JPH10116770A - Method and apparatus for detecting position - Google Patents

Abstract

PROBLEM TO BE SOLVED: To use luminous fluxes which are adjacent as far as possible and which advance along an optical path when the dislocation amount between a reticle mark and a wafer mark is detected via a projection optical system by using alignment light whose wavelength is different from that of exposure light. SOLUTION: By using two luminous fluxes L1, L2 whose frequency is slightly different, the dislocation amount between a reticle mark 14X and a wafer mark 16X is detected via a projection optical system PL by using a two-luminous-flux interference system. The pitch of the reticle mark 14X and that of the wafer mark 16X are set in such a way that diffraction light is not overlapped with each other, the angle of crossing of the two luminous fluxes L1, L2 is changed over by a time-sharing system, and the position of the reticle mark 14X and that of the wafer mark 16X are detected sequentially.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a position detecting method and apparatus for detecting the amount of displacement between two marks, for example, a semiconductor device, an image pickup device (such as a CCD), a liquid crystal display device, or a thin film magnetic head. In a projection exposure apparatus used in a lithography process for manufacturing a mask, a TTR for detecting the amount of misalignment between an alignment mark on a mask and an alignment mark on a photosensitive substrate from above the mask via a projection optical system (Through-the-reticle) type alignment sensor or proximity type exposure apparatus that exposes the mask by closely contacting the mask and the photosensitive substrate. The position of the alignment mark on the mask and the alignment mark on the photosensitive substrate It is suitable for use in an alignment sensor or the like for detecting a shift amount.

[0002]

2. Description of the Related Art Conventionally, in order to form a fine pattern of a semiconductor element or the like, a pattern of a reticle (or a photomask or the like) as a mask is exposed via a projection optical system to a wafer (or a glass plate or the like) as a photosensitive substrate. A projection exposure apparatus such as a stepper for transferring to the above shot areas is used. In such a projection exposure apparatus, in order to perform high-accuracy alignment (alignment) between a circuit pattern already formed in each shot area on a wafer and a reticle pattern to be transferred from the reticle, the reticle and the wafer are aligned. An alignment sensor as a position detecting device for detecting the amount of positional deviation, and an alignment mechanism including a stage mechanism for positioning the reticle and the wafer based on the detection result are provided.

With such a conventional alignment sensor,
One of the methods capable of detecting the amount of misalignment between the reticle and the wafer with the highest accuracy is to use an alignment mark (reticle mark) on the reticle and an alignment mark (wafer mark) on the wafer via a projection optical system from the reticle. This is a TTR method for detecting the amount of displacement. Also, this TT
A type of alignment sensor that can detect the position of each alignment mark with high precision in an alignment sensor applicable to the R method. A diffraction grating alignment mark is irradiated with a coherent light beam from two directions, and the same alignment mark is used from the alignment mark. There is known a two-beam interference method for detecting the position of an alignment mark based on the phase of interference light composed of a plurality of diffracted lights generated in a direction. Two beam interference method
Also called LIA (Laser Interferometric Alignment) method. The two-beam interference method is further divided into a heterodyne interference method using two light beams having slightly different frequencies and a homodyne interference method using two light beams having the same frequency.

[0004] In the alignment sensor, a light beam (non-exposure light) that is insensitive to the photoresist on the wafer is usually used as alignment light. However, if non-exposure light is used as alignment light in a TTR type alignment sensor, the chromatic aberration of the projection optical system is corrected with respect to the exposure light. Conjugation between the reticle and the wafer cannot be maintained, and it becomes difficult to detect the positional deviation between the reticle mark and the wafer mark with high accuracy using a common detection system.

Therefore, for example, in Japanese Patent Application Laid-Open No. 4-7814 (US Pat. No. 5,214,489),
The grating pitch of the reticle mark is set to の of the grating pitch of the wafer mark in consideration of the projection magnification, and the respective optical beat signals obtained from the reticle mark and the wafer mark are compared by using independent detection systems. Detect 2
A light beam interference type alignment sensor is disclosed.
In addition, in this alignment sensor, it has been considered that the projection magnification under the alignment light composed of the non-exposure light is different from the projection magnification under the exposure light.

