JPH1082611A - Apparatus for detecting position of face - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a face position-detecting apparatus which can be adjusted easily in a simple constitution and detect the position of a face highly accurately. SOLUTION: The image of a grating pattern of a grating pattern formation face 8a is, via projection optical systems 9, 10, projected slantwise to a face 1a to be detected on a wafer 1. The grating pattern image is formed from a reflecting light from the face 1a onto a light incident plane 25a via condenser optical system 11, 12, which is relayed to a photodetecting face 17a through a swing and tilt correction prism 25 and relay optical systems 14-16. The grating pattern formation face 8a and face 1a to be detected satisfy a shine-proof condition to the projection optical systems 9, 10, and the face 1a and the light incident plane 25a satisfy the shine-proof condition to the condenser optical systems 11, 12. The projection optical systems 9, 10 are made telecentric to the side of the face 1a, and a numerical aperture of the condenser optical system 11, 12 is set to be larger than that of the projection optical system 9. 10.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a surface position detecting device for detecting a height and an inclination angle of a surface to be inspected, for example, a semiconductor device, an image pickup device (such as a CCD), a liquid crystal display device, a thin film magnetic head, and the like. It is suitable for use in an autofocus sensor for detecting the height of a photosensitive substrate such as a wafer in a projection exposure apparatus used in a photolithography process for manufacturing a semiconductor device.

[0002]

2. Description of the Related Art For example, when manufacturing a semiconductor device, a projection exposure apparatus (stepper or the like) for transferring a circuit pattern formed on a reticle as a mask onto a wafer as a photosensitive substrate via a projection optical system is used. Have been. In such a projection exposure apparatus, the depth of focus of the projection optical system is relatively shallow, and there may be partial unevenness on the wafer. It is necessary to correct each defocus.

As described above, as a sensor for detecting the position (focus position) of the surface of the wafer (surface to be inspected) in the optical axis direction of the projection optical system, a slit image is projected obliquely to the surface of the wafer. And a condensing optical system that condenses the reflected light from the wafer and re-images the slit image, and detects the lateral shift amount of the slit image thus re-imaged. An oblique incidence type autofocus sensor (hereinafter, referred to as an “AF sensor”) is known. In the oblique incidence type AF sensor, when the surface to be inspected moves up and down along the optical axis direction of the projection optical system, the position of the slit image on the surface to be inspected moves along the direction orthogonal to the optical axis of the light collecting optical system. Deviate. Therefore, by measuring the lateral shift amount of the slit image re-imaged by the focusing optical system, it is possible to detect the focus position on the surface to be inspected.

In this case, the surface to be inspected is inclined at a considerably large angle with respect to the optical axes of the projection optical system and the condensing optical system of the AF sensor. When the image is re-imaged, the tilt angle of the image plane of the re-imaged slit image becomes large, and there is a problem that the amount of light incident on the light receiving element becomes small. Therefore, Japanese Patent Application Laid-Open No. 6-97045 discloses an AF sensor in which a prism for correcting the tilt angle is used to reduce the tilt angle of the re-formed slit image. Japanese Patent Application Laid-Open No. 3-183904 (U.S. Pat. No. 5,076,698) discloses a detection device that corrects the tilt angle of an imaged light beam using a grating (diffraction grating).

[0005]

An oblique incidence type AF disclosed in the above-mentioned conventional Japanese Patent Laid-Open Publication No. Hei 6-97045.
In the sensor, it is desirable that both the projection optical system and the condensing optical system be both-side telecentric so that the magnification of the re-formed slit image does not change even if the position of the test surface changes. For this purpose, it is necessary to dispose a telecentric stop in the projection optical system and the condensing optical system with the surface to be measured interposed therebetween. May remain. However, if a positioning error remains between the two diaphragms, there is a disadvantage that a light amount loss occurs.

Scanning means such as a vibrating mirror disposed in the condensing optical system for detecting the amount of lateral displacement of the slit image re-imaged by the condensing optical system by applying the principle of a photoelectric microscope. May cause the slit image to vibrate in the measurement direction with respect to the photoelectric detector. In such a configuration, it is often difficult to place a telecentric stop in the condensing optical system. On the other hand, if a scanning device is provided on the projection optical system side, the slit image moves on the surface to be inspected, so that the state of the lateral shift of the re-imaged image is also changed by the unevenness of the surface to be inspected, and the lateral shift of the image is changed. There is an inconvenience that it is difficult to separate the vibration component into a vibration component of the vibration mirror and a component due to the unevenness of the surface to be measured.

Further, even when the grating disclosed in Japanese Patent Laid-Open No. 3-183904 is used for correcting the tilt angle in the AF sensor, the projection optical system, In addition, there is an inconvenience that it is difficult to dispose a telecentric stop in both the optical system and the condensing optical system. In view of the above, the present invention has a simple configuration without disposing a telecentric stop in series with both the projection optical system and the condensing optical system, is easy to adjust, and has high accuracy in the surface position of the test surface. It is an object of the present invention to provide a surface position detecting device capable of detecting a position.

[0008]

A surface position detecting device according to the present invention is provided with a predetermined pattern (8) formed on a first surface (8a) and an oblique direction with respect to a surface (1a) to be detected. Projection optical system (9, 10) for projecting the image of the predetermined pattern
A condensing optical system (11, 12) for condensing a light beam reflected on the surface to be inspected (1a) and forming an image of the predetermined pattern on the second surface (25a); And a detector (17) for photoelectrically detecting the image of (1), wherein the detector (17) detects a surface position of the test surface (1a) based on an output of the detector. Scanning means (30, 31) for relatively scanning the light-receiving surface of the image and the image of the predetermined pattern, and the first surface and the test surface (1a) are arranged with respect to the main plane of the projection optical system. In addition to satisfying the Scheimpflug condition, the test surface (1a) and the second surface satisfy the Scheimpflug condition with respect to the main plane of the condensing optical system, and the projection optical system includes the test surface (1a). Side is substantially telecentric optics, the focusing optics Those with a numerical aperture greater than a projection optical system.

