JPH09304016A - Surface position detecting device and exposure apparatus including the device - Google Patents

Abstract

(57) An object of the present invention is to provide an oblique incidence type surface position detection device having a simple configuration and easy adjustment. An image of a pattern on a lattice pattern forming surface 8a is projected onto a surface 1a to be inspected via projection optical systems 9 and 10.
The image on the surface 1a to be inspected is re-imaged on the light receiving surface 25a of the tilt correction prism 25 via the condensing optical systems 11 and 12. The grating pattern forming surface 8a and the test surface 1a satisfy the Scheimpflug condition with respect to the projection optical systems 9 and 10, and the test surface 1a and the incident surface 25a satisfy the Scheimpflug condition with respect to the condensing optical systems 11 and 12. And the light-receiving slit S is formed on the incident surface 25a, and the light-receiving surface 17a of the photoelectric detector 17 is arranged in the vicinity of the exit surface 25b of the tilt correction prism 25. Are relatively vibrated by the vibrating mirror 30.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention is for transferring a mask pattern onto a photosensitive substrate in a photolithography process for manufacturing a semiconductor device, an image pickup device (CCD or the like), a liquid crystal display device, a thin film magnetic head or the like. The present invention relates to a surface position detection device which is provided in an exposure device used for the above, and which is suitable for use in an oblique incidence type autofocus sensor for detecting the position of the surface of the photosensitive substrate, and an exposure device including this device.

[0002]

2. Description of the Related Art A projection exposure apparatus for transferring a pattern formed on a reticle as a mask to a photoresist layer in each shot area on a wafer through a projection optical system when manufacturing a semiconductor device, for example. Steppers, etc.) are used. In such a projection exposure apparatus, a projection optical system having a large numerical aperture and, as a result, a considerably shallow depth of focus is used in order to improve resolution. Furthermore, since unevenness may be partially present on the wafer to be transferred due to uneven thickness or foreign matter on the bottom surface, conventionally, each shot area on the wafer has its surface in the optical axis direction. Of the position (focus position) from the image plane (defocus amount)
Is detected, and the defocus amount thus detected is corrected by the autofocus method.

In such a case, an oblique-incidence type autofocus sensor (hereinafter referred to as "AF sensor") for obliquely projecting a slit image or the like onto a surface to be inspected such as a wafer is used as a device for detecting the focal position on the wafer surface. (For example, see Japanese Patent Laid-Open No. 56-42205). In this sensor, when the surface to be inspected moves up and down, the position of the slit image on the surface to be inspected shifts in the direction perpendicular to the optical axis of the oblique incidence optical system. The vertical displacement can be detected.

Further, in Japanese Patent Application Laid-Open No. 6-97045 and US Pat. No. 5,076,698, an image of a striped pattern is obliquely projected on a surface to be inspected and re-imaged by a light receiving system. A plurality of (for example, five) linear patterns are selected from the image of the striped pattern thus selected, and a plurality of linear patterns on the surface to be inspected are selected from the lateral shift amounts of the plurality of linear patterns thus selected. An oblique incidence type multi-point AF sensor that detects the position at a measurement point has also been proposed. This multipoint A
In the F sensor, since the image of the striped pattern is obliquely projected on the surface to be inspected, the image plane of the image of the striped pattern re-imaged by the light receiving system is relative to the optical axis of the light receiving system. Is inclined.
Since the light receiving efficiency of the photoelectric detector is poor as it is, the AF sensor reduces the tilt angle by passing the light flux from the image of the striped pattern re-formed in this way through the tilt correction prism. Moreover, the image of the striped pattern is relayed to the light receiving surface of the photoelectric detector using a relay optical system.

[0005]

In the conventional oblique incidence type multi-point AF sensor as described above, a relay optical system is provided between the prism for tilt correction and the light receiving surface of the photoelectric detector. Conventionally, an image of a striped pattern was formed near the entrance surface of the prism for the tilt correction, and the image of the striped pattern was considerably defocused on the exit surface of the prism. The relay optics were used to form the image on the light receiving surface of the photoelectric detector. However, providing the relay optical system in this way complicates the configuration of the AF sensor, and has a disadvantage that it takes time to adjust the optical system.

In view of the above problems, an object of the present invention is to provide an oblique incidence type surface position detecting device which has a simple structure and is easy to adjust. Another object of the present invention is to provide an exposure apparatus provided with such a surface position detecting device.

[0007]

A surface position detecting device according to the present invention is provided with a predetermined pattern formed on a first surface (8a) and a predetermined pattern from an oblique direction with respect to an inspection surface (1a). A projection optical system (9, 10) for projecting an image of a pattern, and a collection for condensing the light flux reflected by the surface (1a) to be inspected to form an image of the predetermined pattern on the second surface (25a). A light optical system (11, 12) and a detector (1 for photoelectrically detecting an image of a predetermined pattern formed by the light collecting optical system.
7), and a surface position detecting device for detecting the surface position of the surface to be inspected (1a) based on the output of this detector,
The first surface (8a) and the surface to be inspected (1a) satisfy the Scheimpflug condition with respect to the main plane of the projection optical system (9, 10), and the surface to be inspected (1a) and the second surface (25a) are , Scanning means (30, 31) for satisfying the Scheimpflug condition with respect to the principal planes of the condensing optical system (11, 12) and for relatively scanning the light receiving surface of the detector (17) and the image of a predetermined pattern thereof. ) And a deflecting member (25) for tilt correction, which is disposed substantially on the second surface (25a) and deflects the light beam from the condensing optical system (11, 12) in a predetermined direction. The light receiving surface (17a) of the detector (17) is arranged close to the light emitting surface (25b) of the deflecting member.

In this case, the Scheimpflug condition will be described with reference to FIG. 11, for example. That is, in FIG. 11, the fact that the A-plane and the B-plane satisfy the Scheimpflug condition for the optical system for forming the pattern on the A-plane on the B-plane is more accurately in the meridional section. When the intersection of the extension line of the A surface and the object-side main plane of the optical system is H, and the intersection of the extension line of the B surface and the image-side main plane of the optical system is H ', the optical axis from the intersection H Distance to
And the distance L ′ from the intersection H ′ to the optical axis are equal. When the Scheimpflug condition is satisfied, a so-called tilted image forming relationship is established, and the light flux emitted from any point on the A surface is focused on one corresponding point on the B surface. Therefore, an image of the entire point on the surface A is formed on the surface B.

