JPH10199788A - Imaging optical system, aligner, and manufacture of device using the optical system - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide an imaging optical system which can apply technique of limit resolution improvement by the prior art diffracting grating method by using only one projection optical system and improve the limit resolution of a periodic pattern arranged in all directions according to the need. SOLUTION: A reticle 11 is illuminated with an exposure light IL which has passed apertures 6A, 6B of an aperture stop plate 5. A wafer 18 is irradiated with a luminous flux which has passed through the reticle 11, through a modulation grating 14, a projection optical system 15, a demodulation grating 16, and a member 17 for filtering out unnecessary light. The reticle 11 and the wafer 18 are conjugate with respect to the projection optical system 15. The demodulation grating 16 is irradiated with an n-th order diffracted light of the modulation grating 14, through the projection optical system. A luminous flux which is largely inclined in the direction rectangular to the direction (X- direction) which is to improve resolution is eliminated out of the n-th order diffracted light from the demodulation grating 16, by using the member 17 for filtering out unnecessary light.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image forming optical system, an exposure apparatus having the image forming optical system, and a method of manufacturing a device using the image forming optical system. Etc.), used in a lithography process for manufacturing a liquid crystal display element, a thin film magnetic head, or the like, in a case where a mask pattern is transferred onto a photosensitive substrate with high resolution via a projection optical system by so-called super-resolution. It is suitable.

[0002]

2. Description of the Related Art In a projection exposure apparatus such as a stepper used for manufacturing a semiconductor device or the like, it is desired to print a pattern with a degree of integration as high as possible on a wafer (or a glass plate or the like). Therefore, in order to increase the resolution of the exposure, the wavelength of the exposure light (exposure wavelength)
Efforts have been made to shorten the numerical aperture and increase the numerical aperture (NA) of the projection optical system.

In order to further improve the resolving power under the same exposure wavelength and the same numerical aperture of the projection optical system, so-called super-resolution techniques such as a modified light source method (tilt illumination method) and a phase shift reticle method are also used. These technologies have been developed and are being used. By the way, Kose Koji: Applied Physics Vol. 37, No. 9
In p.853 (1968), various methods of super-resolution techniques are classified and introduced. One of the techniques is a “pupil synthesis method”, and one specific technique for implementing the pupil synthesis method is a “diffraction grating method”. This diffraction grating method
W. Lukosz: J. Opt. Soc. Amer. 56, p. 1463 (1966). In this diffraction grating method, when there is an object consisting of a periodic pattern with a fine pitch in a specific direction,
At a position separated from the object by the exposure wavelength or more, the diffraction direction is an angle θ (0 <0) with respect to the periodic direction of the periodic pattern.
A diffraction grating for modulation inclined by θ <π / 2) is arranged. Then, a large number of virtual images of the object are generated by the modulation diffraction grating in a direction inclined by an angle θ with respect to the periodic direction. If the light does not pass through the diffraction grating for modulation, a component of the spatial frequency component of the object that is greater than the numerical aperture of the projection optical system cannot pass through the projection optical system. However, by passing through the modulation diffraction grating, the angle θ with respect to the periodic direction is
A part of the spatial frequency components equal to or larger than the numerical aperture of the projection optical system due to the virtual image formed in the direction inclined only can enter within the numerical aperture of the projection optical system.

Therefore, after passing through the projection optical system, the spatial frequency component within the numerical aperture of the projection optical system due to the zero-order diffracted light is obtained by the demodulation diffraction grating which is also tilted by an angle θ with respect to the periodic pattern. A spatial frequency component equal to or larger than the numerical aperture of the projection optical system by the first-order diffracted light is synthesized. The demodulation diffraction grating may also be separated from the image by an exposure wavelength or more. Further, in the diffraction grating method, an image formed immediately after the demodulation diffraction grating is further subjected to a second projection optical system in order to remove an extra spatial frequency component constituting an overlapping image formed by the demodulation diffraction grating. An image is re-formed by the system, and a slit-shaped filter extending in the period direction of the periodic pattern is arranged at the pupil position of the second projection optical system.

According to this diffraction grating method, the modulation diffraction grating and the demodulation diffraction grating can be separated from an object and an image by a pitch equal to or more than the grating pitch, and are used in a projection optical system for manufacturing semiconductor devices. There is a possibility. Also,
According to this method, it is said that the critical resolution can be improved by about two times. However, since those diffraction gratings are separated from the object or the image by the pitch of the grating or more, the improved critical resolution can be improved by 1 in terms of numerical aperture. It is impossible to correspond to the above.
Therefore, it can be said that this technique is not a super-resolution technique, but a technique of forming an image of a light beam having a numerical aperture of the projection optical system or more with a diffraction grating.

[0006]

According to the conventional diffraction grating method as described above, in order to remove the spatial frequency component of the redundant image, a second image for re-imaging the image formed once is used.
Projection optical system is required. However, when two projection optical systems with extremely small aberrations that can withstand the demand for improvement in resolution are manufactured, the manufacturing cost is significantly increased as compared with a conventional projection exposure apparatus that requires only one projection optical system. There are inconveniences.

In the conventional diffraction grating method, the resolution of only the periodic patterns arranged in a specific direction is improved, and the critical resolution of the periodic patterns arranged in the direction perpendicular to the periodic direction of the periodic pattern is improved. There was a disadvantage that it did not improve. In view of the above, the present invention can apply the technology of improving the limit resolution by the conventional diffraction grating method using only one projection optical system, and can limit the periodic pattern arranged in all directions as necessary. It is an object of the present invention to provide an imaging optical system capable of improving the resolving power, and an exposure apparatus including the imaging optical system. Still another object of the present invention is to provide a method of manufacturing various devices using the imaging optical system as an example of a method of using the imaging optical system.

[0008]

An imaging optical system according to the present invention comprises: a projection optical system (15) for projecting an image of a pattern on a first surface onto a second surface under predetermined illumination light; A modulating means (14) disposed between the first surface and the projection optical system (15) for generating an nth-order diffracted light (n is an integer) of a light beam from the pattern on the first surface; An m-th order diffracted light (m
(16 is an integer), and a demodulation means (1)
6) and an optical member (17) arranged between the second surface and the second surface.
Wherein the second side of the projection optical system is substantially telecentric, and the projection optical system has a numerical aperture for passing the n-order diffracted light from the modulating means (14); The member (17) removes unnecessary light beams of orders that do not satisfy the following relationship from the diffracted light from the demodulation means (16).