According to the present invention, the projection optical system (9,
10) is telecentric on the test surface (1a) side.
This operation will be described with reference to FIG. FIG.
Point P on the surface to be inspected (1a) by the projection optical system (9, 10)
FIG. 10A shows the state of the image forming light beam LA converged on the optical axis, and FIG.
(B) shows a case where the projection optical system is not telecentric. That is, in FIG. 10A, the test surface (1
Off-axis chief ray LA1 of the imaging light beam LA incident on a)
Is also a light beam parallel to the optical axis. In FIG. 10B, the off-axis principal ray LA2 is inclined with respect to the light beam LA1 parallel to the optical axis in the image forming light beam.

At this time, it is assumed that the focusing optical system (11, 12) is in focus on the original surface (1a), and the surface (1a) is moved to the position (1a ') indicated by the dotted line. When displaced,
In the case of the telecentric case shown in FIG. 10 (a), the point P on the surface to be inspected (1a) is formed on the second surface by the condensing optical system by the image forming light beam passing through the point P 'which is a mirror image of the point P. A defocused image at a point Q that is x away from is formed. The displacement of the test surface (1a) at this time is represented by the points P and Q
Is measured by measuring a projected component in a direction perpendicular to the optical axis of the condensing optical system at a distance x from the optical axis. At this time, since the numerical aperture (NA) of the condensing optical system is larger than the numerical aperture of the projection optical system, the image expanded by this defocusing spreads point-symmetrically around the conjugate point of point Q, and An image is formed on two surfaces. When the test surface (1a) is a wafer covered with a resist, the test surface can be regarded as almost a mirror surface, but since it is not a perfect mirror surface, some scattering occurs, and each point on the test surface is affected. May be regarded as a secondary light source in some cases. Even in such a case, the image-forming light beam of the reflected light can be regarded as a light beam that is distributed almost axially symmetrically around the principal light beam LB1 'parallel to the original principal light beam LB1, and the image formed by the condensing optical system eventually becomes a point. Distributed point-symmetrically around the conjugate point of Q, the measured shift amount (projected component of the shift amount x) does not change.

On the other hand, when the projection optical system (9, 10) is not telecentric as shown in FIG. 10 (b), the off-axis principal ray LA2 is not parallel to the optical axis, and the surface (1a) to be measured is indicated by a dotted line. When the light is displaced to the position (1a ') indicated by, the reflected light from the test surface (1a) is reflected by the light-converging optical system on the extension of the principal ray LB2' passing through the point P '(a mirror image of the point P). An image is formed so as to spread at a position centered on the conjugate point of the point Q ′ on the surface to be inspected (1a). Since the point Q ′ is shifted from the point Q in FIG. 10A by an interval Δx, a projected component of the interval Δx in a direction perpendicular to the optical axis of the condensing optical system becomes a detection error. In FIG. 10B, even if the condensing optical system is telecentric, as described above, the surface to be measured (1a) acts as a secondary light source instead of a perfect mirror surface. The distance x 'from the point Q' becomes small. The shift amount of the telecentricity changes according to the distance from the intersection between the surface to be measured (1a) and the optical axis of the projection optical system, and the shift amount becomes an error. That is, if the projection optical system is not telecentric on the test surface (1a), the measurement value changes depending on the position of the measurement point on the test surface (1a), and a measurement error occurs. On the other hand, in the present invention, the projection optical system is configured to be telecentric on the surface to be inspected (1a), and the focusing optical system re-images with a numerical aperture larger than that of the projection optical system. No error occurs.

For example, as shown in FIG. 9, with respect to an optical system that forms a pattern on the surface A on the surface B,
The condition that the surface and the surface B satisfy the Scheimpflug condition means that, in the meridional section of this optical system, the intersection point between the extension line of the surface A and the object-side principal plane of the optical system is H, and the extension of the surface B is If the intersection of the line and the image-side principal plane of the optical system is H ', it means that the distance L from the intersection H to the optical axis is equal to the distance L' from the intersection H 'to the optical axis. When the Scheimpflug condition is satisfied, a so-called tilt image formation relationship is established, and light emitted from any point on the A surface converges to a corresponding point on the B surface.
Therefore, an image of the entire point on the surface A is formed on the surface B.

In the surface position detecting device of the present invention, the first surface and the surface to be inspected (1a) satisfy the image forming relationship of tilt, and the surface to be inspected (1a) and the second surface thereof form a tilt image. Since the image relationship is satisfied, an image of each point on the entire surface of the test surface (1a) is formed on the second surface. In this case, since the projection optical system projects an image of a predetermined pattern from the oblique direction to the surface to be inspected (1a), if the surface to be inspected (1a) partially has irregularities, the second surface The upper image produces a displacement (distortion) corresponding to the unevenness. If this displacement is detected by, for example, the principle of a photoelectric microscope disclosed in Japanese Patent Application Laid-Open No. Sho 56-42205, it is possible to obtain vertical position information of the test surface (1a). The principle of the photoelectric microscope will be briefly described. First, the light receiving surface of the detector (17) and the image on the second surface are relatively scanned by the scanning means (30, 31), and the modulated light is photoelectrically detected on the light receiving surface. Then, when the photoelectric signal detected based on the scanning cycle is synchronously detected, a detection output corresponding to the amount of displacement of the surface to be detected (1a) can be obtained.

Furthermore, the test surface (1a) and the second surface are
Compared to the magnification relationship between surfaces because of the
Then, on the second surface due to the vertical movement of the test surface (1a),
The lateral shift amount of the image of the image becomes large. This is quantitatively explained
I will tell. For example, as shown in FIG.
The incident angle of the principal ray from the projection optical system (9, 10)
θ, the amount of vertical displacement of the test surface (1a) is z, the test surface
(1a) imaging on the second surface along the tilt image plane
Assuming that the magnification is β ', the lateral displacement amount y on the second surface ₁ Is
It can be expressed by the following equation.