According to the present invention, the first surface (8a) and the surface to be inspected (1a) satisfy a so-called tilt image forming relationship, and the surface to be inspected (1a) and the second surface (25a) are Satisfies the tilt image forming relationship, so that each point on the entire surface of the test surface (1a) forms an image on the second surface (25a). In this case, since the irradiation optical system (9, 10) projects a predetermined pattern image onto the surface to be inspected (1a) from an oblique direction, the surface to be inspected (1
If unevenness partially exists in a), a lateral shift occurs in the image on the second surface (25a) corresponding to the unevenness. Further, by using the photoelectric conversion signal obtained from the detector (17) when the scanning means (30, 31) relatively scans the light receiving surface of the detector (17) and the image of the predetermined pattern, for example, A signal corresponding to the surface position of the surface to be inspected (1a) is obtained from the principle of the photoelectric microscope. That is, since the level of the signal obtained by synchronously rectifying the photoelectric conversion signal with the drive signal of the scanning means (30, 31) changes depending on the unevenness of the surface to be inspected (1a), the signal to be inspected (1a ) Vertical position distribution can be detected.

Further, the surface to be inspected (1a) and the second surface (25a)
And are in a tilted image formation relationship, the amount of lateral deviation of the image on the second surface (25a) due to vertical fluctuations of the surface to be inspected (1a) is larger than that in the magnification relationship between the surfaces. This will be described quantitatively. For example, as shown in FIG. 1, the incident angle of the irradiation optical system (9, 10) with respect to the test surface (1a) is θ, the vertical displacement of the test surface (1a) is z, and the test surface (1a). ), The lateral displacement amount y1 on the second surface (25a) can be expressed by the following equation, where δ ′ is the imaging magnification along the tilted imaging surface on the second surface (25a).

Y1 = 2δ'tan θz (1) By the way, if the lateral magnification which is the magnification in the direction perpendicular to the optical axis between the surface to be inspected (1a) and the second surface (25a) is δ, The following relationships are established from the Scheimpflug condition. δ ′ = (δ ² cos ² θ + δ ⁴ sin ² θ) ^1/2 (2) On the other hand, in the conventional technique disclosed in, for example, Japanese Unexamined Patent Publication No. 56-42205, the lateral shift is in the direction perpendicular to the optical axis. Therefore, the detected lateral shift amount y2 is as follows.

Y2 = 2δsin θz (3) Comparing equations (1) and (3), when the incident angle θ is large, (1 The lateral displacement amount y1 in the equation) is larger. If δ = 1 for the sake of simplicity, δ ′ = 1 from the equation (2), and the lateral shift amounts y1 and y
2 is proportional to tan θ and sin θ, respectively. For example, θ
= 80 °, y1 = 11.3z according to the equation (1) and y2 = 2.0z according to the equation (3). Therefore, it is 5.7 when the tilted image formation relationship is satisfied.
The lateral displacement becomes twice as large, and the detection sensitivity and the detection accuracy with respect to the upper and lower sides of the surface to be inspected (1a) are improved.

Moreover, the second surface (25) of the surface (1a) to be inspected
Since the imaging magnification δ ′ for a) is not increased so much, the image of the pattern re-imaged on the second surface (25a) does not increase so much, and the detection system can be downsized. However, when the incident angle θ with respect to the surface to be inspected (1a) is large, the detector (1
When the light receiving surface of 7) is arranged, the surface reflection on the light receiving surface becomes large and the amount of received light may be significantly reduced due to eclipse. In order to avoid this, in the present invention, the second
A deflection member (25) for tilt correction is arranged on the surface (25a), and the light receiving surface of the detector (25) is arranged in the vicinity of the exit surface of the deflection member (25).

In this case, an example of the deflecting member (25) is a prism, but a so-called blazed diffraction grating or the like can also be used. When the deflecting member (25) is a prism, for example, as shown in FIG. 4, the apex angle of the prism is ξ, the refractive index of the prism is n, and the main light flux entering the prism from the condensing optical system. When the incident angle of the light ray is α, it is desirable to determine the shape of the prism so that the apex angle ξ satisfies the following equation.

Ξ ≦ arcsin {(sinα) / n} (4) As shown in FIG. 4, the light ray L incident on the second surface (25a)
1 emits as a light ray L2 from the exit surface (25b) of the prism at an exit angle γ. This exit angle γ is
It is expressed as follows. Here, β is the refraction angle of the light ray L1 on the second surface (25a).

Γ = arcsin {n · sin (β−ξ)} (5) β = arcsin {(sinα) / n} (6) Therefore, when the condition of the equation (4) is satisfied, the apex angle ξ
Is less than or equal to the refraction angle β on the second surface (25a) of the equation (6). As a result, the exit angle γ of the light ray L2 from the exit surface (25b) becomes smaller than the expression (5), and the detector (17)
The light-receiving efficiency at 1 does not decrease. However, the second surface (25
An image of a predetermined pattern is formed on a) and the exit surface (25
Since the light receiving surface of the detector (17) is arranged in b),
If the divergence angle of the image-forming light beam is a certain size, the image-forming light beam spreads on the light receiving surface. Therefore, if there are a plurality of light receiving elements on the light receiving surface, the image forming light flux corresponding to another measurement point may be mixed in the light receiving elements that are close to each other. To prevent this, the second surface (25
It is necessary to make the distance between a) and the emission surface (25b) as small as possible. For this purpose, the apex angle ξ of the prism may be set to a small value in a range that satisfies the expression (4) and a range in which the exit angle γ in the expression (5) does not become too large.

Light having a refractive index larger than 1, that is, larger than air between the light exit surface (25b) of the deflecting member (25) and the light receiving surface (17a) of the detector (17). It is desirable to fill it with a permeable medium (41). As a result, the incident angle of the light beam on the detector (17) becomes smaller, and the light receiving efficiency is improved. Next, an exposure apparatus according to the present invention comprises a surface position detection apparatus according to the present invention, a projection optical system (5) for projecting an image of a mask pattern (R) onto a photosensitive substrate (1), and a photosensitive substrate (1 ) Projection optical system (5)
Surface position adjustment stage (2,
3), and the surface position detection device detects the surface position of the photosensitive substrate (1), and the photosensitive substrate is transferred via the surface position adjustment stage (2, 3) based on the detection result. The surface position of (1) is adjusted.