N + m = 0 (that is, m = −n) (1) The principle of the present invention will be described with reference to FIG. First, the spatial frequency component of the pattern on the first surface is shown in FIG.
(B) The projection optical system (15) distributed in the oval region 21
Is assumed to pass the spatial frequency component in the circular region 24 in FIG. 2 (c).
Components at both ends of 1 cannot contribute to imaging. Therefore, as shown in FIG. 2C, for example, the 0th-order diffracted light, the + 1st-order diffracted light 22, and the -1st-order diffracted light 23 are generated by the modulating means (14), and these diffracted lights are projected. When supplied to the optical system (15), as shown in FIG. 2D, the components 21a, 22a, and 23 in the circular area 24 therein are provided.
Only a passes. Then, using the demodulation means (16), as shown in FIG. 3B, in addition to the 0th-order diffracted light, the -1st-order diffracted light 22a 'of the component 22a and the + 1st-order diffracted light of the component 23a.
When the next-order diffracted light 23a 'is generated, a spatial frequency component including the oval region 21 in FIG. 2B is generated. At this time, n
Are 0, +1, -1 and the corresponding value of m is -n,
That is, 0, -1, and +1. Next, by using the optical member (17), as shown in FIG.
By blocking the 9 light beams, the spatial frequency component in the original oval region 21 is reproduced. That is, the super-resolution effect by the diffraction grating method can be obtained without using the second projection optical system.

In this case, the demodulation means (16) includes, as an example, a diffraction grating disposed in a plane orthogonal to the optical axis of the projection optical system (15), and the diffraction grating in the plane is used to project the projected image. A predetermined direction (x
Direction) with a predetermined pitch d along a direction inclined by an angle θ with respect to the diffraction pattern. At this time, the range of the angle θ is, for example, 0 <θ <π / 2 [ra
d]. Then, assuming that the wavelength of the illumination light is λ, 2
The two angles α1 and α2 are defined as follows.

Α1 = sin ⁻¹ [d · sin θ / (2λ)] (2) α2 = sin ⁻¹ (d · sin θ / λ) (3) At this time, the optical member (17) includes a demodulation unit (16). It is desirable to pass a light beam whose inclination angle with respect to the optical axis of the projection optical system (15) is within the angle α1 or α2 among the light beams from the optical system. At this time, the angle α1 is used for the component 21a of the 0th-order diffracted light in the spatial frequency domain as shown in FIG.
Is orthogonal to the direction (x direction) in which the resolution should be improved.
This is a case where the component 21a has a predetermined spread in the direction, and the component 21a and the ± first-order diffracted light components 22a and 23a are almost in contact with each other. Thus, unnecessary spatial frequency components can be removed while securing an effective component. On the other hand, the angle α2 is used when the component 21a of the 0th-order diffracted light in the spatial frequency domain has little spread in the y direction. In this case, the probability of blocking the effective imaging light flux is reduced. In these cases, the demodulation means (16) and the optical member (17) are synchronized with each other so that π to 2π
[Rad] By making it freely rotatable, the resolution can be improved in all directions.

An example of the optical member includes a plurality of flat plates (31A, 31A, parallel to the direction (x direction) for improving the resolution of the projected image and the optical axis (z direction) of the projection optical system (15). 31B, 31C,...) Are arranged at a predetermined pitch in a direction (y direction) orthogonal to the direction in which the resolution is to be improved. At this time, since the projection optical system (15) is telecentric on the second surface side,
For example, as shown in FIG. 6, assuming that the angle of inclination of the light beam of the spatial frequency component to be removed by the optical member (17) in the direction perpendicular to the optical axis is 2α, the optical member (1
In 7), a light beam having an inclination angle with respect to the optical axis of α or more may be shielded. For this purpose, if the length of the flat plates (31A, 31B,...) In the optical axis direction is h and the interval between them is ds, the following relationship may be satisfied.

Ds ≦ h · tan α (4) Another example of the optical member is a projection optical system (1) for removing the unnecessary light beam from the demodulation means (16).
5) a plurality of prisms (32) having a reflecting surface for reflecting the unnecessary light beam outside the image forming area on the second surface.
To 34). In this case, assuming that the optical member removes a light beam having a tilt angle of α or more in a predetermined direction with respect to the optical axis, for example, as shown in FIG. It is sufficient that the light flux having a surface inclined with respect to the direction (y direction) and having an inclination angle of α or more is directed to another direction by total internal reflection. In other words, it is sufficient that the critical angle is obtained on the surface of at least one prism when the incident angle is α. It is desirable that the surfaces of the prisms that are not perpendicular to the optical axis intersect at a critical angle or more with the light beam emitted at the angle α or more. Even within this range, a coating that significantly increases the reflectance may be applied.

The modulating means (14) and the demodulating means (1)
As 6), not only an amplitude type diffraction grating but also a phase type diffraction grating, an acousto-optic element, etc. can be used. For example, when a phase type diffraction grating is used, the 0th-order diffracted light is included as the diffracted light. Since only the + 1st-order diffracted light and the -1st-order diffracted light can be generated, it is easy to remove the overlapping image. Next, an exposure apparatus according to the present invention includes the above-described imaging optical system of the present invention and illumination optical systems (1 to 4, 7 to 10) for illuminating a mask (11) arranged on the first surface. And a substrate holding member (19) for holding a photosensitive substrate (18) disposed on the second surface thereof, wherein an image of the pattern of the mask is formed on the photosensitive substrate via the imaging optical system. Is transferred to In this case, the mask pattern is transferred onto the photosensitive substrate with a resolution higher than the resolution determined by the numerical aperture of the projection optical system (15) itself.

If the demodulation means (16) is a diffraction grating arranged in a direction rotated by an angle θ with respect to the direction in which the resolution is to be improved, the illumination optical system is arranged in a direction in which the resolution is to be improved. It is desirable to irradiate the mask with a light beam inclined from the optical axis. As a result, the intensity of the light beam emitted from the mask at a large angle is increased, and the image with such improved resolving power becomes brighter.

Further, a method of manufacturing a device according to the present invention is a method of manufacturing a predetermined device using the above-described imaging optical system of the present invention, wherein a pattern for the device is formed on a first surface thereof. Setting a mask (11), and setting a photosensitive substrate (18) on a second surface thereof; and irradiating the mask with illumination light for exposure after the first step. A second step of transferring an image of the pattern of the mask onto the photosensitive substrate via the imaging optical system;
Is included. As a result, the mask pattern is transferred with high resolution.

[0017]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of the present invention will be described below with reference to the drawings. FIG. 1 shows a schematic configuration of a projection exposure apparatus of this embodiment. In FIG. 1, exposure light IL emitted from an exposure light source 1 composed of a mercury lamp at the time of exposure is converged by an elliptical mirror 2 and a shutter (not shown) , And enters the fly-eye lens 4 via the input lens 3. As the exposure light IL, in addition to bright lines such as i-line and g-line of a mercury lamp, excimer laser light such as ArF excimer laser light or KrF excimer laser light, or harmonics of a metal vapor laser or a YAG laser are used. Can be used.