Y ₁ = 2 · β ′ · tan θ · z (1) By the way, if the lateral magnification which is the magnification of the surface to be measured (1a) and its second surface in the direction perpendicular to the optical axis is β, The following relationship is established from the Scheimpflug condition. β ′ = (β ² cos ² θ + β ⁴ sin ² θ) ^1/2 (2) On the other hand, since the lateral displacement of the image is detected in a direction perpendicular to the optical axis, the detected lateral displacement y ₂ is Become like

Y ₂ = 2 · β · sin θ · z (3) Here, when the expressions (1) and (2) are compared, the incident angle θ
Is the greater, the equation (1) is larger in the lateral deviation y ₁ in the case that satisfies the imaging relationship of the tilt. For simplicity, β
If = 1, β ′ = 1 from equation (2), and the lateral displacement amounts y ₁ and y ₂ are proportional to tan θ and sin θ, respectively. For example, when θ = 80 °, according to equation (1), y ₁
= 11.3z, and according to equation (3), y ₂ = 2.0z
Becomes Therefore, the lateral displacement is larger by 5.7 times when the tilt relationship is satisfied, and the surface to be inspected (1a)
The detection sensitivity and the detection accuracy for the vertical displacement of are improved.

Next, in the present invention, it is desirable to provide a tilt angle correcting optical system (25) on the second surface for reducing the inclination angle of the light beam emitted from the condensing optical system with respect to the optical axis. That is, in the present invention, when the incident angle θ of the principal ray from the projection optical system to the test surface (1a) is increased,
Since the rate of change in the amount of lateral displacement with respect to the vertical displacement of the test surface (1a) increases, the detection accuracy can be further improved. However, when the incident angle θ is large,
The incident angle of the light beam incident on the second surface also increases. At this time, for example, if the light receiving surface of the detector (17) is arranged on the second surface, surface reflection on the light receiving surface, blurring of the light beam, etc. will occur, and the amount of light received on the light receiving surface will decrease significantly. May be. Therefore, the tilt angle correcting optical system (25) is arranged on the second surface, and the optical axis of the condensing optical system is deflected to reduce the incident angle of the light beam reaching the light receiving surface. There is no danger of reducing the volume.

Preferably, the tilt angle correcting optical system is a prism. Since the prism body has a smaller dispersion than a grating, for example, a wide-band light beam can be used as detection light. By using a broadband light beam, the influence of thin-film interference and the like can be reduced when the surface to be inspected (1a) is coated with a photoresist.

[0019]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of a surface position detecting device according to the present invention will be described below with reference to FIGS. In the present embodiment, the present invention is applied to an oblique incidence type multi-point autofocus sensor (AF sensor) provided in a projection exposure apparatus such as a stepper.

FIG. 1 shows a projection exposure apparatus of the present embodiment. In FIG. 1, illumination light for exposure from an illumination optical system 40 for exposure is a circuit formed on a pattern surface (lower surface) of a reticle R. Light the pattern. Under the illumination light, an image of the pattern on the reticle R is projected via the projection optical system 5 onto each shot area on the test surface 1a, which is the surface of the wafer 1. A photoresist is applied to the surface 1a to be inspected, the wafer 1 is held by suction on a wafer holder 2, and the wafer holder 2 is movable in the direction of the optical axis AX1 of the projection optical system 5.
It is mounted on the wafer stage 3 via the fulcrums.

The wafer holder 2 (wafer 1) can be moved in the direction (focusing direction) along the optical axis AX1 by projecting and retracting these three fulcrums in parallel. The wafer holder 2 is two-dimensionally inclined by protruding and retracting (performing leveling).
In addition, the wafer holder 2 can be rotated slightly. In addition, the wafer stage 3
When the wafer holder 2 (wafer 1) is positioned in a plane perpendicular to the optical axis AX1, and exposure to a certain shot area on the wafer 1 is completed, the wafer stage 3 is driven to drive the wafer 1 in a step-and-repeat manner. The next upper shot area is set as the exposure area of the projection optical system 5.
The operation of the wafer stage 3 and the three fulcrums is controlled by the drive unit 4.

Here, in order to satisfactorily transfer the image of the circuit pattern on the reticle R onto the test surface 1a of the wafer 1,
For each shot area to be exposed, it is necessary to fit the current exposure area of the test surface 1a within the range of the depth of focus of the projection optical system 5 with respect to the imaging plane. For this purpose, the optical axis A of the projection optical system 5 at each point on the current exposure area of the test surface 1a is
After accurately detecting the position in X1 direction (focus position), the wafer holder 2 (wafer 1) is set so that the surface to be inspected 1a falls within the range of the depth of focus with respect to the imaging plane of the projection optical system 5.
Leveling and movement in the focusing direction may be performed.

For this purpose, the projection exposure apparatus of this embodiment is provided with a multipoint AF sensor for detecting the focus position and the tilt angle of the current exposure area on the surface 1a to be inspected.
In the AF sensor of FIG. 1, illumination light from a light source 6 that supplies white light having a wide wavelength width is converted into a substantially parallel light beam by a condenser lens 7 and then enters a deflection prism 26. The deflecting prism 26 deflects the substantially parallel light beam from the condenser lens 7 by refracting it. On the exit side of the deflecting prism 26, there is provided a transmission type lattice pattern plate 8 on which a plurality of stripe patterns extending in a direction perpendicular to the plane of FIG. 1 are formed. That is, the lattice pattern forming surface 8a on the wafer side of the transmission type lattice pattern plate 8
Is formed with a transmission type lattice pattern in which transmission portions and light shielding portions are alternately arranged at a predetermined pitch.

Instead of the transmission type grating pattern, a reflection type diffraction grating having an uneven shape may be used. Further, a reflection grating pattern in which reflection portions and non-reflection portions are alternately formed is used. Is also good. In this example, since the surface of the wafer 1 (the surface 1a to be inspected) is generally covered with a thin film such as a photoresist, in order to reduce the influence of interference caused by the thin film, the light source 6 has a wavelength width. It is desirable to use a white light source that generates a wide light. In addition, as the light source 6, a light emitting diode or the like that supplies light in a wavelength band having a low photosensitivity to the photoresist may be used. And the deflection prism 2
The illumination light that has reached the grating pattern forming surface 8a after passing through the grating pattern forming surface 8a passes through the projection optical system 5a.
Condenser lens 9 and projection objective lens 1 disposed along an optical axis AX2 that intersects the optical axis AX1 at an angle θ.
The light enters the projection optical systems 9 and 10 consisting of 0. The projection optical systems 9 and 10 arrange the lattice pattern forming surface 8a and the test surface 1a of the wafer 1 in a conjugate arrangement. Then, the test surface 1a
Is coincident with the image forming plane of the projection optical system 5, and the grating pattern forming surface 8a and the test surface 1a are arranged so as to satisfy the Scheimpflug condition with respect to the projection optical systems 9 and 10. . Therefore, the lattice pattern forming surface 8a
Is accurately formed over the entire surface 1a to be inspected.