According to the present invention, the position of the photosensitive substrate (1) in the optical axis direction of the projection optical system (5) is detected with high accuracy by the surface position detecting device having a simple structure and easily adjusted. The surface of the photosensitive substrate (1) can be accurately aligned with the image plane of the projection optical system (5) based on the detection result.

[0019]

BEST MODE FOR CARRYING OUT THE INVENTION An example of an embodiment of a surface position detecting device according to the present invention will be described below with reference to the drawings.
In this example, the present invention is applied to an oblique incidence type multi-point AF sensor (autofocus sensor) provided in a projection exposure apparatus for semiconductor manufacturing. FIG. 1 shows the main part of the projection exposure apparatus of this example. In FIG. 1, the exposure illumination light from the illumination optical system 40 at the time of exposure is a reticle R having a circuit pattern formed on its lower surface (pattern formation surface). Illuminate. The image of the pattern of the reticle R is projected by the projection optical system 5.
Is projected onto the photoresist layer in each shot area on the surface of the wafer 1 via the. Hereinafter, the surface of the wafer 1 will be described as the surface to be inspected 1a.

The wafer 1 is held on a holder 2, and the holder 2 is supported on a wafer stage 3 via three vertically movable fulcrums. The wafer stage 3 positions the wafer 1 in the plane perpendicular to the optical axis AX1 of the projection optical system 5 by the drive unit 4. Further, the driving unit 4 drives the three fulcrums on the wafer stage 3 in parallel to control the position (focus position) of the wafer 1 in the direction parallel to the optical axis AX1 by the autofocus method. Furthermore, the drive unit 4
Also controls (levels) the inclination angle of the holder 2 (wafer 1) by independently driving the three fulcrums.

In order to satisfactorily transfer the circuit pattern on the reticle R onto each shot area on the surface 1a to be tested of the wafer 1, the width of the depth of focus with respect to the image plane formed by the projection optical system 5 for each shot area. It is necessary to accommodate the current exposure area of the surface 1a to be inspected. For this purpose, after accurately detecting the focal positions at a plurality of detection points on the current exposure area of the surface 1a to be measured, the leveling of the holder 2 and the focusing direction of the wafer 1 (direction of the optical axis AX1) are performed. Just move.

Next, the optical system and processing system of the multipoint oblique incidence type AF sensor for detecting the focal positions at a plurality of detection points on the surface 1a of the wafer 1 to be detected will be described. FIG.
In, the illumination light that is non-photosensitive to the photoresist from the light source 6 that supplies white light with a wide wavelength width is converted into a substantially parallel light flux by the condenser lens 7 and is incident on the deflection prism 26. The deflection prism 26 deflects the substantially parallel light flux from the condenser lens 7 by refracting it. A transmission-type grating pattern plate 8 is provided on the exit side of the deflection prism 26, and a grating-pattern forming surface 8a of the transmission-type grating pattern plate 8 facing the wafer 1 is perpendicular to the plane of FIG. A stripe pattern extending in any direction, that is, a pattern in which transmissive portions and light-shielding portions are alternately arranged along the paper surface of FIG. 1 is formed. Instead of the transmissive grating pattern, an uneven reflective diffraction grating may be used, or a reflective grating pattern in which reflective portions and non-reflective portions are alternately formed may be used.

Since the test surface 1a of the wafer 1 is covered with a thin film of photoresist, the light source 6 is a white light source having a wide wavelength width in order to reduce the influence of the thin film interference. desirable. However, as the light source 6, a light emitting diode or the like that emits light in a wavelength band in which the photoresist has low sensitivity may be used. The illumination light that has been deflected by the deflection prism 26 and has reached the lattice pattern forming surface 8a is diffracted by the lattice pattern forming surface 8a, and then is angle θ with respect to the optical axis AX1 of the projection optical system 5.
The light enters the surface 1a to be inspected of the wafer 1 through the projection optical systems 9 and 10 arranged along the optical axis AX2 intersecting with each other. The projection optical systems 9 and 10 include a relay lens 9 and a projection objective lens 1.
0, and includes a lattice pattern forming surface 8a and a test surface 1a.
And are conjugated. At this time, the surface to be inspected 1a and the lattice pattern forming surface 8a meet the Scheimpflug condition with respect to the projection optical systems 9 and 10 in a state where the surface to be inspected 1a matches the image forming surface of the projection optical system 5. Since they are arranged so as to satisfy the condition, the image of the lattice pattern on the lattice pattern forming surface 8a is accurately formed on the entire surface of the surface to be inspected 1a.

The projection optical systems 9 and 10 are so-called double-sided telecentric optical systems, and the points on the grating pattern forming surface 8a and the conjugate points on the surface 1a to be tested have the same magnification on the entire surface. Therefore, in this example, since the lattice pattern on the lattice pattern forming surface 8a has a lattice shape with the longitudinal direction in the direction perpendicular to the plane of FIG. 1, the image formed on the surface 1a to be inspected is It becomes a grid-like pattern at equal intervals with the direction perpendicular to the paper surface of FIG. 1 as the longitudinal direction.

Returning to FIG. 1, the illumination light projected on the surface 1a to be inspected is reflected by the surface 1a to be inspected and the condensing optical systems 11, 1 are collected.
2 is incident. The condensing optical systems 11 and 12 include a light-receiving objective lens 11 and a relay lens 12, and the optical axes AX3 of the condensing optical systems 11 and 12 are the optical axes AX of the projection optical system 5.
1 is arranged so as to be line-symmetric with the optical axis AX2 of the irradiation optical systems 9 and 10. Also, the light-receiving objective lens 1
1 and the relay lens 12, a vibrating mirror 30 as a scanning unit is provided, and the vibrating mirror 30 is a mirror driving unit 3.
1 oscillates periodically around an axis perpendicular to the plane of the paper of FIG. An optical axis obtained by bending the optical axis AX3 with the vibrating mirror 30 (in a state of being at the center of vibration) is represented as an optical axis AX4.