At this time, a large number of secondary light sources are formed on the exit surface of the fly-eye lens 6. An aperture stop plate 5 is disposed on the exit surface of the fly-eye lens 6, and two circular openings 6A and 6B are formed in the aperture stop plate 5 in a predetermined direction symmetrically with respect to the optical axis of the illumination optical system. The exposure light IL that has passed through these openings 6A and 6B is incident on a reticle blind (variable field stop) 8 through a first relay lens 7. The arrangement surface of the reticle blind 8 is conjugate with the pattern surface of the reticle 11, and the illumination area on the reticle 11 is defined by the opening shape of the reticle blind 8. The exposure light IL that has passed through the reticle blind 8 illuminates the pattern surface (lower surface) of the reticle 11 via the second relay lens 9 and the condenser lens 10. The illumination optical system of the present embodiment is configured by the exposure light source 1 to the aperture stop plate 5 and the first relay lens 7 to the condenser lens 10. Further, the aperture stop plate 5 is set at ± 180 ° around the optical axis.
A rotation driving device 39D for rotating is provided, and a main control system 38 for controlling the operation of the entire device is provided with the rotation driving device 39D.
The configuration is such that the rotation angle of the rotary diaphragm plate 5 can be set to a desired angle via the.

The exposure light IL that has passed through the pattern 13 of the reticle 11 is subjected to a modulation diffraction grating (hereinafter referred to as "modulation grating") 14, a projection optical system 15, and a demodulation diffraction grating (hereinafter referred to as "demodulation grating"). 16) and unnecessary light removing member 1
Through 7, the light enters the surface of the wafer 18 on which the photoresist is applied. In this case, the pattern surface of the reticle 11 and the surface of the wafer 18 are conjugate with respect to the projection optical system 15, and the pattern of the reticle 11 is projected onto the wafer 12 at a predetermined projection magnification (usually, for example, 1/4, 1/5, etc.). Reduction ratio)
Is transcribed. In this case, the modulation grating 14,
The resolution of the projection image transferred onto the wafer 18 by the demodulation grating 16 and the unnecessary light removing member 17 is determined by the projection optical system 1.
The resolution is higher than the resolution determined by the numerical aperture of 5 itself. Further, there are provided rotation drive systems 39A, 39B, 39C for rotating the modulation grating 14, the demodulation grating 16, and the unnecessary light removing member 17 within a range of ± 180 ° around the optical axis AX of the projection optical system 15, respectively. The main control system 38 controls the modulation grating 14, the demodulation grating 16, the unnecessary light removing member 1 via the rotation driving systems 39A to 39C and 39D as necessary.
7, and the aperture stop plate 5 is maintained at a predetermined angular relationship ± 1.
It is configured to be able to rotate by 80 °. Hereinafter, the z-axis is set in parallel with the optical axis AX of the projection optical system 15, and the orthogonal coordinate system in a plane perpendicular to the z-axis is described as the x-axis and the y-axis.

At this time, the reticle 11 is held on a reticle stage 12 that can be finely moved in the x direction, the y direction, and the rotation direction. The position of reticle stage 12 is measured with high accuracy by an external laser interferometer (not shown), and main control system 38 controls reticle stage 12 based on the measured values.
Is controlling the position. On the other hand, the wafer 18 is placed on a wafer stage 19 that is movable in the x and y directions via a wafer holder (not shown). Wafer stage 19
Is precisely measured by a laser interferometer (not shown), and the main control system 38 controls the position of the wafer stage 19 based on the measurement result. Then, the wafer 18 coated with the photoresist is loaded on the wafer stage 19 to perform alignment.
The operation of moving each shot area of the wafer 18 into the exposure field of the projection optical system 15 and the exposure operation are repeated in a step-and-repeat manner, so that the image of the pattern on the reticle 11 is It is transferred to the shot area. Thereafter, a circuit pattern corresponding to the pattern of the reticle 11 is formed in each shot area of the wafer 18 through a developing step, a pattern forming step, and the like.

Next, a mechanism for improving the resolving power in this embodiment will be described in detail. Here, as an example,
The pattern to be transferred on the pattern surface of the reticle 11 is x
Periodic pattern 13 arranged at a small pitch in the direction
And FIG. 2B shows an example of the spatial frequency distribution of the pattern 13 in the spatial frequency domain.
As shown in (b), the spatial frequency components of the pattern 13 are distributed in an oval region 21 whose longitudinal direction is the x direction.
FIG. 2C shows the reticle 11 by the projection optical system 15.
The circular area 24 corresponds to the numerical aperture of the projection optical system 15, and the range of the imaging light flux that can pass through the projection optical system 15 is Inside. In this case, since both ends of the oval region 21 corresponding to the spatial frequency distribution of the pattern 13 protrude from the circular region 24, the high spatial frequency component of the pattern 13 in the x direction is left as it is.
Can not pass. Therefore, the direction in which the resolving power of the pattern 13 should be improved is the x direction.

Therefore, in the aperture stop plate 5 of the illumination optical system shown in FIG. 1, two circular openings 6A and 6B are arranged along the x direction so as to correspond to the direction in which the resolving power is to be improved (x direction). The rotation angle is set as follows. As a result, the intensity of the light beam emitted from the reticle 11 at a large diffraction angle in the x direction increases, and a bright image is obtained. The modulation grating 14 is an amplitude type diffraction grating having a pitch d as shown in FIG. 2A, and its pitch direction (periodic direction) is shifted from the wafer 18 with respect to the direction in which the resolving power is to be improved (x direction). It is set in a direction indicated by an arrow 25 rotated counterclockwise by an angle θ (0 <θ <π / 2). Similarly, the demodulation grating 16 is an amplitude type diffraction grating having a pitch d2 as shown in FIG. 3A, and its pitch direction (periodic direction) is counterclockwise as viewed from the wafer 18 with respect to the x direction. It is set in the direction shown by the arrow 27 rotated by θ. Further, the projection optical system 15 of this example is telecentric on the wafer 18 side, and the unnecessary light removing member 17 is arranged in the y direction orthogonal to the direction in which the resolving power is to be improved (x direction), as shown in FIG. The light shielding plates 31A, 31B, 31C,... Are arranged at predetermined intervals. Light shielding plates 31A, 31B, ...
Are respectively parallel to the xz plane, whereby light beams inclined at a predetermined angle or more in the y direction with respect to the optical axis AX of the projection optical system 15 (axis parallel to the z axis) are shielded.

Under such a configuration, the pattern image of the reticle 11 is formed as follows. That is, the exposure light IL that has passed through the two openings 6A and 6B arranged in the x direction of the aperture stop plate 5 illuminates the pattern 13 of the reticle 11 obliquely via the condenser lens 10 and the like.
The exposure light IL that has passed through 3 is diffracted by the modulation grating 14 in a direction rotated by an angle θ with respect to the x-axis.
The 4th to 0th order diffracted light and ± 1st order diffracted light are projected into the projection optical system 15.
It is injected toward. At this time, the diffracted light incident on the projection optical system 15 is, as shown in FIG.
Are the 0th-order diffracted light, the + 1st-order diffracted light 22 in the oval region 21, and the -1st-order diffracted light 23. The + 1st-order diffracted light 22 and the -1st-order diffracted light 23 have spatial frequency components as if they had moved along the oval region 21 along a straight line 26 inclined at an angle θ with respect to the x-axis. In this case, the luminous flux that can pass through the projection optical system 15 out of the diffracted light shown in FIG.
As shown in (d), the components 21a, 21
2a and 23a only.

