for example, by vacuum suction. Further, the stage 11 is rotatable about the normal 1A at the center of the upper surface. With this rotating mechanism, the repeating direction (X direction in FIGS. 2 and 3) of the repeating pattern 22 of the semiconductor wafer 20 can be rotated within the surface of the semiconductor wafer 20. The stage 11 has a horizontal upper surface and does not have a tilt mechanism. For this reason, the semiconductor wafer 20 can always be kept in a horizontal state.

The alignment system 12 illuminates the outer edge of the semiconductor wafer 20 when the stage 11 is rotating, detects the position in the rotation direction of an external reference (for example, a notch) provided on the outer edge, and moves the stage at a predetermined position. 11 is stopped. As a result, the repetitive direction (X direction in FIGS. 2 and 3) of the repetitive pattern 22 of the semiconductor wafer 20 is changed to an illumination light incident surface 3A described later (see FIG. 4).
In contrast, it can be set at an angle of 45 degrees.

  The illumination system 13 is a decentered optical system composed of a light source 31, a wavelength selection filter 32, a light guide fiber 33, a polarizing plate 34, and a concave reflecting mirror 35, and linearly repeats the repetitive pattern 22 of the semiconductor wafer 20 on the stage 11. Illuminate with polarized light L1. This linearly polarized light L1 is illumination light for the repeated pattern 22. The linearly polarized light L <b> 1 is irradiated on the entire surface of the semiconductor wafer 20.

The traveling direction of the linearly polarized light L1 (the direction of the principal ray of the linearly polarized light L1 reaching an arbitrary point on the surface of the semiconductor wafer 20) is substantially parallel to the optical axis O1 of the concave reflecting mirror 35. The optical axis O1 passes through the center of the stage 11 and is inclined by a predetermined angle θ with respect to the normal line 1A of the stage 11. Incidentally, a plane including the traveling direction of the linearly polarized light L1 and parallel to the normal 1A of the stage 11 is an incident surface of the linearly polarized light L1. An incident surface 3 </ b> A in FIG. 4 is an incident surface at the center of the semiconductor wafer 20.

  In this description, for example, the linearly polarized light L1 is p-polarized light. That is, as shown in FIG. 5A, a plane including the traveling direction of the linearly polarized light L1 and the vibration direction of the electric vector (the vibrating surface of the linearly polarized light L1) is included in the incident surface (3A) of the linearly polarized light L1. It is. The vibration plane of the linearly polarized light L <b> 1 is defined by the transmission axis of the polarizing plate 34 disposed in front of the concave reflecting mirror 35.

  The light source 31 of the illumination system 13 is an inexpensive discharge light source such as a metal halide lamp or a mercury lamp. The wavelength selection filter 32 selectively transmits an emission line spectrum having a predetermined wavelength in the light from the light source 31. The light guide fiber 33 transmits the light from the wavelength selection filter 32. The polarizing plate 34 is disposed in the vicinity of the exit end of the light guide fiber 33, and its transmission axis is set to a predetermined direction, and the light from the light guide fiber 33 is linearly polarized according to the transmission axis. The concave reflecting mirror 35 is a reflecting mirror having a spherical inner surface as a reflecting surface, and is arranged so that the front focal point substantially coincides with the exit end of the light guide fiber 33 and the rear focal point substantially coincides with the surface of the semiconductor wafer 20. Then, the light from the polarizing plate 34 is guided to the surface of the semiconductor wafer 20. The illumination system 13 is an optical system that is telecentric with respect to the semiconductor wafer 20 side.

In the illumination system 13, the light from the light source 31 passes through the wavelength selection filter 32, the light guide fiber 33, the polarizing plate 34, and the concave reflecting mirror 35, and is p-polarized linearly polarized light L1 (FIG. 5 (
a)) and enters the entire surface of the semiconductor wafer 20. The incident angles of the linearly polarized light L1 at each point of the semiconductor wafer 20 are the same and correspond to the angle θ formed by the optical axis O1 and the normal line 1A.

  Since the linearly polarized light L1 incident on the semiconductor wafer 20 is p-polarized light (FIG. 5A), the repeating direction (X direction) of the repeated pattern 22 of the semiconductor wafer 20 is incident on the linearly polarized light L1 as shown in FIG. When the angle is set to 45 degrees with respect to (3A), the direction of the vibrating surface of the linearly polarized light L1 on the surface of the semiconductor wafer 20 (the V direction in FIG. 6) and the repeating direction (X direction) of the repeating pattern 22 Is also set to 45 degrees.

  In other words, the linearly polarized light L1 has a repeating pattern 22 in a state where the direction of the vibration surface (V direction in FIG. 6) on the surface of the semiconductor wafer 20 is inclined 45 degrees with respect to the repeating direction (X direction) of the repeating pattern 22. Is incident on the repetitive pattern 22 in a state of crossing diagonally.

  Such an angle state between the linearly polarized light L <b> 1 and the repeated pattern 22 is uniform over the entire surface of the semiconductor wafer 20. Note that the angle state between the linearly polarized light L1 and the repetitive pattern 22 is the same even if 45 degrees is replaced with any of 135 degrees, 225 degrees, and 315 degrees. The reason why the angle formed by the vibration surface direction (V direction) and the repeat direction (X direction) in FIG. 6 is set to 45 degrees is to maximize the sensitivity of the defect inspection of the repeat pattern 22.

When the repetitive pattern 22 is illuminated using the linearly polarized light L1, the elliptically polarized light L2 is generated from the repetitive pattern 22 in the regular reflection direction (FIGS. 1 and 5B). In this case, the traveling direction of the elliptically polarized light L2 coincides with the regular reflection direction. The regular reflection direction is a direction that is included in the incident surface (3A) of the linearly polarized light L1 and is inclined by an angle θ (an angle equal to the incident angle θ of the linearly polarized light L1) with respect to the normal 1A of the stage 11. As described above, since the pitch P of the repeated pattern 22 is sufficiently smaller than the illumination wavelength, no diffracted light is generated from the repeated pattern 22.

Here, the reason why the linearly polarized light L1 is ovalized by the repeated pattern 22 and the elliptically polarized light L2 is generated from the repeated pattern 22 will be briefly described. When the linearly polarized light L1 enters the repetitive pattern 22, the direction of the vibration surface (the V direction in FIG. 6) is divided into two polarization components V _X and V _Y shown in FIG. One polarization component V _X is a component parallel to the repetition direction (X direction). The other polarization component V _Y is a component perpendicular to the repetition direction (X direction). The two polarization components V _X and V _Y are independently subjected to different amplitude changes and phase changes. The reason why the amplitude change and the phase change are different is that the complex reflectivity (that is, the complex amplitude reflectivity) is different due to the anisotropy of the repetitive pattern 22, and this is called structural birefringence. As a result, the reflected lights of the two polarization components V _X and V _Y have different amplitudes and phases, and the reflected light obtained by combining them becomes elliptically polarized light L2 (FIG. 5B).

Further, the degree of ovalization due to the anisotropy of the repeated pattern 22 is the same as the vibration plane (incident plane (3A)) of the linearly polarized light L1 in FIG. 5A among the elliptically polarized light L2 in FIG. ) Can be considered as a polarized light component L3 (FIG. 5C). The magnitude of the polarization component L3 depends on the material and shape of the repetitive pattern 22 and the angle formed by the vibration plane direction (V direction) and the repetitive direction (X direction) in FIG. For this reason, when the angle formed between the V direction and the X direction is maintained at a constant value (for example, 45 degrees), even if the material of the repeated pattern 22 is constant, if the shape of the repeated pattern 22 changes, the degree of ovalization (The size of the polarization component L3) changes.