The illumination light reflected by the surface to be inspected 1a passes through the light-receiving objective lens 11, is reflected by the vibrating mirror 30, and
It is condensed by the condenser lens 12, and the half-wave plate 24
(This action will be described later)
An image of a grid pattern is formed on the top. The light receiving slit S is formed on the incident surface 25a of the tilt correcting prism 25, and the illumination light passing through the opening of the light receiving slit S enters the tilt correcting prism 25. At this time, in a state where the surface to be inspected 1a matches the image forming surface of the projection optical system 5, the surface to be inspected 1a and the entrance surface 25a of the tilt correction prism 25 (the surface on which the light receiving slit S is formed) are collected together. The optical optical systems 11 and 12 are arranged so as to satisfy the Scheimpflug condition. Therefore, in this state, the image of the lattice pattern on the surface 1a to be inspected is accurately re-formed on the entire surface of the incident surface 25a. Further, since the condensing optical systems 11 and 12 are also composed of both-side telecentric optical systems, each point on the test surface 1a and the conjugate point on the incident surface 25a have the same magnification on the entire surface. Therefore,
In the state where the surface to be inspected 1a coincides with the image forming surface of the projection optical system 5, the image projected on the incident surface 25a (light receiving slit S) also has the longitudinal direction in the direction perpendicular to the paper surface of FIG. It becomes a grid pattern of intervals.

That is, in this example, when the surface to be inspected 1a is aligned with the image plane of the projection optical system 5, the grating pattern forming surface 8a, the surface to be inspected 1a, and the incident surface 25a of the tilt correction prism 25 are Each of them has a relationship of Scheimpflug condition, and each surface has the same magnification over the entire surface. Further, the light receiving surface 17a of the photoelectric detector 17 is arranged in proximity to the exit surface 25b of the tilt correction prism 25. The photoelectric detector 17 of this example has a plurality of independent photoelectric conversion elements arranged on a substrate as described later. In FIG. 1, the oscillating mirror 31 includes the condensing optical systems 11 and 12.
However, the vibrating mirror 31 may be located anywhere in the optical path between the surface to be tested 1a and the light receiving slit S. Here, the light receiving slit S and the surface to be inspected 1a
The pattern image projected on the screen and the configuration of the photoelectric detector 17 will be described.

First, FIG. 2 shows a light receiving slit S as a light shielding means. In FIG. 2, the light receiving slit S has five openings Sa1, Sa2, Sa3, Sa in the light shielding portion.
4 and Sa5 are formed. These openings Sa
1 to Sa5 correspond to the detection points of the focal position on the surface 1a to be inspected, and the number and arrangement of the detection points on the surface 1a to be inspected can be arbitrarily set by the number and arrangement of the openings Sa1 to Sa5.

FIG. 3 shows a pattern image projected on the surface 1a to be inspected of the wafer 1. In FIG. 3, a pattern image is formed on the surface 1a to be inspected in a rectangular area almost equal to a shot area to be exposed. An image 80a of the lattice pattern formed on the lattice pattern forming surface 8a is projected. The image 80a is
It is a striped pattern in which slit-like bright portions and dark portions are arranged at a constant pitch in a predetermined direction, and detection areas Da1 and Da that are conjugate with the five openings Sa1 to Sa5 of FIG. 2 in the image 80a.
Based on the striped patterns in 2, Da3, Da4, and Da5, average focus positions are detected in these five detection areas Da1 to Da5. Detection areas Da1 to Da
Since 5 is an area that can be regarded as a point when compared with the entire image 80a, each of the detection areas Da1 to Da5 can be regarded as a “detection point”.

FIG. 5 shows the light receiving surface 1 of the photoelectric detector 17 of this example.
7a shows the arrangement of the photoelectric conversion elements in FIG. 7, and in FIG. 5, five openings Sa1 to S of the light receiving slit S of FIG.
The photoelectric conversion elements S formed by the silicon photodiodes on the light receiving surface 17a of the illumination light that has passed through a5, respectively.
It is configured to be incident on PD1, SPD2, SPD3, SPD4, SPD5. In this case, the size of the photoelectric conversion elements SPD1 to SPD5 and the size of the photoelectric conversion elements SPD1 to SPD5 are set so that the illumination light that has passed through the openings Sa1 is incident on the photoelectric conversion element SPD1, for example. The interval is set.

Next, returning to FIG. 1, since the lattice pattern image projected on the surface to be inspected 1a is re-imaged on the light receiving surface 17a of the photoelectric detector 17, the projection optical system 5 for the surface to be inspected 1a is formed. The vertical movement along the optical axis AX1 causes the lattice pattern image on the light receiving surface 17a to laterally shift. By detecting the amount of lateral deviation, the focus positions at the five detection points on the surface 1a to be measured can be measured. That is, the detection signals from the five photoelectric conversion elements of the photoelectric detector 17 are detected by the detection unit 18 as the detection output SD.
Drive signal to the oscillating mirror 30 from the mirror drive unit 31 is also supplied to the detection unit 18. Then, the detection unit 18 synchronously rectifies the detection output SD with the drive signal, so that the detection output SD is 5 within the predetermined range.
Five focus signals that change substantially linearly according to the focus positions of the detection points (detection areas Da1 to Da5) are generated, and these focus signals are supplied to the correction amount calculation means 19. The method of generating the focus signal by such synchronous rectification is based on the principle of the photoelectric microscope disclosed in, for example, Japanese Patent Laid-Open No. 56-42205 (details will be described later).

In this case, the focus signals are previously calibrated so that they are 0 when the surface 1a to be detected matches the image plane of the projection optical system 5. Therefore, the correction amount calculation means 19 uses the focus signals to calculate the shift of the focal position (defocus amount) and the shift of the tilt angle between the current exposure region of the surface to be inspected 1a and the image plane of the projection optical system 5. Then, the defocus amount and the deviation of the tilt angle are brought close to 0 via the drive unit 4. This is the auto-focusing and auto-leveling operation, and the exposure of the image of the pattern of the reticle R is started when the defocus amount at each detection point falls within the width of the depth of focus. Then, even during the exposure, the operation of bringing the defocus amount and the deviation of the tilt angle closer to 0 by the servo method is continued.