Next, the exposure light passing through the projection optical system 15 is
The light is diffracted again by the demodulation grating 16 in a direction rotated by an angle θ with respect to the x direction. FIG. 3B shows a demodulation grating 16.
FIG. 3B shows the spatial frequency distribution of the exposure light diffracted by the diffraction optical system. In FIG. The -1st-order diffracted light 24A and the + 1st-order diffracted light 24B that have moved back and forth along the straight line 26 inclined by the angle θ are emitted from the demodulation grating 16. As a result,
In FIG. 3B, a −1st-order diffracted light 22a ′ of the component 22a is generated in the −x direction of the 0th-order diffracted light component 21a,
In the + x direction of the component 21a, the + 1st-order diffracted light 2 of the component 23a
3a 'is generated, so that the original oval region 2 in FIG.
1 are restored by super-resolution. However, as can be seen from FIG.
In addition to the one component, unnecessary spatial frequency components in the y direction due to the overlapping image are also generated. Therefore, in order to remove such an overlapped image, the light beam inclined at a predetermined angle or more in the y direction with respect to the optical axis AX by the unnecessary light removing member 17 of FIG. 1 is shielded. From the spatial frequency components of the exposure light after passing through the unnecessary light removing member 17, as shown in FIG. 3C, the components of the regions 28 and 29 that are in contact with the oval region 21 in the y direction are removed. The reticle 1 on the wafer 18
The image of the first pattern 13 is transferred with a high resolution in the x direction. That is, the unnecessary light removing member 17 of the present embodiment can omit the second projection optical system required by the conventional diffraction grating method.

In order to facilitate a quantitative understanding of the operation of the unnecessary light removing member 17, the optical path of the light beam passing through the reticle 11 will be examined by a ray tracing method with reference to FIGS. 1, 4 and 5. I do. At this time, for simplicity, the projection magnification of the projection optical system 15 in FIG. Also, x
The direction is the meridional direction, and the y direction is the sagittal direction. Then, the direction cosine (cx, c
Assuming that the light beam emitted in the direction represented by (y, cz) is a, the traveling direction of the light beam a is represented by a vector having the x component and the y component of the direction cosine as follows.

A = (cx, cy) (5) At this time, the z component cz of the direction cosine is cx ² + cy ² +
From the relationship of cz ² = 1, the other two components cx and cy are always
Can be obtained from Here, the direction in which the resolving power is to be improved is the x direction, and the diffraction directions of the modulation grating 14 and the demodulation grating 16 are inclined by the angle θ with respect to the x direction. Diffracted light b _n
If made, the pitch of the modulation grid 14 is d, the exposure wavelength is lambda, the traveling direction of the diffracted light b _n can be expressed by the following direction cosine.

B _n = (cx ′, xy ′) = (cx− (nd / λ) · cos θ, cy− (nd / λ) · sin θ), (n = 0, ± 1, ± 2,...) (6) Further, when the diffracted light b _n passes through the projection optical system 15 and is diffracted again by the demodulation grating 16 to become the m-th order diffracted light _cmn , the projection optical system 15 is assumed to have the same magnification. Therefore, the pitch d2 of the demodulation grating 16 is set equal to d. The demodulating grating 16 also has a diffraction direction whose angle θ with respect to the x-axis.
And the traveling direction of the diffracted light _cmn is represented by _cmn =
When expressed by the direction cosine (cx ", cy"), each component is as follows.

Cx ″ = cx ′ − (md / λ) · cos θ, cy ″ = cy ′ − (md / λ) · sin θ (m = 0, ± 1, ± 2,...) (7) (6) From the expressions (7) and (7), the traveling direction of the diffracted light _cmn is as follows. c _mn = (cx− (nd / λ) · cos θ− (md / λ) · cos θ, cy− (nd / λ) · sin θ− (md / λ) · sin θ) (8) ) as is clear from equation when the following relation is satisfied, the traveling direction of the diffracted light c _mn are the same as (5) traveling direction of the light beam a from the reticle 11 of the formula (cx, cy) It can be seen that the pattern 13 on the reticle 11 can be reproduced.

N + m = 0 (9) In this case, considering only the strong 0th-order diffracted light and ± 1st-order diffracted light, the combination of the diffraction orders n and m satisfying the expression (9) is (n = 0, m = 0) and (n = 1, m = -1)
And (n = -1, m = 1). In this example,
The pattern of the reticle is reproduced mainly by a combination of these three diffraction orders. However, the diffracted light that does not satisfy the expression (9) is an extra diffracted light that forms an overlapped image, and it is necessary to remove these. Therefore, an unnecessary light between the demodulation grating 16 and the wafer 18 as an optical member for that purpose. A removing member 17 is provided.

[0030] Here, the traveling direction of the light beam a coming from the reticle 11, a = (cx, 0) if represented by the traveling direction of the diffracted light b _n passing through the modulation grid 14 from (6) Next become that way. b _n = (cx− (nd / λ) · cos θ, (−nd / λ) · sin θ) (10) When n = 1, the sine of the angle formed by the diffracted light b _n with the optical axis AX is: It looks like this:

[{Cx− (d / λ) · cos θ} ² + {(d / λ) · sin θ} ² ] ^1/2 (11) Accordingly, the sine of the light beam a from the reticle 11 is If the sine of the incident angle is smaller than the sine, the light beam having a smaller numerical aperture than the light beam emitted from the reticle 11 can pass through the projection optical system 15, which leads to an improvement in resolution. That is, when the following condition is satisfied, improvement in resolution can be expected.

Cx ² −2 (d / λ) · cx · cos θ + (d / λ) ² <cx ² (12) By solving this, the following equation is obtained. d <2λ · cx · cos θ (13) FIG. 4 shows the light beam emitted from the reticle 11 and the modulation grating 1
4 in the respective x and y directions of the diffracted light diffracted by
FIG. 5 is a diagram showing a direction cosine of a direction, and with reference to FIG.
A method for improving the critical resolution of a periodic pattern arranged in a direction will be described. Further, assuming that the numerical aperture NA of the projection optical system 15 in FIG.
Represents a circumference having a radius of sin β (= NA). That is, the inside of the circumference 30 represents the direction cosine of the light beam that can pass through the projection optical system 15. In FIG. 4, the center (optical axis AX) of the circumference 30 is defined as the origin O of the x-axis and the y-axis, and a point separated by x component cx of the direction cosine of the light beam from the reticle 11 in the x direction from the origin O is denoted by A. Assuming that a point at which the line segment OA intersects the circumference 30 is P, of the luminous flux having the direction cosine width like the line segment OA, the light beam of the direction cosine within the range of the line segment PA is usually projection optics. Cannot pass through system 15. However, in this example,
A modulation grating 1 that diffracts a light beam in a direction inclined by an angle θ counterclockwise as viewed from the wafer with respect to the x axis, such as the g axis.
4 is provided, and considering the -1st-order diffracted light diffracted thereby, the x component and the y component of the direction cosine are (0, 0).
The direction cosine of the _−1st-order diffracted light b ₋₁ of the light ray a is (− (d /
λ) · cos θ and − (d / λ) · sin θ). That is, FIG.
Moves the point O corresponding to the ray a to a point O 'along the g-axis.