Further, a telecentric stop 100 as an aperture stop is arranged on the object side focal plane of the projection optical system 10 of the projection optical systems 9 and 10 of this embodiment. In this example, the object-type focal plane of the projection optical system 10 is also the image-side focal plane of the condenser lens 9. Therefore, as shown by a dotted line in FIG. 2, the projection optical systems 9 and 10 composed of the condenser lens 9 and the projection objective lens 10 are so-called double-sided telecentric optical systems, and have a grid pattern forming surface 8a and The conjugate point on the test surface 1a has the same magnification over the entire surface. In this case, since the grid patterns at equal intervals having the longitudinal direction perpendicular to the paper surface of FIG. 1 are formed on the grid pattern forming surface 8a, the image formed on the test surface 1a is:
The grid pattern image is an evenly spaced grid pattern whose longitudinal direction is perpendicular to the paper surface of FIG.

The projection optical systems 9 and 10 need only be telecentric at least on the surface 1a to be detected, and need not be telecentric on the grating pattern forming surface 8a. However, by making the projection optical systems 9 and 10 telecentric also on the grating pattern forming surface 8a side, the grating patterns at positions deviated from the optical axis on the grating pattern forming surface 8a and the corresponding gratings on the test surface 1a Since there is no change in magnification with the pattern image, it is convenient for adjustment. That is, as described later, in this example, the change amount of the focus position is obtained from the lateral shift amount of the grid pattern image at each of a plurality of detection points on the test surface 1a.
By making 10 the telecentric also on the side of the lattice pattern forming surface 8a, the relationship between the lateral shift amount and the focus position becomes the same at all the detection points.

Returning to FIG. 1, the illumination light projected on the test surface 1a is reflected by the test surface 1a, and is condensed optical system 11 comprising a light receiving objective lens 11 and a condensing lens 12. ,
12, the incident surface 25a of the tilt correction prism 25
To form a grid pattern image again. This condensing optical system 11,
Optical axis AX3 ₁ on the incident side of 12, the optical axis A of the projection optical system 5
X1 is provided to be line-symmetric with the optical axis AX2 of the projection optical systems 9 and 10. Further, in the optical path between the incident surface 25a and the test surface 1a, the vibrating mirror 30 is provided as a scanning unit, the illumination light entering the light-receiving objective lens 11 along the optical axis AX3 ₁ is , Vibrating mirror 3
Through 0, it reaches the condenser lens 12 along the optical axis AX3 _2. In this example, the vibrating mirror 30 includes the focusing optical systems 11 and 1.
2 is disposed on a substantially pupil plane (an optical Fourier transform plane with respect to the test surface 1a), and the vibrating mirror 30 includes the test surface 1a and the incident surface 25a (further, a light receiving surface of a photoelectric detector 17 described later). 17a) may be placed anywhere in the optical path.

When the test surface 1a coincides with the image forming surface of the projection optical system 5, the tilt correction prism 25 is placed on a surface conjugate with the test surface 1a with respect to the condensing optical systems 11 and 12.
Are arranged so that the light incident surface 25a is located. At this time, the reflected light from the test surface 1a is condensed on the incident surface 25a of the tilt correction prism 25 to form a lattice pattern image. A light receiving slit S as a light blocking means is provided on the incident surface 25a.

FIG. 3 shows the light receiving slit S. As shown in FIG. 3, the light receiving slit S has five openings S, for example.
a1, Sa2, Sa3, Sa4, Sa5 (hereinafter, "Sa
1 to Sa5), and the reflected light from the test surface 1a via the condensing optical systems 11 and 12 passes through the openings Sa1 to Sa5 of the light receiving slit S, respectively. Then, the light enters the tilt correction prism 25 shown in FIG. The number of the openings Sa1 to Sa5 of the light receiving slit S corresponds to the detection point of the focus position on the test surface 1a.

FIG. 4 shows a state where an image 80a of the lattice pattern forming surface 8a is projected on the surface 1a to be inspected.
, A detection point (detection area) Da1 on the surface 1a to be detected
Da2, Da3, Da4, Da5 (hereinafter, "Da1 to D
a5 ") correspond to the five openings Sa1 to Sa5 of the light receiving slit S in FIG. Then, the focus positions at the plurality of detection points Da1 to Da5 are detected. Therefore, when it is desired to increase the number of detection points on the test surface 1a, the openings Sa1 to Sa5 of the light receiving slit S are required.
Since only the number of detection points needs to be increased, an increase in the number of detection points does not complicate the configuration.

Further, in FIG. 1, when the surface 1a to be inspected on the wafer 1 coincides with the image forming surface of the projection optical system 5, the surface 1a to be inspected and the incident surface 25a of the tilt correction prism 25 are Condensing optical systems 11 and 12 are arranged to satisfy the Scheimpflug condition. Therefore, when the test surface 1a and the imaging surface match, the image 80a of the lattice pattern on the test surface 1a is accurately re-imaged over the entire incident surface 25a.

In this embodiment, no stop is provided in the converging optical systems 11 and 12 for making the converging optical systems telecentric. However, the numerical aperture of the converging optical systems 11 and 12 is different from that of the projection optical system. It is configured to be larger than the numerical aperture of the systems 9, 10. As an example, assuming that the numerical apertures of the projection optical systems 9 and 10 are about 0.04, the numerical apertures of the condensing optical systems 11 and 12 are set to about 0.05. Then, the condensing optical systems 11 and 12 for the principal rays of the illumination light from the projection optical systems 9 and 10
The optical path in the inside is as shown by the dotted line in FIG.

That is, in the condenser optical systems 11 and 12 shown in FIG. 2, the vibrating mirror 30 is arranged near the intersection of the principal rays of the illumination light from the surface 1a to be inspected, and the aperture stop is located at the intersection. There is no. However, the condensing optical systems 11 and 12
Are larger than the numerical apertures of the projection optical systems 9 and 10, and the test surface 1a of the wafer 1 can be regarded as almost a mirror surface, so that the condensing optical systems 11 and 12 are substantially telecentric on the test surface 1a side. It can be considered that the surface 1a to be inspected has the optical axis A
Even when displaced in the direction of X1, the focus position can be accurately detected.