Next, the operation of the tilt correction prism 25 and the like of this example will be described in detail. For that purpose, first, the surface to be inspected 1
The lateral shift amount y of the image of the lattice pattern on the incident surface 25a of the tilt correction prism 25 when the displacement of a in the optical axis AX1 direction is z is obtained. In this case, the incident angle of the principal ray of the incident light on the surface 1a to be inspected is θ, and the condensing optical system 1
The lateral magnifications of 1 and 12 are δ, and the incident surface 25 from the test surface 1a
When the magnification along the image plane of the tilt to a is δ ′, the lateral displacement amount y is as follows from the above equations (1) and (2).

Y = 2 · δ ′ · tan θ · z = 2 (δ ² sin ² θ + δ ⁴ sin ⁴ θ / cos ² θ) ^1/2 · z (7) That is, the lattice pattern image projected on the surface 1 a to be inspected If the incident angle θ of the illumination light is increased, the lateral deviation amount y is also increased, which enables highly accurate focus position detection. By the way, the irradiation optical systems 9 and 10 and the condensing optical systems 11 and 12 satisfy the Scheimpflug condition, respectively. Therefore, the angle formed by the normal line of the transmissive grating pattern plate 8 and the emitted light is φ, and the angle formed by the normal line of the tilt-correction prism 25 and the incident light is α (this angle α is the incident surface 25a of the tilt-correction prism 25). And the lateral magnification of the irradiation optical systems 9 and 10 is δ ₄ , the condensing optical system 11 and
Since the lateral magnification of 12 is δ, the following relationship holds.

Tan φ = δ ₄ tan θ (8) tan α = δ · tan θ (9) By the way, when the light receiving surface 17a is provisionally arranged on the incident surface 25a, the incident angle of the illumination light with respect to the surface 1a to be inspected. When θ is large, the incident angle α of the light flux on the light receiving surface 17a also becomes large according to the equation (9). Therefore, the surface reflection of the photoelectric conversion element on the light receiving surface 17a becomes large, and the light flux may be eclipsed, resulting in a significant decrease in the amount of light received. In this example, in order to avoid such a decrease in the amount of received light, an inclination correction prism 25 having an incident surface 25a is arranged as a deflection member for inclination correction, and the light flux emitted from the condensing optical systems 11 and 12 is arranged. Biased.

FIG. 4 is an enlarged side view showing the tilt correction prism 25. In FIG. 4, the apex angle ξ of the tilt correction prism 25 is the refractive index n of the tilt correction prism 25 and the incident surface 25a. It is set to satisfy the following expression by using the incident angle α of the principal ray of the illumination light from the inspection surface 1a. ξ ≦ arcsin {(sin α) / n} (10) At this time, the light ray L1 incident on the incident surface 25a at the incident angle α.
Let β be the refraction angle when refracted by the tilt-correction prism 25, and let γ be the exit angle when a ray of this refraction angle β exits from the exit surface 25b of the tilt-correction prism 25. The angle β is

Γ = arcsin {nsin (β-ξ)} (11) β = arcsin {(sinα) / n} (12) The photoelectric detector 17 of FIG. A light receiving surface 17a is installed. Further, since the light-receiving slit S is formed on the entrance surface 25a, the light flux transmitted through the opening of the light-receiving slit S reaches each photoelectric conversion element of the photoelectric detector 17. At that time, if the divergence angle of the light flux is a certain size, the light flux that has passed through one opening of the light-receiving slit S on the light-receiving surface 17a of the photoelectric detector 17 is diverged, and the photoelectric light shown in FIG. Conversion element SPD1
~ There is a risk of mixing in the adjacent photoelectric conversion elements in SPD5. In order to avoid it, the entrance surface 25a and the exit surface 2
It is necessary to make the distance to 5b as small as possible. FIG.
, The effective length of the entrance surface 25a (the length of the range used for surface position detection) is w1, the entrance surface 25a and the exit surface 2
Assuming that the maximum distance from 5b is e, the actual air-converted length e ′ of the maximum distance e is as follows.

E ′ = w1 · sin ξ / n (13) In order to reduce the air-converted length e ′, it is necessary to reduce the apex angle ξ. However, as in the equation (11), the emission angle γ becomes large this time, the incident angle of the light beam to the photoelectric detector 17 becomes large, and the light receiving efficiency deteriorates. From the above, (1
If the apex angle ξ is set so that the condition of expression (0) is satisfied and the emission angle γ does not become too large, the photoelectric detector 17 may be installed close to the tilt correction prism 25 without any problem.

Here, in FIG. 4, assuming that the emergence angle γ of the light ray L2 is almost 0, and considering the rise of the image by the tilt-correction prism 25, the emitted light ray L
The tilt angle ρ, which is the inclination of the relayed image plane of the surface 1a to be examined with respect to the plane perpendicular to 2, is expressed by the following equation. ρ = arctan (tan ξ / n) (14) For example, the incident angle θ of the illumination light with respect to the test surface 1a is 85 °,
Assuming that the lateral magnification δ of the condensing optical systems 11 and 12 is 0.5, the incident angle α of the illumination light to the tilt correction prism 25 in FIG.
Is 80.1 °. Then, the tilt correction prism 25
If the refractive index n of is 1.8, the tilt angle ρ = 20.0 °
And the tilt angle ρ approaches 0 °. Further, since the incident angle to the light receiving surface 17a (that is, the exit angle γ at the exit surface 25b) is almost 0 °, it can be considered that there is almost no decrease in the amount of received light. As described above, in this example, since the light beam passing through the light receiving slit S is received by the photoelectric detector 17 via the tilt correction prism 25, the incident angle of the light beam on the light receiving surface 17a becomes small, and the light receiving amount is reduced. It can prevent the deterioration.

In FIG. 1, when the incident angle θ of the illumination light with respect to the surface 1a to be detected is large, the incident angle α with respect to the incident surface 25a also becomes large, and the tilt-correction prism 25
The transmittance of the light flux passing through the P-polarized component and the S-polarized component greatly differ. For example, the incident angle α = 80 °,
When the refractive index n = 1.8, the transmittance of the P-polarized component is 0.79 and the transmittance of the S-polarized component is 0.37. As described above, since the transmittance of the tilt correction prism 25 differs depending on the polarization state of the light flux, the weight of information is different for each polarization component, which may result in inaccurate detection of the focus position of the surface to be inspected 1a. In order to cope with this, in this example, a ½ wavelength plate 24 is arranged between the focusing optical systems 11 and 12 and the tilt correction prism 25, and the ½ wavelength plate 24 rotates the polarization direction by 45 °. Is incident on the tilt correction prism 25. For this reason, the light beam incident on the tilt correction prism 25 is in a state in which the P-polarized component and the S-polarized component are mixed, and the surface position of the surface 1a to be detected is correctly detected.