A ray a having a direction cosine of (sin β, 0)
The direction cosine of the −1st-order diffracted light b ₋₁ is (sin β− (d / λ) ·
cos θ, − (d / λ) · sin θ). That is, in FIG.
The point P corresponding to the light ray a moves to the point Q in parallel with the g axis. Similarly, the direction cosine of the −1st-order diffracted light b ₋₁ of the ray a having the direction cosine of (cx, 0) is (cx− (d / λ) · cos θ, −
(D / λ) · sin θ), and the point A corresponding to the light ray a in FIG. 4 moves to the point B in parallel with the g axis. That is, due to the −1st-order diffraction, the direction cosine of the ray having the direction cosine in the range of the line segment PA moves to the range of the line segment QB, enters the circumference 30 having a radius of sin β, and moves the projection optical system 15 It can be transmitted. Therefore, the resolution can be improved. Also, a ray having a direction cosine in the range of the line segment OP can pass through the projection optical system 15 even as a 0th-order diffracted light, so that a ray whose direction cosine falls within the range of the line segment O′Q due to −1st-order diffraction is projected. Optical system 15
It may not be possible to pass through. Therefore, the length of the line segment O'Q is
It should be sin β. In the triangle OBQ of FIG. 4, the following is obtained by using the theorem of three squares.

{(D / λ) · cos θ−sin β} ² + {(d / λ) · sin θ} ² = sin ² β (14) Solving this gives the following. d / λ = 2 cos θ · sin β (15) This is a relational expression between the numerical aperture NA of the projection optical system 15 (= sin β) and the angle θ indicating the diffraction direction of the modulation grating 14. It is configured to satisfy Expression (15). In this case, since cx> sin β, (1
When the expression (5) is satisfied, the expression (13) is satisfied and the resolution is improved.

Then, the + 1st-order diffraction at the demodulation grating 16 whose pitch direction is the direction inclined at an angle θ in a counterclockwise direction as viewed from the wafer side with respect to the x-axis causes the line segment QB in FIG.
The direction cosine of a ray having a direction cosine is in the range of the line segment PA, and the luminous flux of the direction cosine in the range of the original line segment OA can be reproduced. However, the direction cosine of the 0th-order diffracted light by the demodulation grating 16 is It remains in the range of QB. Therefore, light rays having a direction cosine in the range of the line segment QB need to be removed so as not to form an image on the wafer.

In this case, it is necessary to prevent a light ray having a direction cosine of-(d / λ) · sin θ or less (more than an absolute value) in the y direction from being imaged on the wafer.
Considering the spread of the direction cosine of the direction, at least-｛d
/ (2λ)} · sin θ It is necessary to prevent light rays having a direction cosine of less than from being imaged on the wafer. Also, as before,
Considering the case where the + 1st-order diffracted light from the modulation grating 14 is converted into the -1st-order diffracted light and the 0th-order diffracted light by the demodulation grating 16, a light beam having a direction cosine of + (d / λ) · sin θ or more in the y direction is formed on the wafer. It can be seen that it is necessary not to image on top. In this case as well, considering the spread of the direction cosine of the light beam in the y direction, at least + {d / (2λ)} · sin
It is necessary to prevent light rays having a direction cosine of θ or more from being imaged on the wafer. That is, in the unnecessary light removing member 17,
It is sufficient to remove the light beam whose direction cosine cy in the y direction satisfies the following range so as not to form an image on the wafer.

Cy ≧ + {d / (2λ)} · sin θ or cy ≦ − {d / (2λ)} · sin θ (16) For this reason, as shown in FIG.
Reference numeral 7 is configured by arranging light shielding plates 31A, 31B,... Parallel to the zx plane at regular intervals in the y direction. When the distance between the light shielding plates 31A, 31B,... In the y direction is ds, and the length of the light shielding plates in the optical axis direction is h, the light rays R1, R2, R3,. The maximum value α ′ of the inclination angle in the y direction satisfies h · tan α ′ = ds. At this time, the direction cosine in the y direction is ± {d / (2λ)} · s
In order to prevent light rays outside the range of inθ from passing, the maximum value α ′ of the inclination angle is sin ⁻¹ [｛d / (2
λ)} · sin θ] (this is α) or less, so the interval ds needs to be determined so that the following equation is satisfied.

Ds ≦ h · tan [sin ⁻¹ {d · sin θ / (2λ)}] = h · tan α (17) At this time, if ds = h · tan α, then y
The light beam inclined at an angle α or more in the direction
A, 31B,... Do not form an image on the wafer. Further, since the interval ds needs to be sufficiently larger than the wave optical resolution limit of about 0.1 μm, the interval ds needs to be 10 μm or more. On the other hand, if the interval ds is too large, illuminance unevenness occurs depending on the image forming position in the y direction.
Smaller than the distance d.
s is desirably 100 μm or less. Therefore, it is desirable that the distance ds be in the range of 100 μm to 10 μm.

As a specific configuration of such an unnecessary light removing member 17, a member formed by etching metal to form the above-mentioned light shielding plates 31A, 31B,... Can be used. Further, a light-shielding band may be used instead of the light-shielding plate. As such a light-shielding band, one having a structure in which a light-shielding medium is embedded in a medium transparent to a used wavelength can be considered. However, when placing a light-shielding band inside a transparent medium,
If the light-shielding band absorbs light and generates heat, the refractive index of the transparent medium fluctuates, which may fluctuate the aberration of the light beam imaged on the wafer. It is desirable to scatter without.

It is to be noted that the direction cosine in the y direction is ±
As the unnecessary light removing member 17 for preventing light rays outside the range of {d / (2λ)} · sin θ from being imaged on the wafer, an optical system combining a plurality of prisms is used, It is also possible to use a phenomenon that a light beam that is going to be emitted at an angle or more is totally reflected. FIG. 7 shows a plurality of prisms that can be used as the unnecessary light removing member 17. In FIG. 7, light rays R1 to R4 from the demodulation grating 16 in FIG. 1 have a plane perpendicular to the optical axis AX directed to the incident side. The light is incident on the first prism 32. In the order from the first prism 32 toward the wafer 18, the surface inclined at the same angle as the inclined surface of the first prism 32 is set to the incident side, and the surface is inclined in the opposite direction at the same angle as the inclined angle. The second prism 33 having the first side as the exit side, and the third side in which the plane inclined by substantially the same angle as the inclined side on the exit side of the second prism 33 is set as the entrance side, and the plane perpendicular to the optical axis AX is set as the exit side. Are disposed. The inclination directions of the surfaces of the prisms 32 to 34 that are not perpendicular to the optical axis AX are all the y direction.
A light beam having a large inclination angle with respect to the optical axis AX whose direction cosine is ± {d / (2λ)} · sin θ or more in the direction
Do not reach the top.