A relationship between the shape of the repeated pattern 22 and the size of the polarization component L3 will be described. As shown in FIG. 3, the repetitive pattern 22 has a concavo-convex shape in which the line portions 2A and the space portions 2B are alternately arranged along the X direction, and is formed according to the design value with an appropriate exposure focus. , line width D _B is equal to the line width D _a and the space portion 2B of the line portion 2A, the volume ratio of the line portion 2A and the space portion 2B is substantially 1: 1. In the case of such an ideal shape, the size of the polarization component L3 is the largest. In contrast, when the exposure focus deviates from an appropriate value, becomes different and the line width D _B of the line width D _A and the space portion 2B of the line portion 2A, substantially the volume ratio of the line portion 2A and the space portion 2B 1 : 1 off. At this time, the size of the polarization component L3 is smaller than the ideal case. The change in the magnitude of the polarization component L3 is illustrated in FIG. The horizontal axis in FIG. 8 is the line width D _A of the line portion 2A.

In this way, when the repetitive pattern 22 is illuminated using the linearly polarized light L1 in a state where the vibration plane direction (V direction) in FIG. 6 is inclined by 45 degrees with respect to the repetitive direction (X direction) of the repetitive pattern 22, The elliptically polarized light L2 (FIGS. 1 and 5 (b)) generated in the specular direction is the degree of ovalization (
The magnitude of the polarization component L3 in FIG. 5C corresponds to the shape of the repeated pattern 22 (volume ratio between the line portion 2A and the space portion 2B) (FIG. 8). The traveling direction of the elliptically polarized light L2 is included in the incident surface (3A) of the linearly polarized light L1, and is inclined with respect to the normal 1A of the stage 11 by an angle θ (an angle equal to the incident angle θ of the linearly polarized light L1).

Next, the light receiving system 14 will be described. As shown in FIG. 1, the light receiving system 14 is a decentered optical system including a concave reflecting mirror 36, an imaging lens 37, a polarizing plate 38, and an image sensor 39.
The concave reflecting mirror 36 is a reflecting mirror similar to the concave reflecting mirror 35 of the illumination system 13 described above, and its optical axis O2 passes through the center of the stage 11 and has an angle θ with respect to the normal 1A of the stage 11. It is arranged to tilt only. Therefore, the elliptically polarized light L <b> 2 from the repeated pattern 22 travels along the optical axis O <b> 2 of the concave reflecting mirror 36. The concave reflecting mirror 36 reflects the elliptically polarized light L <b> 2 and guides it toward the imaging lens 37, and collects it on the imaging surface of the imaging device 39 in cooperation with the imaging lens 37.

However, a polarizing plate 38 is disposed between the imaging lens 37 and the concave reflecting mirror 36. The direction of the transmission axis of the polarizing plate 38 is set so as to be orthogonal to the transmission axis of the polarizing plate 34 of the illumination system 13 described above (crossed Nicol state). Therefore, only the polarization component L4 (FIG. 1) corresponding to the polarization component L3 in FIG. 5C of the elliptically polarized light L2 can be extracted by the polarizing plate 38 and guided to the image sensor 39. As a result, a reflection image of the semiconductor wafer 20 by the polarization component L4 is formed on the imaging surface of the imaging element 39.

  The image sensor 39 is, for example, a CCD image sensor or the like, photoelectrically converts the reflected image of the semiconductor wafer 20 formed on the imaging surface, and outputs an image signal to the image processing device 15. The brightness and darkness of the reflected image of the semiconductor wafer 20 is substantially proportional to the light intensity of the polarization component L4 (the magnitude of the polarization component L3 in FIG. 5C), and the shape of the repeated pattern 22 (the line portion 2A and the space portion 2B). It changes according to the volume ratio (see FIG. 8). The reflected image of the semiconductor wafer 20 is brightest when the repeated pattern 22 has an ideal shape (volume ratio is 1: 1). Note that the brightness and darkness of the reflected image of the semiconductor wafer 20 appears for each shot area.

  The image processing device 15 captures a reflected image of the semiconductor wafer 20 based on the image signal output from the image sensor 39. Note that the image processing apparatus 15 stores a reflection image of a non-defective wafer in advance for comparison. A non-defective wafer is one in which the repetitive pattern 22 is formed on the entire surface in an ideal shape (volume ratio is 1: 1). It is considered that the luminance information of the reflected image of the non-defective wafer shows the highest luminance value.

  Therefore, when the image processing apparatus 15 captures the reflection image of the semiconductor wafer 20 that is the test substrate, the image processing apparatus 15 compares the luminance information with the luminance information of the reflection image of the non-defective wafer. Then, based on the amount of decrease in the luminance value of the dark portion of the reflected image of the semiconductor wafer 20 (the amount of decrease Δ in FIG. 8), the defect of the repeated pattern 22 (change in the volume ratio between the line portion 2A and the space portion 2B). Is detected. For example, if the amount of decrease in luminance value is larger than a predetermined threshold (allowable value), it is determined as “defect”, and if it is smaller than the threshold, it is determined as “normal”.

  As described above, the image processing device 15 may have a configuration in which the reflection data of the non-defective wafer is stored in advance and the array data of the shot area of the wafer and the threshold value of the brightness value are stored in advance. .

  In this case, since the position of each shot area in the captured reflection image of the wafer is known based on the array data of the shot area, the luminance value of each shot area is obtained. Then, a pattern defect is detected by comparing the brightness value with a stored threshold value. A shot region having a luminance value smaller than the threshold value may be determined as a defect.

  As described above, according to the surface inspection apparatus 30, the linearly polarized light L <b> 1 is used, and the vibration plane direction (V direction) in FIG. 6 is repeatedly inclined with respect to the repeating direction (X direction) of the repeating pattern 22. While illuminating the pattern 22, the defect of the repeated pattern 22 is detected based on the light intensity of the polarization component L4 (the magnitude of the polarization component L3 in FIG. 5C) of the elliptically polarized light L2 generated in the regular reflection direction. Therefore, even if the pitch P of the repeated pattern 22 is sufficiently small compared to the illumination wavelength, the defect inspection can be reliably performed. That is, even if the linearly polarized light L1 that is illumination light is not shortened in wavelength, it is possible to reliably cope with repetitive miniaturization of the pitch.

  Furthermore, in the surface inspection apparatus 30, the luminance value of the reflected image of the semiconductor wafer 20 is reduced by setting the angle formed by the vibration surface direction (V direction) and the repetition direction (X direction) in FIG. The amount (the amount of decrease Δ in FIG. 8) can be grasped greatly, and the defect inspection of the repeated pattern 22 can be performed with high sensitivity.

The surface inspection apparatus 30 is not limited to the case where the pitch P of the repetitive pattern 22 is sufficiently small compared to the illumination wavelength, and the same is true even when the pitch P of the repetitive pattern 22 is about the same as the illumination wavelength or larger than the illumination wavelength. In addition, the defect inspection of the repeated pattern 22 can be performed. That is, the defect inspection can be surely performed regardless of the pitch P of the repeated pattern 22. This is because the ellipticalization of the linearly polarized light L1 due to the repeated pattern 22 occurs depending on the volume ratio between the line portion 2A and the space portion 2B of the repeated pattern 22, and does not depend on the pitch P of the repeated pattern 22.

Further, in the surface inspection apparatus 30, if the volume ratio between the line portion 2A and the space portion 2B of the repetitive pattern 22 is the same, the amount of decrease in the luminance value of the reflected image (the amount of decrease Δ in FIG. 8) becomes equal. For this reason, regardless of the pitch P of the repetitive pattern 22, if the change amount of the volume ratio is the same, the detection can be performed with the same sensitivity. For example, as in the repeated pattern 22 shown in FIGS. 9A and 9B, when the pitch P is different and the volume ratio between the line portion 2A and the space portion 2B is the same, the defect inspection can be performed with the same sensitivity. As can be seen from the comparison between FIGS. 9A and 9B, the smaller the pitch P, the more reliably the minute shape change (deviation amount δ from the design value of the line width D _A of the line portion 2A) is reliably detected. can do.