Further, in the condensing optical systems 11 and 12 of this embodiment, no telecentric diaphragm is provided, but the image-side focal point of the light-receiving objective lens 11 and the object-side focal point of the condenser lens 12 match. Have been. Therefore, the condensing optical system 1
Since points 1 and 12 can be substantially regarded as telecentric on both sides, each point on the test surface 1a and a conjugate point on the incident surface 25a of the tilt correction prism 25 have substantially the same magnification over the entire surface.
Therefore, when the surface 1a to be inspected coincides with the imaging plane of the projection optical system 5, the image projected on the light receiving slit S is also shown in FIG.
Is a grid-like pattern at equal intervals with the longitudinal direction perpendicular to the plane of the drawing.

That is, in this embodiment, when the surface 1a to be inspected and the image forming surface of the projection optical system 5 coincide with each other, the lattice pattern forming surface 8a, the surface 1a to be inspected, and the tilt correction prism 2
The five incident surfaces 25a satisfy the Scheimpflug condition, and each surface has substantially the same magnification over the entire surface. Next, the lateral displacement y of the grating pattern image on the entrance surface 25a of the tilt correction prism 25 when the displacement of the test surface 1a in the direction of the optical axis AX1 is z is obtained. Here, the incident angle of the chief ray of the incident light on the surface 1a to be inspected is θ, the lateral magnification of the condensing optical systems 11 and 12 is β, and the surface 1
Assuming that the magnification along the image plane of the tilt from the lens a to the incident surface 25a of the tilt correction prism 25 is β ′, the above-mentioned (1)
From equation (2), the lateral shift amount y is as follows.

Y = 2 · β ′ · tan θ · z = 2 (β ² sin ² θ + β ⁴ sin ⁴ θ / cos ² θ) ^1/2 · z (4) That is, a lattice pattern projected on the surface 1a to be measured If the incident angle θ of the imaging light beam is increased, the lateral shift amount y also increases, and it becomes possible to detect the surface position with higher accuracy. Incidentally, the projection optical systems 9 and 10 and the condensing optical systems 11 and 12 each satisfy the Scheimpflug condition. Therefore, the angle between the normal of the transmission grating pattern plate 8 and the reflected light is γ,
The angle between the incident light and the normal to the tilt correction prism 25 is α (this angle α is the tilt correction prism 25
), The lateral magnification of the projection optical systems 9 and 10 is β ₄ , and the lateral magnification of the condensing optical systems 11 and 12 is β, the following relationship holds. .

Tan γ = β ₄ · tan θ (5) tan α = β · tan θ (6) By the way, when a light receiving surface of a photoelectric detector described later is directly disposed on the incident surface 25a, the surface to be detected is Incident angle θ for 1a
Is large, the angle of incidence of the light beam on the light receiving surface also increases. For example, when the light receiving surface of the silicon photodiode SPD is arranged on the incident surface 25a and the incident angle of the light beam on the silicon photodiode SPD is large,
The surface reflection at the silicon photodiode SPD increases, and the luminous flux is disturbed, so that the amount of received light may be significantly reduced.

In this embodiment, in order to avoid such a decrease in the amount of received light, as shown in FIG. 1, a tilt correction prism 25 as a deflecting optical system is arranged on the incident surface 25a, and the condensing optical system 11, The light emitted from the light source 12 is deflected. The tilt correction prism 25 has a predetermined apex angle ξ, as shown in the cross-sectional view of FIG.
Are determined to be substantially parallel to the normal to the exit surface 25b of the tilt correction prism 25.

In FIG. 1, on the exit side of the tilt correction prism 25, relay optical systems 14 to 16 including a plane mirror 15, a relay lens 14, and a relay lens 16 are arranged. The light beam incident on the tilt correction prism 25 is
After being subjected to the refraction action by the tilt correction prism 25, it is guided to the plane mirror 15. The light beam reflected by the plane mirror 15 is converged on the light receiving surface 17a of the photoelectric detector 17 by the relay optical systems 14 to 16, and a lattice pattern image is formed on the light receiving surface 17a. That is, the relay optical system 14
16 is a light receiving surface 17 for further conjugate images of the lattice pattern image formed on the incident surface 25a of the tilt correction prism 25.
a.

Here, the operation of the tilt correction prism 25 will be described with reference to FIG. 5, when a light beam parallel to the optical axis AX3 ₂ on the exit side of the converging optical system 11, 12 in FIG. 1 and L1, the incident angle of the light beam L1 with respect to the tilt correction prism 25 from FIG. 1 becomes alpha. In FIG. 5, the refractive angle of the tilt correction prism 25 is n, and the refraction angle is ξ.
Then, the following equation is established.

Ξ = sin ⁻¹ (sin α / n) (7) In this example, the apex angle of the tilt correction prism 25 is
The angle of refraction is set to ξ. This allows
The light beam is emitted from the exit surface 25b of the tilt correction prism 25 L2 becomes perpendicular to the exit surface 25b, the incident side of the optical axis AX4 ₁ also its exit surface 2 of the relay optical system 14 to 16
5b.

At this time, the angle between the light beam L2 entering the tilt correction prism 25 and the light beam L2 exiting is
If the apex angle 小 さ い is small using the refractive index n of the tilt correction prism 25, then (sin ⁻¹ (nsin ξ) −ξ). This angle is considerably smaller than the angle α formed between the light beam L1 incident on the tilt correction prism 25 via the condenser optical systems 11 and 12 and the incident surface 25a.

Considering the floating of the image by the glass of the tilt correction prism 25, the tilt angle ρ, which is the inclination of the relayed image plane of the test surface 1a with respect to the plane perpendicular to the emitted light beam L2, is It is shown by the formula. ρ = tan ⁻¹ (tan ξ / n) (8) Now, assuming that the magnification of the relay optical systems 14 to 16 is 1, the tilt angle ρ is directly used as the light receiving surface 17 of the photoelectric detector 17.
and a, the optical axis AX4 of the exit side of the relay optical system 14 to 16 ₂
, That is, the angle of incidence of the principal ray on the light receiving surface 17a.