That is, in FIG. 7, light rays R2 and R3 having a direction cosine in the y direction within a range of ± {d / (2λ)} · sin θ are incident on the wafer 18 via the prisms 32, 33 and 34. However, since the ray R1 is a ray having a direction cosine in the y direction of {d / (2λ)} · sin θ or more, the ray R1 is incident on the exit-side surface of the prism 32 at a critical angle or more, is totally reflected, and the ray R It becomes' 1 and does not reach the surface of the wafer 18. Also,
Since the ray R4 is a ray having a direction cosine in the y direction of − ｛d / (2λ)} · sin θ or less, the ray R4 is incident on the exit-side surface of the prism 33 at an incidence angle equal to or greater than the critical angle, and is totally reflected. As a result, the light ray R'4 does not reach the surface of the wafer 18.

When the numerical aperture NA of the projection optical system 15 is small, in order to prevent light rays having an inclination angle of not less than sin ^-1 [{d / (2λ)} · sin θ] from being imaged on the wafer, the prisms 32 to
If the inclination angle of 34 is not increased, the critical angle will not be as described above. When the critical angle of the prisms 32-34 is increased,
It is necessary to increase the distance from the projection optical system 15 to the wafer 18, and it is difficult to design such a projection optical system 15. Therefore, it is desirable to apply a coating on the inclined surface of the prism that has significantly different angular characteristics such that total reflection occurs even within a critical angle.

The prisms 32, 33, 34
May be used as a polarizing beam splitter. These polarization beam splitters transmit a polarization component (P-polarized light) in a direction (y direction) parallel to the direction of inclination of the surface inclined with respect to the optical axis AX, and a direction (x direction) perpendicular to the direction of inclination. (S-polarized light). However, in this example, the inclination direction (y direction) of the polarizing beam splitter is perpendicular to the periodic direction (x direction) of the periodic pattern, and P
When the polarized light forms an image on the wafer, the polarization component in the x direction is removed, so that the resolution is improved as compared with ordinary natural light, which is convenient.

In the above embodiment, the modulation grating 1
Although a normal amplitude type diffraction grating is used as 4 and the demodulation grating 16, it is desirable to use a phase type diffraction grating from the viewpoint of diffraction efficiency. When the phase-type diffraction grating is used as described above, almost no zero-order diffracted light is generated, so that an image is formed as described below with reference to FIG.

FIG. 5 shows the x-axis and y-axis of the light beam emitted from the reticle 11 corresponding to FIG. 4 and the diffracted light diffracted by the modulation grating 14 (a phase type diffraction grating is used here). FIG. 5 is a diagram illustrating the direction cosine of the direction. In FIG. 5, the direction cosine of a light beam that can pass through the projection optical system 15 inside the circumference 30 having a radius of sin β (= NA) is illustrated. In FIG. 5, among the light fluxes emitted from the reticle 11 and extending in the direction cosine only in the x direction and within the range of the line segment OA, the light flux having the direction cosine in the range of the line segment PA is equal to or more than sin β. Since there is a direction cosine in the x direction, the light cannot pass through the projection optical system 15 as it is. On the contrary,
Since the 0th-order diffracted light by the modulation grating 14 does not exist, a light flux having a direction cosine in the line segment OA is diffracted by the modulation grating 14 in only two directions of the -1st-order diffracted light and the + 1st-order diffracted light. The direction cosine of the -1st-order diffracted light of the light beam having the direction cosine moves along the g-axis inclined by an angle θ with respect to the x-axis and falls within the line segment O′B.

Here, assuming that the pitch of the modulation grating 14 as a phase type diffraction grating is d and the exposure wavelength is λ, the light ray a having the direction cosine (0, 0) in FIG. The cosine is (-｛d / (2λ)} · cos θ,-｛d /
(2λ)} · sinθ) diffracted light b _-1 next to the point O moves to point O '. The ray a having the direction cosine of (sin β, 0) is −
By the first-order diffraction, the direction cosine is (sin β− ｛d / (2
λ)} · cosθ, - { d / (2λ)} · sinθ) diffracted light b _-1 next, the point P is shifted to the point Q. Similarly, the direction cosine is (cx,
0) has a direction cosine of (cx− {d / (2λ)} · c
os θ, − {d / (2λ)} · sin θ), and becomes diffracted light b ₋₁ ,
Point A moves to point B.

That is, the -1st-order diffracted light of the light beam whose direction cosine is the line segment OA is the light beam whose line segment O'B is the direction cosine. As is clear from FIG. 5, OO '= OB = sin β
It is. Also, since the line segment O'B is parallel to the x-axis,
The point B is symmetric with respect to the point O ′ with respect to the y axis, and the direction cosine of the point B is ({d / (2λ)} · cos θ, − {d / (2λ)} · sin
θ), the following equation holds.

O′B = cx = 2 ｛d / (2λ)} · cos θ = (d / λ) · cos θ (18) Accordingly, in the light beam emitted from the reticle 11, the maximum resolution can be improved. It can be seen that the numerical aperture is (d / λ) · cos θ. OO ′ = d / (2λ) = sin β
Therefore, if cx is the maximum numerical aperture capable of improving the resolving power in the light beam emitted from the reticle 11, cx = 2
・ Sin β ・ cos θ That is, the resolution is the projection optical system 1
Although the numerical aperture NA cannot be increased to twice or more, the resolving power can be improved as compared with the case where only the projection optical system 15 is used, as long as the following equation is satisfied.

2 · cos θ> 1 (19) However, among the diffracted light beams transmitted through the demodulation grating 16 composed of the phase type diffraction grating, the light beams having the direction cosine of the line segment O′B are respectively + 1st-order diffracted light beams. Is a light beam having a direction cosine of the line segment OA, and a -1st-order diffraction is a light beam having a direction cosine of the line segment O ″ C. However, since there is no zero-order diffracted light,
There is no light beam having the direction cosine of the line segment O'B. Therefore, the unnecessary light removing member 17 for removing the redundant image is used.
Then, it is sufficient to remove the light flux having the direction cosine of the line segment O "C, and considering the spread of the light flux in the y direction, the light flux having the y direction cosine of at least the line segment O'B is formed on the wafer. That is, as in the case of the amplitude type diffraction grating, ± {d / (2λ)} in the y direction.
It is only necessary to prevent rays of direction cosine equal to or greater than sin θ from being imaged on the wafer. However, even if the pitch d of the amplitude type and the phase type diffraction grating is the same, the direction cosine of the diffraction direction of the ± 1st-order diffracted light of the phase type diffraction grating is half of that of the amplitude type diffraction grating.