  In the surface inspection apparatus 30, even when the pitch P of the repetitive pattern 22 is different, the inspection can be performed while the semiconductor wafer 20 is kept in a horizontal state (without adjusting the tilt of the stage as in the prior art). Therefore, it is possible to reliably reduce the preparation time until the defect inspection is started (that is, to capture the reflection image of the semiconductor wafer 20), and the working efficiency is improved.

  Furthermore, in the surface inspection apparatus 30, since the stage 11 does not have a tilt mechanism, the apparatus configuration is simplified. In addition, an inexpensive discharge light source can be used as the light source 31 of the illumination system 13, and the overall configuration of the surface inspection apparatus 30 is inexpensive and simple.

In the surface inspection apparatus 30, even when a plurality of types of repetitive patterns are formed on the surface of the semiconductor wafer 20 and repetitive patterns having different pitches P and different repetitive directions (X directions) are mixed, the entire surface of the semiconductor wafer 20 is present. It is possible to easily inspect defects of all the repetitive patterns simply by fetching the reflected images at once and examining the amount of decrease in the luminance value at each location. Incidentally, the repeating patterns having different repeating directions are a repeating pattern 25 in the 0 degree direction and a repeating pattern 26 in the 90 degree direction, as shown in FIG. These repeating patterns 25 and 26 are 90 degrees different from each other in the repeating direction (X direction). However, the angle formed between each repeating direction (X direction) and the direction of the vibration plane of the linearly polarized light L1 (V direction) is 45 degrees.

  Further, in the surface inspection apparatus 30, the linearly polarized light L 1 is incident obliquely on the surface of the semiconductor wafer 20 (see FIG. 1), so that the edge shape asymmetry of the line portion 2 A of the repetitive pattern 22 (for example, the edge shape collapses). It is also possible to obtain defect information related to the directionality of For this purpose, the repeat direction (X direction) of the repetitive pattern 22 of the semiconductor wafer 20 is rotated 180 degrees by the stage 11, the reflected image of the semiconductor wafer 20 is captured in the state before and after that, and the luminance difference at the same location is examined. become.

FIG. 11 illustrates the relationship between the repetitive pattern 22 having an asymmetric edge shape and the incident direction of the linearly polarized light L1. For example, FIG. 11A shows a state before 180 ° rotation, and illumination light is incident from the broken edge (E ₁ ) side of the edges E ₁ and E ₂ of the line portion 2A. FIG. 11B shows a state after rotation of 180 degrees, and illumination light is incident from the unbroken edge (E ₂ ) side of the _two edges E ₁ and E ₂ . Then, the brightness value of the reflected image captured in each state reflects the edge shape of the edges E ₁ and E _{2 in} the incident direction, and in this example, the case of FIG. The brightness value increases. Therefore, the asymmetry of the edge shape of the line portion 2A can be found by examining the luminance difference between the reflected images before and after being rotated 180 degrees. The defect inspection may be performed by combining the reflection images before and after being rotated by 180 degrees.

When the linearly polarized light L1 is incident obliquely on the surface of the semiconductor wafer 20 (see FIG. 1, incident angle θ), the elliptically polarized light L2 (FIG. 5B) generated from the repeated pattern 22 can be said strictly. For example, it is slightly rotated around its traveling direction. For this reason, it is preferable to finely adjust the direction of the transmission axis of the polarizing plate 38 of the light receiving system 14 in consideration of the rotation angle. In the state after fine adjustment, the orientations of the transmission axes of the two polarizing plates 34 and 38 are not exactly 90 degrees, but such an angle is also in the “vertical (or orthogonal)” category, I can say that. By finely adjusting the orientation of the transmission axis of the polarizing plate 38, the inspection accuracy can be improved. As a fine adjustment method, for example, an image is captured by reflecting linearly polarized light L1 on a surface without a repetitive pattern, and the direction of the transmission axis of the polarizing plate 38 is rotated so that the luminance value of the image is minimized. Conceivable.

  In the above description, an example in which the linearly polarized light L1 is p-polarized light has been described. However, the present invention is not limited to this. It may be s-polarized light instead of p-polarized light. The s-polarized light is linearly polarized light whose vibration surface is perpendicular to the incident surface. For this reason, as shown in FIG. 4, when the repetitive direction (X direction) of the repetitive pattern 22 of the semiconductor wafer 20 is set at an angle of 45 degrees with respect to the incident surface (3A) of s-polarized light that is linearly polarized light L1, The angle formed by the direction of the vibrating surface of the s-polarized light on the surface of the semiconductor wafer 20 and the repeating direction (X direction) of the repeating pattern 22 is also set to 45 degrees.

  The p-polarized light is advantageous for acquiring defect information related to the edge shape of the line portion 2A of the repetitive pattern 22. The s-polarized light is advantageous for efficiently capturing defect information on the surface of the semiconductor wafer 20 and improving the SN ratio.

  Furthermore, not only p-polarized light and s-polarized light, but also linearly polarized light whose vibration surface has an arbitrary inclination with respect to the incident surface may be used. In this case, the repeating direction (X direction) of the repeating pattern 22 is set to an angle other than 45 degrees with respect to the incident surface of the linearly polarized light L1, and the direction of the vibrating surface of the linearly polarized light L1 on the surface of the semiconductor wafer 20 and the repeating pattern are set. It is preferable to set the angle formed by the 22 repeat directions (X direction) to 45 degrees.

FIG. 12 is a diagram showing the illumination optical system 13 of the first embodiment according to the present invention.
The lamp house LS contains a light source 31 such as a halogen lamp (not shown), a metal halide lamp, or a mercury lamp, a wavelength selection filter 32, an ND filter (not shown) for adjusting the amount of light, etc. Is extracted as illumination light L 1 and is incident on the light guide fiber 33. The illumination optical system 13 includes a light guide fiber 33, a polarizing plate 34, a polarization compensator 9, and a concave reflecting mirror 35. The illumination light L1, which is a divergent light beam emitted from the light guide fiber 33, is converted into substantially parallel light by the spherical concave reflecting mirror 35, and illuminates the wafer 20 placed on the stage 11. A polarizing plate 34 is disposed in the vicinity of the emission portion of the light guide fiber 33, and the illumination light L1 emitted from the light guide fiber 33 is linearly polarized. The light that has been linearly polarized by the polarizing plate 34 is collimated by the concave reflecting mirror 35 through the polarization compensation plate 9 described later, and the linearly polarized collimated light illuminates the wafer 20. In order to improve throughput, it is extremely advantageous to collect images of the entire wafer surface in a lump. In this embodiment, as described above, the luminous flux from the light source is expanded and collimated by the concave reflecting mirror 35. In addition, the entire surface of the wafer can be illuminated.

  In the present embodiment, the polarization compensation plate 9 is disposed in the illumination optical system 13 between the polarizing plate 34 and the concave reflecting mirror 35. First, a description will be given of the polarization state of a light beam incident on and reflected from the concave reflecting mirror 12 when no polarization compensator is provided.

In FIG. 12, the illumination light L1 diverged in accordance with the numerical aperture of the light guide fiber 33 is converted into predetermined linearly polarized light by the polarizing plate 34 as described above, and the principal ray AX1 of the divergent light beam is the light from the concave reflecting mirror 35. This is a so-called off-axis optical system that is incident on a portion shifted from the axis O35.