For example, when the refractive index n of the tilt correction prism 25 is 1.8 and the incident angle α is 80.1 °, the tilt angle ρ is 20.0 °, and the tilt angle approaches 0 °. Here, the angle of incidence of the light beam on the light receiving surface 17a is also 20.0.
However, with this degree, it can be considered that there is almost no decrease in the amount of received light. As described above, in this example, the tilt angle correction prism 25 is used as the deflection optical system, and the incident angle of the light beam incident on the light receiving surface 17a is reduced. A decrease in the amount of received light can be prevented.

The incident surface 2 of the tilt correction prism 25
It is desirable that 5a and the light receiving surface 17a of the photoelectric detector 17 satisfy the Scheimpflug condition with respect to the relay optical systems 14 to 16. Further, since the numerical apertures of the relay optical systems 14 to 16 are also set to be larger than the numerical apertures of the projection optical systems 9 and 10, the condensing optical systems 11 and 16 extending from the test surface 1a to the light receiving surface 17a. 12 and relay optical system 14
The numerical aperture of the optical system consisting of .about.16 is set to be larger than the numerical aperture of the projection optical systems 9, 10. As a result, even if the position of the test surface 1a changes, the position can be accurately detected without providing an aperture stop such as a telecentric stop in the relay optical systems 14 to 16. Therefore, the degree of freedom in designing the optical system is increased, and the adjustment of the optical system is facilitated. In this example, the grating pattern forming surface 8a and the light receiving surface 2
Since only one telecentric aperture 100 is provided between the aperture stop 5a and the aperture 5a, there is an advantage that the loss of light amount is small.

When the wavelength band of the illumination light from the light source 6 is wide as in this embodiment, it is desirable to use a prism such as the tilt correction prism 25 as the deflection optical system. On the other hand, when a diffraction grating is used as the deflecting optical system, for example, the divergence of the deflected light beam becomes larger than when a prism is used, and the relay optical systems 14 to 16 are used.
It is not preferable because it is necessary to increase the numerical aperture.

When the incident angle α of the light beam incident on the test surface 1a is large, the transmittance of the light beam passing through the tilt correction prism 25 is significantly different between the P-polarized light and the S-polarized light. For example, the incident angle α = 80 ° and the refractive index n = 1.8
In the case of (1), the transmittance of P-polarized light is 0.79, and the transmittance of S-polarized light is 0.37. As described above, since the transmittance of the tilt correction prism 25 varies depending on the polarization state of the light beam, the weight of information for each polarization component varies, and
It may not be possible to accurately detect the position a. Therefore, in this example, as shown in FIG. 1, a half-wavelength plate 24 is provided in the optical path between the condensing optical systems 11 and 12 and the tilt correction prism 25.
Place. Then, the luminous flux whose polarization direction is rotated by 45 ° by the half-wave plate 24 is converted into a tilt correction prism 2.
5 is incident. Therefore, the tilt correction prism 2
5 enters a state where P-polarized light and S-polarized light are mixed, and the position of the test surface 1a is correctly detected.

It should be noted that instead of the half-wave plate 24, 1 /
A four-wavelength plate may be used. When a quarter-wavelength plate is used, the light beam incident on the tilt correction prism 25 is circularly polarized, and the position of the surface 1a to be detected is detected by using a half-wavelength plate. It is performed exactly as if it were used. Further, it is preferable that the exit surface 25b of the tilt correction prism 25 and the light beam L2 (see FIG. 5) exiting from the exit surface 25b be substantially perpendicular. In this example, the exit surface 25b and the light beam L2 are made substantially perpendicular to each other by making the apex angle of the tilt correction prism 25 equal to the refraction angle の of the light beam L2 with respect to the light beam L1 incident on the tilt correction prism 25. On the other hand, if the exit surface 25b and the light beam L2 do not become substantially perpendicular, aberrations such as astigmatism are generated when the relay optical systems 14 to 16 relay an image on the incident surface 25a. Not preferred.

On the incident surface 25a of the tilt correction prism 25 shown in FIG. 1, five openings S are formed as shown in FIG.
Since the light receiving slit S having a1 to Sa5 is provided, the lattice pattern image re-imaged on the incident surface 25a is:
Partially shielded from light. That is, the opening S of the light receiving slit S
Only the luminous flux from the lattice pattern image formed in the areas a1 to Sa5 passes through the tilt correction prism 25. Then, the images of the openings Sa1 to Sa5 of the light receiving slit S formed on the incident surface 25a are relayed to the light receiving surface 17a of the photoelectric detector 17 by the relay optical systems 14 to 16.

FIG. 6 shows the light receiving surface 17 of the photoelectric detector 17.
6, five silicon photodiodes SPD1, SPD are located at positions corresponding to the openings Sa1 to Sa5 of the light receiving slit S in FIG. 3 on the light receiving surface 17a in FIG.
2, SPD3, SPD4, SPD5 (hereinafter, "SPD1
-SPD5 "). On these silicon photodiodes SPD1 to SPD5, images of openings Sa1 to Sa5 are slit images SL1 and SL.
2, SL3, SL4, SL5 (hereinafter, "SL1 to SL
5 "). This slit image SL
1 to SL5 are silicon photodiodes SPD1
-Image is formed so as to be smaller than SPD5. In this example, a photoelectric detector 17 provided with a plurality of silicon photodiodes on a light receiving surface is used. However, as the photoelectric detector 17, a two-dimensional CCD image sensor, an optical fiber and a plurality of photodetectors are used. A detector combined with a multiplier may be used.

When the test surface 1a moves up and down along the optical axis AX1 of the projection optical system 5, the lattice pattern image projected on the incident surface 25a of the tilt correction prism 25 becomes
A lateral shift occurs in the pitch direction of the lattice pattern image by an amount corresponding to the vertical amount of the test surface 1a. In this embodiment, the amount of lateral displacement is detected by using the vibration mirror 30 and the photoelectric detector 17 according to the principle of the photoelectric microscope disclosed in, for example, Japanese Patent Application Laid-Open No. Sho 56-42205, to thereby determine the amount of each deviation on the surface 1a to be inspected. The position (focus position) in the direction of the optical axis AX1 at the detection position is detected.