In place of the half wave plate 24, 1 /
A four-wave plate may be used, and when a quarter-wave plate is used, the incident light flux on the tilt correction prism 25 becomes circularly polarized light, and like the case where a half-wave plate is used, the surface 1a to be inspected is changed. The surface position is accurately detected. In this example, the surface position is detected by the principle of the photoelectric microscope as described above.
Therefore, an example of the surface position detection method based on the principle of the photoelectric microscope in this example will be described in detail. In FIG.
The mirror driving unit 31 vibrates the vibrating mirror 30 in the condensing optical systems 11 and 12 in the arrow direction at a predetermined cycle T based on a signal from an internal oscillator. With the vibration of the vibrating mirror 30, the incident surface 25a of the tilt correction prism 25
The lattice pattern image formed on the (light receiving slit S) also vibrates. At this time, the period of vibration of the grid pattern imaged on the incident surface 25a is specified to be 1/2 or less of the pitch of the grid pattern, and the openings Sa1 to Sa1 of the light receiving slit S shown in FIG. The width of Sa5 in the detection direction is also defined to be 1/2 or less of the pitch of the grid pattern image. Along with the vibration of the lattice pattern image, the amount of light flux passing through the openings Sa1 to Sa5 of the incident surface 25a (light receiving slit S) changes, and the light fluxes are converted into photoelectric conversion elements of the photoelectric detector 17 of FIG. 5, respectively. It is incident on SPD1 to SPD5.

A method of processing a detection signal from one photoelectric conversion element SPD1 will be described below, but detection signals from other elements are processed in the same manner. Almost all of the light flux that has passed through the opening Sa1 of the light receiving slit S is the photoelectric conversion element S.
It is incident on PD1. The amount of light passing through the opening Sa1 changes as the vibrating mirror 30 vibrates. This change in light quantity will be described with reference to FIG. FIG. 8 is an enlarged view of the vicinity of the opening Sa1 of the light-receiving slit S. In FIG. 8, an image ST of the grid pattern has a dark portion shown by hatching and a bright portion other than that repeated at a pitch PST. It is a statue. The image ST of the grid pattern vibrates in the direction indicated by the arrow in FIG. 9, that is, in the short side direction of the opening Sa1 due to the vibration of the vibrating mirror 30.

At this time, the width WSa1 of the opening Sa1 is
The pitch P of the image ST of the grid pattern formed on this
It is specified that ST satisfies the following conditions. WSa1 ≦ PST / 2 (15) Further, the amplitude AST of the image ST of the grid pattern vibrating by the vibrating mirror 30 is defined so as to satisfy the following condition.

AST ≦ PST / 2 (16) When the width WSa1 of the opening Sa1 does not satisfy the expression (15), or the amplitude AST of the image ST of the lattice pattern.
Does not satisfy the expression (16), the change in the amount of light at the opening Sa1 due to the vibration of the vibrating mirror 30 becomes small, and the detection accuracy decreases, which is not preferable.

The position of the opening Sa1 is determined by the position of the surface 1a to be measured.
And the image forming plane of the projection optical system 5 coincide with each other, the center of the opening Sa1 is determined so as to coincide with the vibration center of one bright portion (or dark portion) in the image ST of the grid pattern. ing. Since the width WSa1 of the opening Sa1 is 1/2 or less of the pitch PST of the image ST of the grid pattern and the amplitude AST of the image ST is also 1/2 or less of the pitch PST, the grid pattern When the image ST vibrates,
The amount of light received on the photoelectric conversion element SPD1 changes. Then, the photoelectric conversion element SPD1 outputs the detection output SD1 corresponding to the change in the light intensity, that is, the light modulation, to the detection unit 18.

Similarly, other photoelectric conversion elements SPD2 to SPD
Also in D5, the detection outputs SD2 to SD5 corresponding to the amount of light having passed through the corresponding openings Sa2 to Sa5 are detected by the detection unit 18.
Output to Further, the mirror driving section 31 also outputs an AC signal having the same phase as the vibration cycle T of the vibrating mirror 30 to the detecting section 18. Then, in the detection unit 18, the detection outputs SD1 to S are based on the phase of the AC signal of the period T.
Synchronous rectification (synchronous detection) of D5 is performed, and the focus signals FS1 to FS5, which are the detection outputs, are corrected by the correction amount calculation unit 19.
Is output to The focus signals FS1 to FS5 are
This is a so-called S-curve signal, and when the detection areas Da1 to Da5 of FIG. 3 on the surface to be detected 1a are located on the image forming surface of the projection optical system 5, that is, the detection outputs SD1 to SD5 respectively vibrate the vibration mirror 5. When it changes in a cycle of 1/2 of the cycle T, it becomes 0 level. Further, the detection area D of the surface 1a to be inspected
When a1 to Da5 are located above the image plane of the projection optical system 5, a positive level is set. When detection areas Da1 to Da5 are located below the image plane of the projection optical system 5, a negative level is set. Show. In this way, the focus signals FS1 to FS1
FS5 indicates an output value corresponding to the focus position of the surface to be inspected 1a, and changes substantially linearly with a change in the focus position within a predetermined range. Then, as described above, the correction amount calculation unit 19 and the drive unit 4 perform autofocus and autoleveling.

As described above, in the multi-point oblique incidence type AF sensor of this example, the image of the lattice pattern re-formed by the condensing optical systems 11 and 12 is passed through the light receiving slit S and the tilt correction prism 25. The light is directly received by the photoelectric detector 17. Therefore, as compared with the case where a relay optical system is provided between the tilt correction prism 25 and the photoelectric detector 17, for example, the focus position of the surface to be inspected 1a of the wafer 1 can be detected with a simple configuration. Moreover, since there is no relay optical system, adjustment of the optical system is easy.