  Here, for explanation, regarding the concave reflecting mirror 35, a plane including the principal ray AX1 of the linearly polarized light L1 incident on the concave reflecting mirror and the perpendicular of the portion of the concave reflecting mirror on which the principal ray enters is used as the concave reflecting mirror. It is defined as a reference incident surface A4 for the incident linearly polarized light L1. In addition, an axis of the reference incident surface that is parallel to the principal ray and intersects with the concave reflecting mirror perpendicularly is defined as an optical axis O35 of the concave reflecting mirror.

  As described above, the light beam incident on the concave reflecting mirror 35 is a divergent light beam. For this reason, according to Frenel's reflection formula, a difference in transmittance occurs between the p component and the s component of the polarized light, and as a result, the polarization plane rotates.

  The polarization plane rotation behavior will be described below. Consider a case where a linearly polarized divergent light beam having a vibration surface (p-polarized light) parallel to the reference incident surface A4 is incident on the concave reflecting mirror 35. FIG. 13 shows this state. In FIG. 12, the light from the light source is incident only on the effective diameter of the concave reflecting mirror 35. However, in FIG. 13, the concave reflecting mirror is centered on the optical axis O35 and includes a circle (including the effective diameter of the concave reflecting mirror 35). This is expressed as a dotted line), and the diameter of the incident light beam is also shown enlarged to illuminate the entire circle. At this time, of the surface of the concave reflecting mirror 35, the rotation of the polarization plane is the portion that intersects the reference incident surface A4 and the portion that intersects with the surface that includes the optical axis O35 and is perpendicular to the reference incident surface A4. Although it does not occur, rotation occurs in other parts of the concave reflecting mirror 35. As shown in FIG. 13, the plane of polarization of the concave reflecting mirror 35 rotates symmetrically with respect to the reference incident plane. Further, the rotation of the vibration plane of the polarization is generated in line symmetry with respect to the concave reflecting mirror 35 including the optical axis O35 and the plane perpendicular to the reference incident plane. The amount of rotation of the polarized light is larger as the position is away from the optical axis O35 of the concave reflecting mirror. This is because the incident angle of the incident light beam increases as the position is farther from the optical axis of the concave reflecting mirror, that is, the position is away from the normal incidence.

  Here, as shown in FIG. 12, when a divergent light beam enters from a position shifted from the optical axis O35 of the concave reflecting mirror 35 (a region 35 surrounded by a solid line in FIG. 13 corresponds to a light beam incident region of the concave reflecting mirror 35). The leftmost light of the light beam incident on the concave reflecting mirror 35 has an inclination such that the incident angle is the smallest and the rightmost light has the largest incident angle. (The incident angle is the angle between the incident light and the normal of the concave reflecting mirror surface).

  In this way, the incident angle of light with respect to the concave reflecting mirror is different in the plane (has an inclination), so there is a slight difference in the rotation of the polarization plane in the plane. Causes unevenness in extinction ratio.

The azimuth angle αr of the polarized light of the reflected light when the linearly polarized light having the azimuth angle αi is incident on the concave reflecting mirror is expressed by the equation (1).
tanαr = rs / rp · exp (i · (Δs−Δp)) tanαi
= Rs / rp · exp (i · Δ) tanαi (1)
rp and rs are amplitude reflectivities of two components (hereinafter referred to as “p component” and “s component”) that vibrate in directions perpendicular to each other in a plane perpendicular to the traveling direction of light, and Δp and Δs are p. It is a phase difference caused by reflection of each of the component and s component, and is a value determined by the complex refractive index of the reflecting surface and the incident angle (see the principle of Born-Wolf optics III, metal optics, etc.). The reflecting surface of the concave reflecting mirror is a metal such as aluminum, and the phase difference Δ and the amplitude reflectances rp and rs in the equation (1) vary depending on the incident angle.

In the present embodiment, the optical member between the polarizing plate 34 and the wafer 20 is only the concave reflecting mirror 35 having an aluminum reflecting surface, and the rotation of the polarization plane of the light beam reflected by the concave reflecting mirror is at most several. It is as small as about 50 degrees (if the rotation angle of the polarization plane is 3 °, this corresponds to a phase change of 1/60 of the wavelength λ of the illumination light).

  Further, as described above, since the divergent light beam L1 is incident on a portion off the optical axis of the concave reflecting mirror 35, the linearly polarized light beam incident on the surface of the concave reflecting mirror 35 rotates symmetrically across the incident surface A4. Occurs. The amount of rotation increases as the distance from the optical axis O35 of the concave reflecting mirror 35 increases. Accordingly, the rotation amount has an inclination in the direction of the optical axis O35.

  In order to eliminate the uneven rotation of the polarization plane in the plane of the illumination light due to the rotation of the minute polarization plane distributed with an inclination, in this embodiment, the polarization compensator 9 is formed with the polarizing plate 34 and the concave surface. It arrange | positions between the reflective mirrors 35. FIG. The polarization compensation plate 9 is a parallel plate made of glass, and is inclined with respect to the optical axis AX1 of the illumination light L1. The operation of the polarization compensation plate 9 will be described below.

  A light beam L1 emitted from the light guide fiber 11 and converted into linearly polarized light through the polarizing plate 34 enters the polarization compensator 9. Here, since the light beam L1 is a divergent light beam and the polarization compensator 9 is disposed to be inclined with respect to the optical axis AX1, the incident angle of the light beam incident on the polarization compensator 9 is the cross-sectional direction of the light beam. With a slope.

The azimuth angle of the polarization of the transmitted light of the light beam incident on the polarization compensator at the azimuth angle α′i is expressed by equation (2).
tanαi = ts / tp · exp (i · (Δs−Δp)) · tanα'i
= Ts / tp · exp (i · Δ) · tan α'i (2)
In this case, ts and tp are amplitude transmittances at the transmission surfaces of the s component and the p component, respectively, and Δp and Δs are phase differences caused by the transmission of the components of the s component and the p component. ts, ts, Δp, and Δs are functions of the refractive index of the glass and the incident angle.

  In an ideal phase plate, the phase difference Δ and the rotation amount δ of the polarization plane have a relationship of Δ = 2δ, and it can be considered that the present embodiment also satisfies this relationship. Therefore, the illumination light L1 transmitted through the polarization compensator 9 changes the phase difference Δ between the p component and the s component according to the incident angle of the incident light, and the plane of polarization rotates. FIG. 13B shows the state of polarization rotation generated by the polarization compensator 9.

  The polarization plane of the illumination light L1 at the time of illuminating the wafer 20 is the sum of the rotation of the polarization plane generated by the polarization compensation plate 9 and the rotation amount of the polarization plane generated by the concave reflecting mirror 35. Therefore, the polarization compensator 9 is tilted with respect to the optical axis AX1 of the illumination optical system, and is arranged so as to cause rotation of the polarization plane having an inclination opposite to that of the polarization plane rotation amount generated by the concave reflecting mirror 35. By doing so, the value of the rotation amount of the polarization plane (that is, the azimuth angle α of the light over the entire surface of the illumination light beam) is aligned.

  In FIG. 12, as described above, the linearly polarized light beam incident on the concave reflecting mirror 35 is symmetric with respect to the incident surface A4, and the rotation amount increases as the distance from the optical axis O35 of the concave reflecting mirror 35 increases. (Ie, having an inclination about the optical axis O35 of the concave reflector). In order to give a tilt in the opposite direction to this tilt, the polarization compensator is arranged in a tilt direction opposite to the tilt direction of the concave reflecting mirror with respect to the light beam, so that a cross section of the light beam reflected by the concave reflecting mirror 35 is obtained. The distribution of the rotation amount of the polarization plane in the direction can be made substantially uniform.