Therefore, as is apparent from a comparison of FIG. 5 with FIG. 4, the phase type diffraction grating does not form an image on the wafer with a direction cosine in the y direction larger than that of the amplitude type diffraction grating. When the unnecessary light removing member 17 is formed of a light shielding plate or a light shielding band, the width h of the light shielding plate or the light shielding band in the optical axis direction can be reduced.
If a plurality of prisms shown in FIG. 7 are used as 7, the inclination angle of each prism can be reduced. That is, it becomes easy to form an optical member for preventing an unnecessary overlapping image from being formed on the wafer.

For example, by using a plurality of prisms 32 to 34 shown in FIG.
Considering the case where the inclined light beam is totally reflected at a critical angle of a predetermined prism, if the inclination angle of the prism is ψ and the refractive index is Np, the following relationship is established. sin (φ + ψ) = 1 / Np (20) Specifically, when Np = 1.5, φ + ψ = 41.81 °. For example, the numerical aperture NA of the projection optical system 15 is 0.5
Assuming that {d / (2λ)} · sin θ = 0.14 when the amplitude type diffraction grating is used as shown in FIG.
Light rays having a direction cosine of 0.14 or more in the y direction may be prevented from being imaged on the wafer. Thus, the angle φ is as follows.

Sin φ = 0.14 / Np = 0.14 / 1.5 (21) When this is solved, φ = 5.36 °, and the inclination angle of the prism is given by ψ = 36.45 °.
However, similarly, the numerical aperture NA of the projection optical system 15 is set to 0.1.
In the case of 56, using a phase type diffraction grating as shown in FIG.
If {d / (2λ)} · sin θ = 0.28, then 0.
Since it is sufficient to prevent light rays having a direction cosine of 28 or more in the y direction from being imaged on the wafer, the angle φ is as follows.

Sin φ = 0.28 / Np = 0.28 / 1.5 (22) When this is solved, φ = 10.76 °. Therefore, the inclination angle of the prism is given by ψ = 31.05 from equation (20).
゜, and a smaller inclination angle than that of the amplitude type diffraction grating is sufficient. Therefore, when a phase-type diffraction grating is used, the area occupied by these prisms in the optical axis direction can be reduced, and the design of the projection optical system 15 can be given a corresponding amount.

As described above, in this embodiment, the resolution in the x direction can be improved, but the pattern to be transferred may include a periodic pattern arranged at a fine pitch in various directions. In such a case, in FIG. 1, during exposure, the main control system 38 controls the rotation driving devices 39D, 39A, 3
9B, 39C, the apertures 6A, 6B of the aperture stop plate 5
, The diffraction direction of the modulation grating 14, the diffraction direction of the demodulation grating 16, and the arrangement direction of the light shielding plate of the unnecessary light removing member 17 are rotated at least 180 ° while maintaining the above-described positional relationship. In addition, you may make it rotate ± 180 degrees, ie, 360 degrees. As a result, the resolving power of the patterns arranged in all directions can be increased to be higher than the resolving power determined by the numerical aperture of the projection optical system 15.

In the above embodiment, the reticle 11 is illuminated with the exposure light inclined in the direction in which the resolving power is to be improved (the x direction in FIG. 1) by using the aperture stop plate 5.
Thus, the intensity of the light flux emitted from the pattern 13 of the reticle 11 at a large diffraction angle in the x direction on the pupil plane of the projection optical system 15 does not decrease, and a high resolution can be obtained. When the aperture stop plate 5 is used as described above, in order to further increase the inclination angle of the exposure light to the reticle 11, as disclosed in Japanese Patent Application Laid-Open No. 5-188577,
A phase type diffraction grating whose pitch direction is the direction in which the resolving power is to be improved (x direction) may be arranged on the incident side of the reticle 11. Assuming that the pitch of the phase type diffraction grating is dp, the exposure light from the illumination optical system is further illuminated from the direction increased or decreased by λ / (2dp) as the direction cosine.

FIG. 8 shows a light source image on the pupil plane (optical Fourier transform plane for the reticle 11) of the illumination optical system when illumination is performed using such a phase type diffraction grating. In FIG. 8, a circular light source image (not shown) in the case where the aperture of the aperture stop plate 5 is a normal circular aperture is converted into a direction cosine in the x direction by the phase type diffraction grating, ± λ /
(2dp) shifted to the light source images P _-1 and P _1, respectively.
It has become. In FIG. 8, the pupil on the entrance side of the projection optical system 15 (the passable area determined by the numerical aperture on the entrance side of the projection optical system) is P ₀ . Further, let the numerical aperture (NA _IL ) on the exit side of the illumination optical system be sin γ and the numerical aperture on the entrance side of the projection optical system PL be sin β, and the following equation is satisfied.

Sin β <λ / (2dp) <sin β + sin γ (23) In this case, λ / (2dp) is approximated to (sin β + sin γ)
The pupil P of the projection optical system 15₀And the illumination optics
Light source image P by phase type diffraction grating on pupil plane _-1, P₁
And overlap only at the ends. Therefore, the illumination light of FIG.
Positions of apertures 6A and 6B of aperture stop plate 5 on the pupil plane of the academic system
With a small circular opening Q in the overlap region of FIG.₁₁, Q₁₂
(Or Q_-11, Q_-12), The reticle
Only the exposure light emitted from 11 at a large angle in the x direction
It can pass through the projection optical system 15 and has a fine pitch in the x direction.
Turns are projected with high resolution.

Another example of an optical element that preferentially forms an image of a tilted exposure light on a wafer, such as a phase-type diffraction grating provided on the reticle 11, is as follows.
As shown in FIG. 9, there is a light-shielding mask 36 provided to face the emission side of the reticle 11. In FIG.
On the pattern surface of the reticle 11, light shielding patterns 35A, 35B, 25C,... Are formed at a predetermined pitch in the x direction. On the other hand, light-shielding patterns 35A, 35B, 35 are provided at positions separated by a predetermined distance from the pattern surface of reticle 11.
A mask 36 having the same arrangement as that of C and having light-shielding patterns 37A, 37B, 37C,... Using the mask 36,
In order to preferentially pass only the inclined light flux of the exposure light emitted from the reticle 11, the distance between the exit surface of the reticle 11 and the light-shielding mask 36 is set to H, and the light-shielding of the mask 36 is performed. The pitch of the pattern is dp ',
Assuming that the numerical aperture NA _IL on the exit side of the illumination optical system is sin γ, the following relationship should be satisfied.