In the above-described embodiment, as shown in FIG. 9, on one opening Sa1 (photoelectric conversion element SPD1).
One bright line or dark line (hereinafter referred to as “slit image”) of the lattice pattern image is formed on the upper side), but a plurality of slit images of the lattice pattern image are formed on one opening. You may form. FIG. 10 shows an example in which a plurality of slit images are formed on one opening Sa6 on the light receiving slit S. In FIG. 10, three slits W6 of width W6 in the image of the lattice pattern are formed on the opening Sa6. Slit images SL6a, SL
6b and SL6c are formed at a pitch P6. At this time, the pitch P6 is the pitch PST of FIG.
It is desirable that the following relation be satisfied.

P6 = m · PST (17) Further, the width W6 of the slit images SLa, SLb, SLc is
It is desirable to satisfy the following relationships. W6 ≦ PST / 2 (18) Here, when the pitch P6 of the slit images SL6a, SL6b, SL6c deviates from the equation (17), each slit image SL
6a, SL6b, and SL6c have different light quantity changes due to the vibration of the vibrating mirror 30, which unfavorably reduces the reliability of the detected value. In addition, the slit images SL6a, S
When the width W6 of L6b and SL6c deviates from the equation (18),
This is not preferable because the change in the amount of light accompanying the vibration of the vibrating mirror 30 becomes small.

According to the configuration of FIG. 10, in addition to the advantage that the amount of light received on the photoelectric conversion element corresponding to the opening Sa6 is increased, the corresponding detection area on the surface 1a to be detected is widened and the averaging effect is obtained. Is obtained. Further, in this example, the light receiving slit surface S is arranged on the incident surface 25a of the tilt correction prism 25, but this is the photoelectric detector 17
It may be on the light receiving surface 17a. At this time, the sizes of the openings Sa1 to Sa5 of the light receiving slit S are determined by the light receiving surface 17
It is desirable to be provided so as to be smaller than the photoelectric conversion elements SPD1 to SPD5 on a. In this example, the plurality of photoelectric conversion elements SPD1 to SPD5 each of which is composed of a silicon photo diode shown in FIG.
It may be provided on a by a photo-etching method. In this case, there is an advantage that even if the number of the openings Sa1 to Sa5 of the light receiving slit S is significantly increased, it is possible to cope with it.

The surface of the wafer 1 as the surface to be inspected 1a is not always flat depending on the process. In such a case, the number of detection areas (detection points) on the surface 1a to be detected may be increased. In order to increase the number of detection points in this way, the number of openings of the light-receiving slit S corresponding to the detection points may be increased. As described above, since the AF sensor of this example can increase the number of detection points on the surface 1a to be tested with a simple configuration, the focus position over a wide range on the surface 1a to be tested can be accurately measured. Moreover, since they can be detected at the same time, there is no reduction in throughput.

Next, in the example of FIG. 1, the tilt correction prism 2
5, the light receiving surface 1 of the photoelectric detector 17 is close to the exit surface 25b of
7a is arranged, but as shown in FIG. 6, a light transmissive medium 41 having a refractive index larger than that of air, that is, a refractive index larger than 1 is filled between the exit surface 25b and the light receiving surface 17a. You may. As the medium 41, a light-transmissive resin adhesive or the like can be used. In the configuration example of FIG. 6, the light ray L2 emitted from the emission surface 25b of the tilt correction prism 25
The incident angle with respect to the photoelectric detector 17 is the light ray L when there is no medium 41, that is, when there is an air layer in place of the medium 41.
It is smaller than 2 '. Therefore, the photoelectric detector 1
The light amount loss in 7 is smaller.

Further, in the above embodiment, the tilt correction prism 25 is used as the deflection member for tilt correction, but a diffraction grating, a decentering lens or the like may be used instead. FIG. 7A shows an example in which a blazed and phase type diffraction grating 42 is used as a deflection member for tilt correction, and in FIG.
Similar to the example of FIG. 1, the image of the grid pattern projected on the surface 1a of the wafer W to be inspected is re-imaged thereon. And
The phase type diffraction grating 42 is arranged on the bottom surface of the light receiving slit S, and the light receiving surface of the photoelectric detector 17 is arranged so as to face the exit surface of the diffraction grating 42. Further, the diffraction grating 42 is formed so that the 0th-order light becomes almost 0, and the diffracted light of a predetermined order in the direction to reduce the tilt angle of the incident light beam is incident on the photoelectric detector 17. . In this example, since the tilt angle of the light beam with respect to the photoelectric detector 17 can be made small, and the light receiving surface of the photoelectric detector 17 and the surface on which the image of the lattice pattern is formed (here, the light receiving slit S) can be substantially brought into close contact with each other, each detection is performed. There is an advantage that the light fluxes corresponding to the points can be prevented from mixing with each other.

Next, FIG. 7B shows an example in which a decentering lens 43 is used as a deflection member for tilt correction.
In (b), the eccentric lens 4 is provided on the bottom surface of the light receiving slit S.
3 is arranged, and the light receiving surface of the photoelectric detector 17 is arranged so as to face the exit surface of the decentering lens 43. In this case, the eccentric lens 43 is a lens obtained by cutting out the upper half from the optical axis of an axially symmetric lens, and the light flux passing through the light-receiving slit S and focused by the eccentric lens 43 has a small tilt angle and thus the photoelectric detector. It is incident on the light receiving surface of 17. Therefore, the light receiving efficiency is improved.

Further, in the example of FIG. 1, the tilt correction prism 2
Although the light receiving surface 17a of the photoelectric detector 17 is arranged on the exit surface 25b of No. 5, the light flux of the exit surface 25b may be guided to the light receiving surface by an optical fiber. FIG. 8 shows an example using such an optical fiber. In FIG.
A tilt correction prism 2 having a light receiving slit S formed in a
5, the optical fibers 44A, 44B, 44C on the exit surface 25b
Is installed at one end. It should be noted that the light receiving slit S shown in FIG.
When the five openings Sa1 to Sa5 are formed as shown in FIG. 5, the optical fibers 44A, 44B, 4 are arranged so as to receive the light flux passing through these openings Sa1 to Sa5.
Five 4C are also installed. Then, the optical fibers 44A, 4
The other ends of 4B and 44C are installed on the light receiving surfaces of photomultipliers 45A, 45B and 45C as photoelectric detectors, respectively. In this example as well, the photo multiplier 45
By processing the photoelectric conversion signals of A, 45B, and 45C, the surface position of the surface 1a to be detected is detected as in the example of FIG. Moreover, in the example of FIG. 8, since the optical fiber is used, there is an advantage that the photomultipliers 45A, 45B and 45C which are large in size but have high sensitivity can be used as the photoelectric detector.