Here, as shown in FIG. 16, it is preferable to provide the polarization compensation plate 9 with a position adjustment 40 for arbitrarily setting the tilt angle and the tilt direction. With this configuration, slight variations due to the apparatus can be adjusted. Further, for example, depending on the rotation amount or distribution change of the polarization plane rotation in the concave reflecting mirror 35 generated when the illumination wavelength is changed by the light source 31 and the wavelength selection filter, either one of the inclination angle and the inclination direction or By adjusting both, the phase compensation amount can be adjusted. Further, for example, fine adjustment in accordance with the state of the apparatus such as the adjustment state of the apparatus is possible. By the position adjusting mechanism, the adjustment including the fine adjustment can be actually performed in the phase difference change region of several tenths of the illumination wavelength λ.

  As described above, in the illumination optical system 13, by disposing the polarization compensation plate 9, the tilt of the rotation amount of the polarization plane in the cross-sectional direction of the light beam by the concave reflecting mirror 12 can be corrected. Thus, it is possible to irradiate illumination light having a uniform polarization plane rotation direction.

  As described above, since the concave reflecting mirror 35 is used off-axis to make the divergent light beam into a parallel light beam, the incident angle with respect to the normal is different at each point of the concave reflecting mirror 35 (the concave reflection in FIG. 12). The values of the mirrors 35 are inclined to the left and right), and the polarization plane is inclined. The rotation angles of these polarization planes are very small. In order to make the distribution of the rotation angles uniform, the polarization compensator 9 needs to have a phase difference having a very small inclination. As described above, when the light beam is incident on the glass surface at an angle, a difference occurs in the phase difference generated between the p-polarized light and the s-polarized light. By utilizing this, in Example 1, the inclination of the phase difference necessary to make the distribution of the rotation angle uniform is generated by inclining the parallel plane of the glass to the non-parallel light beam, and off-axis A slight disturbance of the rotational distribution of the vibration plane of polarized light in the optical system using the concave reflecting mirror can be compensated uniformly. Further, if the tilt angle and tilt direction of the parallel plane are configured to be adjustable, the amount of tilt of the phase difference can be adjusted on the order of several tenths to one hundredth of the wavelength λ. Subtle adjustments are possible according to the state of the.

  Further, in this embodiment, since a parallel plate of glass is used, there is an advantage that it is not affected by a processing error (generally about 10%) unlike a normal phase plate.

  In Example 2, the light receiving optical system of the surface inspection apparatus of the present invention will be described with reference to FIG. The light receiving optical system 14 of the present invention is composed of a concave mirror 36 that makes light from the wafer 20 incident and condensed, a polarization compensation plate 10 to be described later, a polarizing plate 38, a lens 37, and an imaging device 39. .

  The wafer 20 is illuminated with, for example, a linearly polarized light beam L1 in which the amount of rotation of the polarization plane is aligned in the cross-sectional direction of the illumination light by the polarization illumination optical system provided with the polarization compensation plate 9 as described in the first embodiment. Is done. As described in the explanation of the principle, the wafer is arranged so that the arrangement direction of the repeated patterns has an angle of 45 ° with respect to the vibration plane of the linearly polarized light of the illumination light. The state of polarization of the specularly reflected light L2 generated from the wafer 20 changes due to structural birefringence caused by the state of a repeated pattern (for example, pattern shape, pitch, edge shape, etc.) formed on the wafer surface.

The specularly reflected light L2 from the wafer 20 is guided and collected by a light receiving optical system 14 including a concave reflecting mirror 36 and a lens 37, and an image of the wafer 2 by the specularly reflected light L2 is captured on the imaging surface of the image sensor 39. To form. The image sensor 39 is, for example, a two-dimensional CCD camera.

  A polarizing plate 38 is disposed between the concave reflecting mirror 36 and the lens 37 so as to transmit linearly polarized light orthogonal to the linearly polarized light of the illumination light L1. A polarization compensation plate 10 is disposed between the polarizing plate 38 and the concave reflecting mirror 36.

  In Example 1, the inclination of the amount of rotation of the polarization plane in the cross-sectional direction of the beam of divergent light incident on the concave reflecting mirror 35 is corrected by the polarization compensator 9 so that the amount of rotation of the polarization plane is aligned. Based on the same principle, in the light receiving optical system 14, the rotation amount of the polarization plane in the polarization compensator 10 is aligned with the inclination of the rotation amount of the polarization plane in the cross-sectional direction of the convergent light beam reflected and emitted from the concave reflecting mirror 36. It is corrected as follows.

  In FIG. 14, regarding the concave reflecting mirror 36, a plane including the principal ray AX <b> 2 of the linearly polarized light L <b> 2 emitted from the concave reflecting mirror 36 and parallel to the optical axis O <b> 36 of the concave reflecting mirror is the linearly polarized light L <b> 2 emitted from the concave reflecting mirror 36. The incident surface is defined as a reference incident surface A5. On the other hand, the collimated light beam L2 reflected from the wafer 20 is incident on a portion of the concave reflecting mirror 36 that is off the optical axis O36 and receives a converging action, so that the light receiving optical system 14 is a so-called off-axis optical system. Yes.

  Here, for the sake of explanation, as in the first embodiment, with respect to the concave reflecting mirror 36, the principal ray AX2 of the linearly polarized light L2 emitted from the concave reflecting mirror and the perpendicular of the portion of the concave reflecting mirror from which the principal ray AX2 emerges Is defined as a reference incident surface A5 of the linearly polarized light L2 emitted from the concave reflecting mirror 36. Further, an axis of the reference incident surface that is parallel to the principal ray and intersects with the concave reflecting mirror perpendicularly is defined as an optical axis O36 of the concave reflecting mirror.

  The rotation of the polarization plane of the convergent light beam exiting from the concave reflecting mirror 36 is in accordance with the description made with reference to FIG. In the concave reflecting mirror 36, the polarization vibration plane rotates in line symmetry with respect to the reference incident surface A 5 in the plane of the concave reflecting mirror 36. The amount of rotation increases as the position is farther from the optical axis O36 of the concave reflecting mirror. This is because the convergent light beam L2 emitted from the concave reflecting mirror 36 is emitted from a position deviated from the optical axis O36 of the concave reflecting mirror 36, and therefore the rightmost light of the light beam emitted from the concave reflecting mirror 36 in FIG. This is because the light exit angle is the smallest, and the leftmost light has a slope such that the light exit angle is the largest. (The exit angle is the angle between the exit light and the normal of the concave reflecting mirror surface). As described above, since the light emission angle with respect to the concave reflecting mirror is different (inclined) in the plane, there is a slight difference in the rotation of the polarization plane in the plane. appear.

  The polarization compensation plate 10 is a glass parallel plane plate, similar to the polarization compensation plate 9 of the first embodiment, and is inclined with respect to the principal ray AX2 of the regular reflection light L2. Since the specularly reflected light L2 is convergent light, the magnitude of the incident angle with respect to the polarization compensator 10 that is arranged to be inclined with respect to the optical axis AX2 has an inclination in the cross-sectional direction of the light beam.

  For this reason, the polarization compensator 10 can also generate a phase difference between the p component and the s component having an inclination corresponding to the incident angle, similarly to the polarization compensator 9. If the angle of the polarization compensator is determined such that the phase difference between the p component and the s component has a reverse slope with respect to the slope of the phase difference distribution between the p component and the s component generated in the concave reflecting mirror 41. The rotation of the polarization plane can be almost aligned.