Returning to FIG. 1, the detection of the focus position based on the principle of the photoelectric microscope will be described in detail. In FIG. 1, a vibrating mirror 30 is provided in the optical path of the condenser optical systems 11 and 12. Then, the mirror driving unit 31
The vibrating mirror 30 is vibrated in a direction indicated by an arrow in FIG. 1 at a predetermined cycle T based on a signal from an internal oscillator. With the vibration of the vibration mirror 30, the tilt correction prism 25
The lattice pattern image formed on the incident surface 25a also vibrates. At this time, the amplitude of the vibration of the lattice pattern image formed on the incident surface 25a is one of the pitch of the lattice pattern image.
/ 2 or less, and the openings Sa1 to Sa1 of the slit S
The width of Sa5 is also defined to be 以下 or less of the pitch of the lattice pattern image. With the vibration of the lattice pattern image, the incident surface 2
5a, that is, the amount of detection light passing through the region of the openings Sa1 to Sa5 of the light receiving slit S changes. The transmitted light is transmitted to the silicon photodiodes SPD1 to SPD1 on the photoelectric detector 17 by the relay optical systems 14 to 16.
The light reaches the light receiving surface of the PD 5.

Hereinafter, for simplicity of explanation, FIG.
Only one silicon photodiode SPD1 will be described. The light transmitted through the area of the opening Sa1 of the light receiving slit S forms a slit image SL1 on the silicon photodiode SPD1. This slit image SL1
Varies with the vibration of the vibrating mirror 30.
Next, a change in brightness of the slit image SL1 will be described in detail with reference to FIG.

FIG. 7 is an enlarged view of the vicinity of the opening Sa1 of the light receiving slit S. In FIG. 7, the grating pattern image ST is formed on the light receiving slit S at a pitch PST. Vibrates in the direction of the arrow in FIG. Here, the opening S
It is desirable that the width WSa1 of a1 satisfies the following relationship with the pitch PST of the lattice pattern image ST formed on the opening Sa1.

WSa1 ≦ PST / 2 (9) It is desirable that the amplitude AST of the grating pattern image ST vibrated by the vibrating mirror 30 satisfies the following relationship. AST ≦ PST / 2 (10) Here, when the width WSa1 of the opening Sa1 does not satisfy the above equation (9), or when the amplitude AST of the lattice pattern image ST does not satisfy the above equation (10). Is not preferable because the change in the amount of light at the opening Sa1 due to the vibration of the vibrating mirror 30 is small, and the detection accuracy is reduced. In this example, the width WSa1 of the opening is the lattice pattern image S
T is set to be equal to or less than の of the pitch PST, and the amplitude AST of the grid pattern image
Since it is defined to be equal to or less than の of ST, there is no possibility that a change in the amount of light due to the vibration of the vibrating mirror 30 becomes small.

The position of the opening Sa1 is determined by the position of the test surface 1a.
The center of the opening Sa1 is provided so as to coincide with the vibration center of the lattice pattern image ST when the image coincides with the image forming plane of the projection optical system 5. Then, the width WSa1 of the opening Sa1 is １ ／ of the pitch PST of the lattice pattern image ST.
And the amplitude AST of the grating pattern image ST is also equal to or less than の of the pitch PST. Therefore, when the grating pattern image ST vibrates, the amount of light received on the silicon photodiode SPD1 changes. Then, the silicon photodiode SPD1 changes the light intensity,
That is, the detection output SD1 corresponding to the light modulation is output to the detection unit 18 in FIG.
Output to

Similarly, in the other silicon photodiodes SPD2, SPD3, SPD4, SPD5, the slit images SL2, SL3, SL4, respectively
Detection output SD2, SD3, SD according to light modulation of SL5
4 and SD5 are output to the detector 18. Further, an AC signal having the same phase as the oscillation period T of the oscillation mirror 30 is output from the mirror driving unit 31 to the detection unit 18. Here, the detection unit 1
8, with reference to the phase of the AC signal having the period T,
Synchronous rectification of the detection outputs SD1, SD2, SD3, SD4, and SD5 (hereinafter abbreviated as "SD1 to SD5"), that is, synchronous detection, is performed, and the detection output signals FS1, FS2, and FS are output.
3, FS4, FS5 (hereinafter abbreviated as “FS1 to FS5”) are output to the correction amount calculation unit 19. The detection output signals FS1 to FS5 are called so-called S-curve signals, and are used when the detection points Da1 to Da5 on the surface 1a to be detected are located on the imaging surface of the projection optical system 5, that is, the detection outputs SD1 to SD5.
Are changed to a zero level when they change in a cycle of １ ／ of the oscillation cycle T of the oscillation mirror 30. In addition, the test surface 1a
Are positive levels when the detection points Da1 to Da5 are displaced above the imaging plane of the projection optical system 5, and the detection points Da1 to Da5
When Da5 is displaced below the image plane of the projection optical system 5, it indicates a negative level. As described above, the detection output signals FS1 to FS5 indicate output values corresponding to the displacement (focus position) of the test surface 1a.

After that, the correction amount calculating section 19 to which the detection output signals FS1 to FS5 have been output, outputs the detection output signals FS1 to FS5.
From the positive and negative levels of S1 to FS5, the focus positions of the respective detection points Da1 to Da5 on the test surface 1a are calculated, respectively.
An average inclination of the test surface 1a and an average focus position of the test surface 1a are obtained. Then, the amount of correction of the inclination and the amount of correction of the focus position such that the test surface 1a is located within the range of the depth of focus with respect to the imaging plane of the projection optical system 5 are calculated. After that, the correction amount calculation unit 19 transmits these correction amounts to the drive unit 4. Then, the drive unit 4 drives the three fulcrums on the wafer stage 3 based on the correction amount of the tilt, performs the leveling of the wafer holder 2, and based on the correction amount of the focus position, the three fulcrums. The fulcrum is driven to focus the wafer holder 2 (wafer 1). As described above, in this example, the focus positions of a plurality of detection points on the test surface 1a are simultaneously detected by the multipoint AF sensor, and the test surface 1a is adjusted to the image forming surface based on the detection output. it can.

Although one slit image is formed on each of the silicon photodiodes SPD1 to SPD5 in FIG. 6, a plurality of slit images are formed on one silicon photodiode SPD6 as shown in FIG. Slit images (eg, three slit images SL6a, SL6b,
SL6c) may be formed. At this time, the slit image S
It is desirable that the pitch P6 of L6a, SL6b, SL6c satisfies the following expression with respect to the pitch PST of the lattice pattern image.