[0006]

As described above, the two-beam interference type alignment sensor can detect the position of the alignment mark with extremely high accuracy. However, with a conventional two-beam interference type alignment sensor,
In order to detect the amount of displacement between the reticle mark and the wafer mark, alignment light is irradiated so as not to overlap with the reticle mark and the wafer mark, and the positions of the two marks are detected individually and in parallel. And the optical system is complicated.

Further, in order to detect the positions of two marks in parallel by the two-beam interference method in such a manner, the reticle mark and the reticle mark are spatially prevented from interfering with each other from the reticle mark and the wafer mark. It is necessary to shift the detection position of the wafer mark from a conjugate positional relationship. As a result, the optical path of the light beam incident on the two marks and the optical path of the interference light from the two marks are not the same, and the two marks have different effects such as air fluctuation, which may result in a detection error. there were.

Further, recently, excimer laser light such as a KrF excimer laser (wavelength: 248 nm) or an ArF excimer laser (wavelength: 193 nm) has been used as exposure light. However, even in this case, He-N, which is a light source easily available at present, is used as the alignment light.
Wavelength 600-800n from e-laser, semiconductor laser, etc.
m of red light is used. As a result, large chromatic aberration (axial chromatic aberration, lateral chromatic aberration, and the like) is generated in the projection optical system with respect to the alignment light, and the correction is necessary in order to perform TTR-type alignment. In particular, when there is a relative positional deviation between the reticle mark and the wafer mark in a state where the chromatic aberration of magnification is not corrected, the positional deviation amount of the alignment light and the positional deviation amount of the exposure light are different, and a detection error occurs. There was an inconvenience. For example, when the wavelength of the exposure light is 248 nm, a magnification error of 5% with respect to the alignment light may occur in the projection optical system.
In this case, assuming that the displacement amount of the two marks in the alignment light is 2 μm, this displacement amount is 2.1 μm (or 1.9 μm) in the exposure light, and the detection error is 0.1 μm (100 nm). ).

In this connection, the above-mentioned Japanese Patent Application Laid-Open No. 4-7814 is disclosed.
In the two-beam interference type alignment sensor disclosed in Japanese Patent Application Laid-Open Publication No. H11-157, the pitch of interference fringes of the two light beams irradiated on the reticle mark and the light emitted on the wafer mark are considered in consideration of the magnification error of the projection optical system with respect to the alignment light. And the pitch of the interference fringes of the two light beams. Specifically, this means that an independent detection system is provided for the reticle mark and the wafer mark, and the observation magnification is changed for the reticle mark and the wafer mark. However, having two optical systems in which the observation magnification is changed between the reticle mark and the wafer mark as described above has a disadvantage that the entire optical system is complicated and the manufacturing cost is increased.

In view of the above, the present invention provides a simple method that uses a light beam that travels along an optical path as close as possible when detecting the amount of displacement between a reticle mark and a wafer mark using alignment light of a predetermined wavelength. It is an object of the present invention to provide a position detection method capable of accurately detecting the amount of displacement between these two marks via a simple optical system. Furthermore, the present invention
Another object of the present invention is to provide a position detecting device capable of performing the position detecting method.

[0011]

According to the position detecting method of the present invention, a pattern formed on a mask (R) is transferred onto a photosensitive substrate (W) under illumination light (IL) for exposure. Using the position detection light (L) in a predetermined wavelength range, the grid-like positioning mark (14X) on the mask (R) and the grid-like positioning mark (16) on the photosensitive substrate (W) are used.
X) in a position detection method for detecting the amount of positional deviation from
Pitch of alignment mark (14X) on mask (R) and alignment mark (16X) on photosensitive substrate (W)
Are determined so that the diffracted lights from the two alignment marks (14X, 16X) by the position detection light do not overlap, and the position detection light is simultaneously transmitted to the two alignment marks (14X, 16X). The two alignment marks (14X,
16X), based on the phase of the diffracted light from the two alignment marks.