  As shown in FIG. 16, it is preferable that the polarization compensation plate 10 is also provided with a position adjusting mechanism 41 so that the inclination angle and the inclination direction arranged in the light receiving optical system 14 can be set freely. With this configuration, it is possible to perform fine adjustment in accordance with the status of the apparatus, such as the adjustment state of the apparatus. In addition, as in the case of the illumination optical system 1, it is possible to perform correction including fine adjustment in the phase difference change region of several tenths of the illumination wavelength λ with this configuration. Further, similarly to the case of the illumination optical system 13, it is possible to cope with a change in the rotation distribution of the polarization plane that occurs when the wavelength of the illumination light is changed.

  As described above, also in the light receiving optical system 14, by arranging the polarization compensation plate 10, the inclination of the rotation amount of the polarization plane in the cross-sectional direction of the light beam by the concave reflecting mirror 36 can be corrected. The regular reflection light L2 can be guided to the image sensor 39 without changing the distribution of the polarization plane. Therefore, an apparatus with high detection accuracy can be provided.

  In the embodiment, the optical member between the wafer 20 and the polarizing plate 38 is only the concave reflecting mirror 36 having an aluminum reflecting surface. The rotation of the polarization plane by the concave reflecting mirror 36 is small and about several degrees (corresponding to a phase change of about 1/60 of the wavelength λ of the illumination light when the rotation angle of the polarization plane is 3 degrees).

  As described above, in this embodiment, the concave reflecting mirror 36 is used off-axis to convert the parallel light beam into a converging light beam, and therefore the exit angle with respect to the normal line differs at each point of the concave reflecting mirror 36 (FIG. 14). The value is inclined on the left and right of each concave reflecting mirror 36), and the rotation of the polarization plane is inclined. The rotation angles of these polarization planes are very small. In order to make the rotation angle distribution uniform, the polarization compensator 10 needs to have a phase difference having a very small inclination. As described in the first embodiment, when a light beam is incident on the glass surface with an angle, a difference occurs in the phase difference generated between the p-polarized light and the s-polarized light. Utilizing this fact, also in the second embodiment, the necessary retardation of the phase difference is generated by arranging the parallel plane of the glass in a non-parallel light beam. In addition, if the tilt angle and tilt direction of the parallel plane can be adjusted, the amount of tilt of the phase difference can be adjusted on the order of several tenths to one hundredth of the wavelength λ. Subtle adjustments can be made according to the state of Further, there is an advantage that it is not affected by a processing error (generally about 10%) unlike a normal phase plate.

In Example 3, a surface inspection apparatus including the illumination optical system described in Example 1 and the light receiving optical system described in Example 2 will be described. FIG. 15 shows a surface inspection apparatus of the present invention.
The configuration of the illumination optical system 13 is the same as that of the first embodiment. A polarizing plate 34 is disposed in the vicinity of the emission portion of the light guide fiber 33, and the illumination light L1 emitted from the light guide fiber 33 is linearly polarized. The light that has been linearly polarized by the polarizing plate 34 is collimated by the concave reflecting mirror 35 through the polarization compensation plate 9 described later, and the linearly polarized collimated light L 1 illuminates the wafer 20. In order to improve throughput, it is extremely advantageous to collect images of the entire wafer surface in a lump. In this embodiment, as described above, the luminous flux from the light source is expanded and collimated by the concave reflecting mirror 35. In addition, the entire surface of the wafer can be illuminated.

  The linearly polarized collimated light L 1 incident on the wafer 20 is reflected by the wafer surface and enters the light receiving optical system 14. The configuration of the light receiving optical system 14 is the same as that of the second embodiment. The light beam L2 reflected by the wafer 20 is incident on the concave reflecting mirror 36 and receives a converging action. The convergent light beam is a polarizing plate in which the polarizing compensator 10 described later and the polarizing plate 34 are arranged in a crossed Nicols relationship. Then, an image of the surface of the wafer 20 is formed on the imaging surface of the imaging device 39 disposed at a position conjugate with the surface of the wafer 20 by the imaging lens 37.

  For example, a plurality of chip areas 21 as shown in FIG. 2 are arranged in the XY direction on the surface of the wafer 20, and a repeated pattern 22 is formed in each chip area 21. The repetitive pattern 22 is a resist pattern (for example, a wiring pattern) in which a plurality of line portions 2A and space portions 2B are arranged at a constant pitch P along the short direction (X direction) as shown in FIG. .

  The stage 11 places the wafer 20 on which the above-described pattern is formed, and fixes and holds the wafer 20 by vacuum suction or the like. Further, the stage 11 is configured to be rotatable around a predetermined rotation axis orthogonal to the stage surface by a stage rotation mechanism 16. With this stage rotation mechanism 16, the angle formed by the longitudinal direction of the repeated pattern formed on the surface of the wafer 20 with respect to the linearly polarized vibrating surface of the light beam L1 that illuminates the wafer 20 can be set to an arbitrary angle.

In the surface inspection apparatus of FIG. 15, an alignment system for detecting the orientation of the pattern formed on the surface of the wafer 20 placed on the stage 11 between the concave reflecting mirror 35 and the concave reflecting mirror 36. 12 is arranged, and an angle formed by a preset linearly polarized vibrating surface of the light beam L1 and the longitudinal direction Y of the repeated pattern 22 is detected, and the stage rotating mechanism 16 is used to detect the illumination optical system 13 and the light receiving optical system 14. The direction of the longitudinal direction Y of the repeated pattern can be adjusted. The principle of defect inspection in this embodiment is the same as that described in the explanation of the principle of the surface inspection apparatus.

  By configuring the surface inspection apparatus 30 using the illumination optical system described in the first embodiment and the light receiving optical system described in the second embodiment, the rotation of the polarization plane by the two concave reflecting mirrors 35 and 36 is prevented. Since the polarization compensators 9 and 10 are arranged and the phase change is appropriately applied as described above to align the rotation angle of the polarization plane, the extinction ratio by the two polarizing plates 34 and 38 arranged in crossed Nicols. It is possible to configure a surface inspection apparatus that suppresses the above.

  In the surface inspection apparatus of FIG. 15, as described above, the image of the wafer 20 obtained through the illumination optical system and the light receiving optical system in which the extinction ratio is kept small is arranged at a position conjugate with the surface of the wafer 20. The image is picked up by the image pickup device 39 and converted into a digital image. The image obtained here has different luminance values depending on the pattern region based on the shape and pitch of the repeated pattern formed on the wafer surface, the shape of the side surface, and the like. The digital image is sent to the image processing device 15, and image processing of the image captured by the image sensor 39 is performed to extract a luminance value for each pattern area. In parts where there is an abnormality in the defocus or exposure amount of the exposure device, the amount of rotation of the polarization plane is different compared to the part where exposure has been performed normally. Arise. The image processing device 15 detects this based on the extracted luminance value, and performs defect inspection.

  Further, if the illumination optical system 13 needs to illuminate the wafer with a linearly polarized light beam whose polarization azimuths are completely aligned, the polarization compensator 9 is provided, for example, by providing a predetermined phase plate behind the polarizing plate. It is sufficient to give a phase difference that causes a rotation opposite to the rotation of the polarization plane by the concave reflecting mirror 35.

  As described above, in the surface inspection apparatus of this embodiment, the rotation of the respective polarization planes by the concave reflecting mirror on the illumination side of the wafer and the concave reflecting mirror on the light receiving side from the wafer is performed in-plane by the polarization compensators 9 and 10, respectively. The rotation of the polarization plane can be aligned. Therefore, since the non-uniformity of the extinction ratio due to the illumination optical system and the light receiving optical system can be eliminated, the extinction ratio is improved over the entire surface of the light beam cross section, and noise is reduced. It can be detected with detection accuracy.