Dp ′ / (2 tan γ) ≦ H <dp ′ / [2 tan (γ−δ)] (24) where the angle δ is an angle satisfying the relationship 0 <δ <γ.
When the light-shielding mask 36 whose light and dark phases are reversed in this way is used, there is diffraction due to the mask 36.
The efficiency of preferentially transmitting the inclined illumination light is lower than that of the phase type diffraction grating arranged on the incident side of the mask. In the above embodiment, the image forming optical system of the present invention is applied to a projection exposure apparatus. However, the image forming optical system of the present invention can be applied to an optical microscope and the like. As described above, the present invention is not limited to the above-described embodiment, and can take various configurations without departing from the gist of the present invention.

[0060]

According to the image forming optical system of the present invention, since an optical member for removing unnecessary light from the diffracted light from the demodulating means is provided, the conventional optical system uses only one projection optical system. The technique of improving the critical resolution by the diffraction grating method can be applied, and there is an advantage that the critical resolution of a periodic pattern can be improved. Further, there is an advantage that the critical resolution of periodic patterns arranged in all directions can be improved by rotating the modulating means, the demodulating means, the optical member and the like as necessary.

The demodulating means includes a diffraction grating disposed in a plane orthogonal to the optical axis of the projection optical system, and the diffraction grating is provided in the plane in a predetermined direction for improving the resolving power of the projected image in the plane. And a diffraction pattern having a predetermined pitch d along a direction inclined by an angle θ with respect to the angle .lambda., Where .lambda. Is the wavelength of the illumination light, and two angles .alpha.1 and .alpha.2 are defined by equations (2) and (3) When the optical member passes a light beam having an angle of inclination α1 or α2 with respect to the optical axis of the projection optical system in the light beam from the demodulating means, unnecessary overlapping images are removed. And a clear projected image can be obtained.

Further, the optical member sets a plurality of flat plates parallel to the optical axis of the projection optical system in a direction perpendicular to the direction in which the resolution of the projected image is to be improved and in a direction perpendicular to the direction in which the resolution is to be improved. When the optical members are formed at a pitch, the optical members can be easily and inexpensively formed. Further, the optical member has a reflecting surface for reflecting the unnecessary light beam outside the imaging area on the second surface of the projection optical system in order to remove the unnecessary light beam from the demodulation means. In the case where the prisms are included, there is an advantage that an effective light flux is not blocked.

According to the exposure apparatus of the present invention, since the image forming optical system of the present invention is provided, the pattern on the mask can be transferred onto the photosensitive substrate with a resolution higher than the resolution of the projection optical system itself. Further, according to the device manufacturing method of the present invention, the pattern on the mask can be transferred onto the photosensitive substrate with high resolution using the imaging optical system of the present invention.

[Brief description of the drawings]

FIG. 1 is a partially cutaway perspective view showing a projection exposure apparatus according to an example of an embodiment of the present invention.

FIG. 2 is an explanatory diagram of an operation of a modulation grating 14 of FIG.

FIG. 3 is an explanatory diagram of an operation of a demodulation grating 16 of FIG. 1;

FIG. 4 is a diagram showing a direction cosine of a direction perpendicular to the optical axis of each diffracted light when the modulation grating 14 and the demodulation grating 16 are amplitude type diffraction gratings in the embodiment of the present invention.

FIG. 5 is a diagram showing a direction cosine of a direction perpendicular to the optical axis of each diffracted light when the modulation grating 14 and the demodulation grating 16 are phase diffraction gratings in the embodiment of the present invention.

FIG. 6 is a side view showing an example of the unnecessary light removing member 17 of FIG.

FIG. 7 is a side view showing a plurality of prisms that can be used as the unnecessary light removing member 17 of FIG.

FIG. 8 is an explanatory diagram of an example of an optical element that preferentially forms an image of an inclined illumination light on a wafer.

FIG. 9 is an explanatory diagram of another example of an optical element that preferentially forms an image of an inclined illumination light on a wafer.

[Explanation of symbols]

 Reference Signs List 5 aperture stop plate 11 reticle 14 diffraction grating for modulation (modulation grating) 15 projection optical system 16 diffraction grating for demodulation (demodulation grating) 17 unnecessary light removing member 18 wafer 19 wafer stage 31A, 31B, ..., 31E light shielding plate 32 , 33, 34 Prism 38 Main control system 39A-39D Rotary drive

Claims (6)

[Claims] A projection optical system for projecting an image of a pattern on a first surface onto a second surface under predetermined illumination light; and a projection optical system disposed between the first surface and the projection optical system. First
A modulating means for generating an n-th order diffracted light (n is an integer) of a light beam from a surface pattern; and an m-th order diffracted light (m) of a light beam from the projection optical system disposed between the projection optical system and the second surface. Is an integer), and an optical member disposed between the demodulation means and the second surface. The second surface side of the projection optical system is substantially telecentric. The projection optical system has a numerical aperture through which the n-th order diffracted light from the modulating means passes, and the optical member includes (n +
m = 0). An imaging optical system for removing unnecessary light beams of orders that do not satisfy the relationship of m = 0). 2. The imaging optical system according to claim 1, wherein said demodulating means includes a diffraction grating arranged in a plane orthogonal to an optical axis of said projection optical system, and said diffraction grating is provided in said plane. And a diffraction pattern having a predetermined pitch d along a direction inclined by an angle θ with respect to a predetermined direction in which the resolving power of the projection image is to be improved, where λ is the wavelength of the illumination light and two angles α1 And α2
Is defined as follows: α1 = sin ⁻¹ [d · sin θ / (2λ)] α2 = sin ⁻¹ (d · sin θ / λ) The optical member performs the projection within the light beam from the demodulation unit. An imaging optical system characterized in that a light beam whose inclination angle with respect to the optical axis of the optical system is within an angle α1 or α2 is transmitted. 3. The imaging optical system according to claim 1, wherein the optical member has a plurality of flat plates parallel to a direction in which a resolving power of a projected image is to be improved and an optical axis of the projection optical system. Are arranged at a predetermined pitch in a direction perpendicular to the direction in which the resolution is to be improved. 4. The imaging optical system according to claim 1, wherein said optical member is provided on said second surface of said projection optical system in order to remove said unnecessary light beam from said demodulating means. An imaging optical system, comprising: a plurality of prisms each having a reflection surface for reflecting the unnecessary light beam outside the imaging region. 5. The imaging optical system according to claim 1, wherein the illumination optical system illuminates a mask disposed on the first surface, and the illumination optical system is disposed on the second surface. An exposure apparatus comprising: a substrate holding member configured to hold a photosensitive substrate; and transferring an image of the pattern of the mask onto the photosensitive substrate via the imaging optical system. 6. A method for manufacturing a predetermined device using the imaging optical system according to claim 1, wherein a pattern for the device is formed on the first surface. A first step of setting a mask and setting a photosensitive substrate on the second surface; and irradiating the mask with illumination light for exposure after the first step to form an image of a pattern of the mask. Transferring the image onto the photosensitive substrate via an image optical system.