P6 = m · PST (11) where m is an integer. Also, each slit image SL6
It is desirable that the width W6 of a, SL6b, and SL6c satisfy the following expression. W6 ≦ PST / 2 (12) Here, if the interval P6 between the slit images SL6a, SL6b, and SL6c deviates from the above equation (11), the vibration of the vibrating mirror 30 is caused for each of the slit images SL6a, SL6b, and SL6c. This is not preferable because the change in the amount of light becomes different and the detection value becomes inaccurate. Also, the slit image S
The width W6 of L6a, SL6b, and SL6c is as described in (12) above.
If the value is out of the range of the expression, the change in the amount of light accompanying the vibration of the vibrating mirror 30 becomes small, which is not preferable.

According to the above-described configuration, in addition to the advantage that the amount of light received on the silicon photodiode SPD6 increases, the corresponding detection area on the test surface 1a is widened and an averaging effect is obtained. There are advantages. In this example, the silicon substrate provided on the light receiving surface 17a
Although the light through the light receiving slit S is incident on the photodiodes SPD1 to SPD5, it is not always necessary to pass through the light receiving slit S. When not passing through the light receiving slit S, the silicon photodiodes SPD1 to SPD1
The area of the PD 5 is a light receiving area on the surface 1a to be inspected. In other words, silicon is placed at a position corresponding to the detection point on the surface 1a to be detected.
Photodiodes SPD1 to SPD5 may be arranged.

In this embodiment, the light receiving slit S
Is the second surface (incident surface 25a of tilt correction prism 25)
However, this may be on the light receiving surface 17 a of the photoelectric detector 17. At this time, it is desirable that the size of the opening portion Sa of the light receiving slit S be smaller than the size of the light receiving surface 17a. In this example, a plurality of silicon photodiodes may be provided on the light receiving surface 17a by a photo etching method.
In this case, the openings Sa1 to Sa5 of the light receiving slit S
There is an advantage that even if the number increases significantly, it can be handled.

The surface of the wafer as the test surface 1a is not always flat depending on the process. In such a case, the number of detection areas (detection points) on the test surface 1a may be increased. Here, in order to increase the number of detection points on the test surface 1a, the number of openings of the light receiving slit S corresponding to the detection points may be increased. As described above, the AF sensor of the present embodiment can increase the number of detection points on the test surface 1a with a simple configuration. Can be. Moreover, even if the number of detection points increases, they can be detected at the same time, so that the throughput does not decrease.

It should be noted that the present invention is not limited to the above-described embodiment, and it is needless to say that various configurations can be adopted without departing from the gist of the present invention.

[0065]

According to the surface position detecting device of the present invention, since the projection optical system is a telecentric optical system and the numerical aperture of the condensing optical system is larger than the numerical aperture of the projection optical system, a simple configuration is possible. Therefore, there is an advantage that the surface position of the test surface can be detected with high accuracy even if the position of the test surface changes.

Further, since the number of aperture stops is small, there is an advantage that the degree of freedom in designing the optical system is increased and the light amount loss is small. When the tilt angle correcting optical system for reducing the inclination angle of the light beam emitted from the condensing optical system with respect to the optical axis is provided on the second surface, the incident angle of the light beam incident on the light receiving surface of the detector Is reduced, the light receiving efficiency of the light beam is increased, and the SN ratio of the obtained detection signal is advantageously improved.

When the tilt angle correcting optical system is a prism body, there is an advantage that a wide band light beam having little influence of thin film interference or the like can be used as a detection light beam because the dispersion is small.

[Brief description of the drawings]

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

FIG. 2 shows a projection optical system 9, 10 shown in FIG.
FIG. 7 is an optical path diagram showing that the side is telecentric.

FIG. 3 is a plan view showing an opening of a light receiving slit S of FIG. 1;

FIG. 4 is a perspective view showing a relationship between a lattice pattern image projected on a test surface 1a in FIG. 1 and a detection area.

FIG. 5 is an enlarged view showing the tilt correction prism 25 of FIG. 1;

6 is a plan view showing a relationship between a silicon photodiode on a light receiving surface 17a of the photoelectric detector 17 of FIG. 1 and a slit image.

FIG. 7 is an enlarged view showing an opening on the light receiving slit S of FIG. 3;

FIG. 8 is an enlarged view showing an example in which a plurality of slit images are formed on one silicon photodiode.

FIG. 9 is a diagram provided for explaining conditions of Scheimpflug.

FIG. 10A is a diagram showing a change in an optical path due to up and down of a test surface when a projection optical system is telecentric, and FIG. 10B is a diagram showing a change in an optical path due to up and down of a test surface when the projection optical system is not telecentric; FIG.

[Explanation of symbols]

 R Reticle 1 Wafer 1a Test surface 5 Projection optical system 8 Transmissive grating pattern plate 8a Grid pattern forming surface (first surface) 9 Condensing lens 10 Projection objective lens 11 Light receiving objective lens 12 Condensing lens 14, 16 Relay Lens 17 Photoelectric detector 17a Light receiving surface 25 Tilt correction prism 25a Incident surface (second surface) 30 Vibrating mirror 31 Vibrating mirror driver S Light receiving slit 100 Telecentric diaphragm

Claims (3)

[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 measured, and a light reflected by the surface to be measured. A condensing optical system that condenses the light beam and forms the image of the predetermined pattern on the second surface; and a detector that photoelectrically detects the image of the predetermined pattern. A surface position detection device that detects a surface position of the surface to be detected based on an output, wherein a scanning unit that relatively scans a light receiving surface of the detector and an image of the predetermined pattern is provided; And the test surface satisfy the condition of Scheimpflug with respect to the main plane of the projection optical system, and the test surface and the second surface satisfy the condition of Scheimpflug with respect to the main plane of the condensing optical system. And the projection optical system is substantially on the test surface side. A telecentric optical system, the light converging optical system, the surface position detecting apparatus characterized by having a numerical aperture larger than the projection optical system. 2. The surface position detecting device according to claim 1, wherein a tilt angle correcting optical system for reducing a tilt angle of a light beam emitted from the condensing optical system with respect to an optical axis is provided on the second surface. A surface position detecting device, comprising: 3. The surface position detecting device according to claim 2, wherein the tilt angle correcting optical system is a prism.