  Here, as shown in FIG. 16, it is preferable that the polarization compensation plate 9 and the polarization compensation plate 10 are provided with position adjustments 40 and 41 for arbitrarily setting the inclination angle and the inclination direction, respectively. With this configuration, slight variations due to the apparatus can be adjusted. Further, for example, according to the rotation amount and distribution change of the polarization plane rotation in the concave reflecting mirror 35 that occurs when the polarizing plates 34 and 38 are rotated or when the illumination wavelength is changed by the light source 31 and the wavelength selection filter. Thus, the phase compensation amount can be adjusted by adjusting one or both of the tilt angle and the tilt direction. Further, for example, fine adjustment in accordance with the state of the apparatus, such as the adjustment state of the apparatus, is possible. By the position adjusting mechanism, the adjustment including the fine adjustment can be actually performed in the phase difference change region of several tenths of the illumination wavelength λ.

(Modification)
In the present embodiment, a surface inspection apparatus that performs defect detection using structural birefringence has been described. However, as a modification, the technique of the present invention can be applied to a hole pattern inspection method using polarized light. In this case, not only specularly reflected light but also diffracted light from the pattern is used. When performing inspection using diffracted light, for example, in FIG. 15, a stage 11 is provided with a tilt mechanism (not shown). By this tilt mechanism, the stage 11 is tilted about the rotation axis AX11 perpendicular to the paper surface, and the angle is adjusted so that an arbitrary order diffracted light generated from the repeated pattern on the wafer 20 can be taken in by the light receiving optical system. On the other hand, when performing defect detection based on structural birefringence, a specularly reflected light beam is mainly used.

  In the above-described embodiment, a polarization compensator is inserted in both the illumination optical system and the light receiving optical system to eliminate both the rotation unevenness of the polarization plane of the illumination optical system and the rotation unevenness of the polarization plane of the light receiving optical system. However, the polarization compensator may be disposed in either the illumination optical system or the light receiving optical system to simultaneously correct the phase shift caused by the concave reflecting mirror 35 and the concave reflecting mirror 36. Of course, if it cannot be completely corrected due to the off-axis direction or angle, the rotation angle of the polarization plane can be more reliably ensured by inserting the illumination system and the light receiving system separately as in the embodiment, and making independent corrections. Can be aligned and the extinction ratio is improved. The arrangement of the polarization compensator may be selected according to the arrangement position of the concave reflecting mirror and the required specifications. As for the arrangement of the concave reflecting mirror, in the first and second embodiments, as described with reference to FIG. 13, the incident surface A4 (A5) and the optical axis O35 of the concave reflecting mirror on the plane perpendicular to the incident surface. Of the reflecting surfaces defined by the surface including the light beam, the light beam is reflected by using a surface whose diameter direction overlaps the incident surface. However, the present invention is not limited to the embodiment. For example, the light may be reflected in a region that does not include both the incident surface and a surface that is perpendicular to the incident surface and includes the optical axis of the concave reflecting mirror. In this case, the polarization rotation angle is larger than that of the plane including the incident surface, but the polarization compensator tilted so as to eliminate the tilt of the rotation amount of the generated vibration plane is disposed as in the first and second embodiments. By doing so, it can correct | amend so that distribution of rotation of the vibration surface in the cross-sectional direction of a light beam may become uniform.

Moreover, although the polarization compensator 9 and the polarization compensator 10 are parallel plane plates made of glass, since they are inserted into the divergence and convergence system and tilted, aberration (as) due to tilt occurs. In order to reduce the aberration, it is desirable to reduce the tilt angle and reduce the thickness. In order to reduce the inclination angle, the refractive index of the plane parallel plate may be increased. A general low refractive index optical glass such as BK7 has a refractive index of 1.
However, there are some flint-type high refractive index glasses having a refractive index close to 2.0, and these may be used. Or you may use what vapor-deposited (coated) the thin film of the high refractive index substance on the glass surface of a low refractive index. The phase change in this case is obtained by adding the phase change due to the refractive index difference at the interface between the vapor deposition material and air and the phase change due to the refractive index difference at the interface between the vapor deposition material and glass. The difference in refractive index between the vapor deposition material and glass is smaller than the refractive index difference between the vapor deposition material and air, and the phase change due to the interface between the vapor deposition material and air is dominant. Accordingly, an effect equivalent to that of a plane parallel plate made of high refractive index glass can be obtained. Depending on the combination of the thickness of the glass and the vapor deposition material, a reflection increasing film may be formed and the transmittance may be lowered. If this is inconvenient, a high refractive index substance may be deposited on a high refractive index glass. If the refractive indexes of glass and vapor deposition material are the same, the effect is equivalent to that of glass only. High-refractive-index glass is prone to burn, and thin film deposition also prevents burn.

The polarization compensators 9 and 10 may be made of plastic instead of glass as long as the material is transparent to the inspection light. The material of the parallel plate may be selected according to the required specifications.
Furthermore, the polarization compensation plates 9 and 10 are not glass parallel plane plates but are bonded to each other so that two crystal planes having birefringence, for example, quartz or other parallel plane plates are orthogonal to each other as shown in FIG. You may also use a new one. The rotation of the polarization plane caused by the concave reflecting mirror can be canceled out by the thickness of the crystal and the phase difference caused by the difference in refractive index between ordinary rays and extraordinary rays. When the thickness of each of the two crystals is t1, t2, and the refractive index difference between the ordinary ray and the extraordinary ray of the crystal is Δn, the phase difference φ is
φ = 2π / λ · (t1−t2) Δn

Further, since the polarization compensation plates 9 and 10 are arranged in the pupil space, the angle formed with the optical axis of the light beam corresponds to the position on the wafer. The beam diameter in the pupil space corresponds to the numerical aperture (NA) on the wafer. When the beam diameter is large as shown in FIG. 18, the above-mentioned parallel plane plate or crystal of glass is preferably laminated, but when the beam diameter is thin as shown in FIG. 20, a Babinet compensation plate may be used. The Babinet compensation plate is a parallel flat plate in which two wedge prisms such as a crystal having birefringence, for example, quartz, are bonded so that the crystal axis directions thereof are orthogonal to each other as shown in FIG. The resulting phase difference is the same as in the previous equation, but the t1-t2 value changes at the position through the compensator depending on the wedge angle. Therefore, a different phase difference can be given for each angle, and the rotation of the polarization plane caused by the concave reflecting mirror can be canceled as before.

  However, as shown in FIG. 18, even when the light beam diameter is large, the light beam having the same angle (light beam that illuminates, reflects, or diffracts the same position on the wafer. Indicated by a solid line, a dotted line, and a one-dot chain line) Since the phase difference changes, the extinction ratio cannot be improved uniformly.

  Accordingly, a compensation plate may be selected as appropriate depending on the size of the light beam diameter. When a birefringent crystal is used, it becomes difficult to use when the incident angle is large, but it can also be used when the divergence angle of the device is relatively small. What is necessary is just to select according to the conditions of an apparatus.

  In all the embodiments of the present invention, the configuration using the concave reflecting mirror in the illumination optical system and the light receiving optical system has been described. However, the reflecting mirror is not limited to the concave reflecting mirror, and the diverging light or the convergent light is arranged obliquely. Even when the light enters the refractive optical system or the catadioptric optical system, an inclination of the rotation amount of the polarization plane in the cross-sectional direction of the light beam occurs. Even in such a case, it goes without saying that the amount of rotation of the polarization plane can be made uniform in the cross-sectional direction of the light beam by employing the configuration of the present invention.