The present invention is not limited to the above-mentioned embodiments, and it goes without saying that various configurations can be made without departing from the gist of the present invention.

[0057]

According to the surface position detecting apparatus of the present invention, the second
Since the light receiving surface of the detector is arranged close to the exit surface of the deflection member for tilt correction that deflects the light beam from the condensing optical system on the surface, the tilt angle can be made small to efficiently cover the detector. Since the light flux from the inspection surface can be made incident and the Scheimpflug condition is satisfied, two-dimensional surface position information can be obtained. Further, as compared with the case where the relay optical system is arranged after the deflecting member, there are advantages that the configuration is simple and small, and the adjustment of the optical system is easy.

Next, when the deflecting member is a prism, the structure is simple, and even when white light is used as the detection light beam because the dispersion is small, there is little mixing of light beams corresponding to different detection points. There are advantages. Also,
The vertex angle of the prism is ξ, the refractive index of the prism is n,
When the incident angle of the principal ray of the light beam entering the prism from the condensing optical system is α, and the apex angle ξ satisfies the relationship of equation (4), the tilt angle of the light beam entering the detector can be made small and , Since the distance between the re-imaging surface of the pattern image projected on the surface to be inspected and the light receiving surface of the detector is kept small, the mixture of light beams corresponding to different detection points is extremely small, and the surface position There is an advantage that the detection accuracy of is improved.

Further, when a light transmissive medium having a refractive index of more than 1 is filled between the light exit surface of the deflecting member and the light receiving surface of the detector, the light receiving efficiency of the detector is further improved. . Further, according to the exposure apparatus of the present invention, since the focal position of the photosensitive substrate is detected with high accuracy by the surface position detection apparatus of the present invention, the surface of the photosensitive substrate is accurately detected as the image plane of the projection optical system. Exposure can be performed in combination.

[Brief description of drawings]

FIG. 1 is a configuration diagram showing a projection exposure apparatus provided with an example of an embodiment of a surface position detection apparatus according to the present invention.

FIG. 2 is an enlarged view showing an array of openings of a light receiving slit S shown in FIG.

3 is a perspective view showing an image of a grid pattern projected on a surface 1a to be tested of the wafer 1 of FIG.

FIG. 4 is an enlarged side view showing the tilt correction prism 25 of FIG.

5 is an enlarged view showing an arrangement of photoelectric conversion elements of the photoelectric detector 17 of FIG.

6 is an enlarged view of a main part showing an example in which a medium having a refractive index larger than 1 is filled between the tilt correction prism 25 and the photoelectric detector 17. FIG.

7A is an enlarged view of a main part showing a case where a diffraction grating is used as a deflecting member, and FIG. 7B is an enlarged view of a main part showing a case where a decentering lens is used as a deflecting member.

FIG. 8 is a configuration diagram of a main part showing an example in which an optical fiber is installed between the tilt correction prism 25 and a photoelectric detector.

9 is one opening Sa1 on the light receiving slit S of FIG.
FIG. 6 is an enlarged view showing a relationship between the image of the grid pattern projected on the image and the image of the grid pattern projected on the image.

FIG. 10 is a diagram showing another example of an image of a lattice pattern.

FIG. 11 is an explanatory diagram of Scheimpflug conditions.

[Explanation of symbols]

 1a Surface to be inspected 5 Projection optical system 8 Transmission grating pattern plate 9 Relay lens 10 Objective lens for projection 11 Objective lens for light reception 12 Relay lens 17 Photoelectric detector 25 Aorientation correction prism 30 Vibration mirror 31 Mirror drive unit S Light receiving slit Sa1 Sa5 openings Da1 to Da5 detection areas (detection points)

─────────────────────────────────────────────────── ─── Continuation of the front page (51) Int.Cl. ⁶ Identification code Office reference number FI technical display location H01L 21/027 G02B 7/11 M // G02B 21/26 H01L 21/30 526B

Claims (5)

[Claims] 1. A predetermined pattern formed on a first surface, a projection optical system for projecting an image of the predetermined pattern from a direction oblique to the surface to be inspected, and a reflection on the surface to be inspected. A condensing optical system that condenses the light flux to form an image of the predetermined pattern on the second surface, and a detector that photoelectrically detects the image of the predetermined pattern formed by the condensing optical system. And a surface position detecting device for detecting a surface position of the surface to be inspected based on an output of the detector, wherein the first surface and the surface to be inspected are shines with respect to a main plane of the projection optical system. Satisfies the proof condition, the test surface and the second surface satisfy the Scheimpflug condition with respect to the main plane of the condensing optical system, and the light receiving surface of the detector and the image of the predetermined pattern are formed. Scanning means for relatively scanning, substantially on the second surface And a deflection member for tilting correction for deflecting the light beam from the condensing optical system in a predetermined direction, and the light receiving surface of the detector is arranged in the vicinity of the light beam emission surface of the deflection member. A surface position detecting device. 2. The surface position detecting device according to claim 1, wherein the deflecting member is a prism. 3. The surface position detecting device according to claim 2, wherein the apex angle of the prism is ξ, the refractive index of the prism is n,
The shape of the prism is determined so that the apex angle ξ satisfies the relation of ξ ≦ arcsin {(sinα) / n}, where α is the incident angle of the principal ray of the light beam entering the prism from the condensing optical system. A surface position detecting device characterized by: 4. The surface position detecting device according to claim 1, 2, or 3, wherein a light transmission having a refractive index larger than 1 is provided between an exit surface of the light flux of the deflecting member and a light receiving surface of the detector. A surface position detecting device characterized by being filled with a permeable medium. 5. The surface position detecting device according to claim 1, a projection optical system for projecting an image of a mask pattern onto a photosensitive substrate, and an optical axis of the projection optical system of the photosensitive substrate. A surface position adjusting stage for adjusting the position in the direction, and an exposure apparatus, which detects the surface position of the photosensitive substrate by the surface position detecting device, and through the surface position adjusting stage based on the detection result. An exposure apparatus for adjusting the surface position of the photosensitive substrate.