1 is a diagram illustrating an overall configuration of a surface inspection apparatus 30. FIG. 2 is an external view of the surface of a semiconductor wafer 20. FIG. FIG. 6 is a perspective view for explaining an uneven structure of a repeated pattern 22. It is a figure explaining the inclination state of the incident surface (3A) of the linearly polarized light L1 and the repeating direction (X direction) of the repeating pattern 22. FIG. It is a figure explaining the vibration direction of linearly polarized light L1 and elliptically polarized light L2. It is a figure explaining the inclination state of the direction (V direction) of the vibration surface of the linearly polarized light L1, and the repeating direction (X direction) of the repeating pattern 22. FIG. It is a diagram for explaining a state where divided into the repeat direction (X direction) polarization component parallel to the V _X perpendicular polarization component V _Y. 6 is a diagram for explaining the relationship between the size of a polarization component L3 and the line width D _A of a line portion 2A of a repetitive pattern 22. FIG. It is a figure which shows an example of the repeating pattern 22 from which the pitch P differs and the volume ratio of the line part 2A and the space part 2B is the same. It is a figure explaining the repeating patterns 25 and 26 from which a repeating direction differs. It is a figure which shows the relationship between the repeating pattern 22 with an asymmetrical edge shape, and the incident direction of the linearly polarized light L1. It is a figure which shows the whole structure of the surface inspection apparatus of 1st Example. (A) It is a figure which shows the mode of the polarization | polarized-light rotation in the concave reflecting mirror 35. FIG.

(B) It is a figure which shows the mode of the polarization | polarized-light rotation in the polarization compensator.
It is a figure which shows the whole structure of the surface inspection apparatus of 2nd Example. It is a figure which shows the whole structure of the surface inspection apparatus of 3rd Example. It is a figure which shows the modification of the surface inspection apparatus of 3rd Example. It is a figure which shows the structure of a polarization compensator. It is a schematic diagram of the light beam passing through the polarization compensation plate. It is a figure which shows the structure of a polarization compensator. It is a schematic diagram of the light beam passing through the polarization compensation plate.

Explanation of symbols

9, 10 Polarization compensation plate 30 Surface inspection device 11 Stage 12 Alignment system 13 Illumination system 14 Light receiving system 15 Image processing device 16 Stage rotation mechanism 20 Semiconductor wafer 21 Chip region 22, 25, 26 Repeat pattern 22
31 Light source 32 Wavelength selection filter 33 Light guide fibers 34 and 38 Polarizing plates 35 and 36 Concave reflecting mirror 37 Imaging lens 39 Imaging element L1 Illumination light L2 Reflected light

Claims (10)

Light source means for emitting a linearly polarized divergent light beam for illuminating a substrate to be inspected on which a repeated pattern is formed;
The principal ray of the linearly polarized divergent light beam is disposed between the light source and the test surface so as to enter a position shifted from the optical axis, and guides the linearly polarized divergent light beam to the substrate to be inspected. An optical member;
A light receiving means for receiving linearly polarized light whose polarization direction is orthogonal to the linearly polarized light out of the light flux from the inspection substrate;
Polarized light generated due to the optical member disposed in the optical path between the light source means and the light receiving means, being tilted in a direction opposite to the tilt direction of the optical member with respect to the principal ray of the divergent light beam A polarization correction member composed of at least one parallel plate for correcting surface disturbance,
A surface inspection apparatus for inspecting a surface of the substrate to be inspected based on light received by the light receiving means. Light source means for emitting a linearly polarized light beam for illuminating a substrate to be inspected on which a repeated pattern is formed;
The light beam from the substrate to be inspected is arranged so that the principal ray of the light beam from the substrate to be inspected enters a position shifted from the optical axis, and the light beam from the substrate to be inspected is incident on the light beam. An optical member that emits a principal ray of the light beam as a convergent light beam having a predetermined emission angle;
A light receiving means for receiving linearly polarized light orthogonal to the predetermined linearly polarized light among the convergent light flux from the optical member;
Polarized light generated due to the optical member disposed between the light source unit and the light receiving unit in the optical path, tilted in a direction opposite to the tilt direction of the optical member with respect to the principal ray of the convergent light beam. A polarization correction member composed of at least one parallel plate for correcting surface disturbance,
A surface inspection apparatus for inspecting a surface of the substrate to be inspected based on light received by the light receiving means. Light source means for emitting a linearly polarized divergent light beam for illuminating a substrate to be inspected on which a repeated pattern is formed;
Located between the light source and the test surface, the linearly polarized divergent light beam is incident on a position where the principal ray of the light beam is shifted from the optical axis, and the linearly polarized divergent light beam is incident on the substrate to be inspected. A first optical member for guiding;
The light beam from the substrate to be inspected is disposed at a position where the light beam from the substrate to be inspected enters, the light beam from the substrate to be inspected is incident on a position where the principal ray of the light beam is shifted from the optical axis, and the convergent light beam has a predetermined emission angle A second optical member that is ejected and imaged on a predetermined surface;
Extraction means for extracting linearly polarized light orthogonal to the linearly polarized light out of the convergent light beam from the second optical member;
A light receiving means for receiving an image of the inspection substrate formed through the second optical member and the extracting means;
In the optical path between the light source means and the light receiving means, the optical member is inclined in the direction opposite to the inclination direction of the optical member with respect to the principal ray of the divergent light beam or the inclination direction of the optical member with respect to the principal ray of the convergent light beam. A surface inspection apparatus comprising: a polarization correction member comprising at least one parallel plate that is arranged in a manner that corrects a polarization plane disturbance of the light beam caused by the first and second optical members. The surface inspection apparatus according to claim 1, wherein the optical member includes a concave mirror .   The surface inspection apparatus according to claim 1, further comprising a holding unit that holds at least one of an inclination direction and an inclination angle of the polarization correction member in an adjustable manner.   6. The surface inspection apparatus according to claim 1, wherein the polarization correction member is a parallel plate of glass arranged to be inclined with respect to a surface of the optical member.   The polarization correction member is a parallel flat plate of two birefringent crystals that are arranged so as to be inclined with respect to a plane perpendicular to the optical axis of the optical member and are bonded so that the crystal axes thereof are orthogonal to each other. The surface inspection apparatus according to claim 1, wherein the surface inspection apparatus is provided.   The polarization correction member is disposed with an inclination with respect to a plane perpendicular to the optical axis of the optical member, and is bonded to form a parallel plate so that the crystal axes thereof are orthogonal to each other. 6. The surface inspection apparatus according to claim 1, wherein the surface inspection apparatus is a wedge-shaped birefringent crystal. Light source means for emitting a linearly polarized divergent beam;
The principal ray of the linearly polarized divergent light beam is disposed between the light source and the test surface so as to enter a position shifted from the optical axis, and guides the linearly polarized divergent light beam to the substrate to be inspected. An optical member;
In the optical path between the light source means and the substrate to be inspected, the optical member is tilted in a direction opposite to the tilt direction of the optical member with respect to the principal ray of the divergent light beam, and is generated due to the optical member. A polarization illumination device comprising: a polarization correction member comprising at least one parallel plate for correcting disturbance of a polarization plane of the light beam. It is arranged so that the principal ray of the light beam from the substrate to be inspected is incident on a position shifted from the optical axis at the position where the light beam from the substrate to be inspected, and a predetermined polarization component from the substrate to be inspected An optical member that makes the incident light beam incident and emits the convergent light beam with a predetermined emission angle;
A light receiving means for receiving linearly polarized light out of the light flux from the optical member;
In the optical path between the substrate to be inspected and the light receiving means, the optical member is disposed in an inclined direction opposite to the inclination direction of the optical member with respect to the principal ray of the convergent light beam, and is generated due to the optical member. A light receiving device comprising: a polarization correction member comprising at least one parallel plate for correcting disturbance of a polarization plane of the light beam.

Images