According to the present invention, for example, as shown in FIG. 3, a pattern on a mask (R) is projected onto a projection optical system (P).
L), the position of the projection exposure apparatus that transfers the image to the photosensitive substrate (W) via the two alignment marks (14).
X, 16X) can be detected using the position detection light (L1, L2) passing through a coaxial (along the same optical axis) adjacent optical path. That is, assuming that the two-beam interference method is used, first, as shown in FIG. 3A, the position detection light (L) is adjusted to the pitch of the alignment mark (16X) on the photosensitive substrate (W).
When the intersection angle of (1, L2) is set, the interference light (L) composed of, for example, ± 1st-order diffracted light from the alignment mark (16X)
W), but the direction of the diffracted light from the alignment mark (14X) on the mask (R) is the direction of the interference light (LW).
, It is possible to detect only the interference light (LW). Next, as shown in FIG. 3B, when the intersection angle of the position detection light (L1, L2) is set in accordance with the pitch of the alignment mark (14X) on the mask (R), the alignment mark Interference light (LR) composed of, for example, ± 1st-order diffracted light is generated from (14X), but the direction of the diffracted light from the alignment mark (16X) is different from the direction of the interference light (LR). Interference light (LR)
Only can be detected. Therefore, the interference light (LW, L
R), two alignment marks (14X, 1)
6X) can be accurately detected, and the amount of displacement between the two marks under the position detection light can be accurately detected from the difference between the detection results.

In this case, the irradiation of the position detection light to the two marks can be performed by a common optical system, so that the optical system is simplified. Further, the projection magnification under the position detection light (alignment light) and the projection magnification under the illumination light for exposure (exposure light) are determined in advance, and the detection result is used as the illumination light for exposure. The amount of displacement under the illumination light for exposure is accurately detected by converting into the amount of displacement under the condition (1).

In this case, assuming that the pattern of the mask (R) has been transferred onto the photosensitive substrate via the projection optical system (PL), the alignment mark (14) on the mask (R) is assumed.
X) the pitch of Pr, the pitch of the alignment mark (16X) on the photosensitive substrate (W) as Pw, and the photosensitive substrate from the mask (R) of the projection optical system (PL) under the position detection light. When the projection magnification for (W) is β, it is desirable to set the pitches Pr and Pw using the arbitrary integers m and n of 1 or more so that the following two relationships are satisfied.

Pw ≠ m · β · Pr (1) n · Pw ≠ β · Pr (2) More specifically, the pitch Pw is set to m · β · Pr or β · Pr.
For example, a value different from the value determined by Pr / n by several percent may be set. Thus, the zero-order light and the first-order or higher-order diffracted light from the light beam incident on the alignment mark (14X) on the mask (R) are incident on the alignment mark (16X) on the photosensitive substrate (W). There is no combination traveling in the same direction in the diffracted light obtained by this. vice versa,
Even if the zero-order light and the first-order or higher-order diffracted light from the alignment mark (16X) on the photosensitive substrate (W) are diffracted by the alignment mark (14X) on the mask (R), they travel in the same direction. There is no diffracted light. That is, the diffracted lights from the two alignment marks do not mix with each other.

The above equations (1) and (2) represent the mask (R) in the projection exposure apparatus for transferring the pattern on the mask (R) to the photosensitive substrate (W) via the projection optical system (PL). ) Is used to detect the alignment mark (14X) on the photosensitive substrate (W) and the alignment mark (16X) on the photosensitive substrate (W), but the mask (R) is used without using the projection optical system (PL). The present invention can also be applied to an exposure apparatus such as a proximity method that exposes the upper pattern and the photosensitive substrate (W) in close or close contact with each other. In the case of the exposure apparatus of the proximity type or the like, the projection optical system (PL) does not exist, but since the mask (R) and the photosensitive substrate (W) are arranged close to each other, the projection optical system (PL) Equations (1) and (2) can be treated assuming that the projection magnification β is 1. Therefore, even in the case of an exposure apparatus of the proximity type or the like, unnecessary diffracted light from two alignment marks is not mixed, and only necessary diffracted light can be obtained.

Therefore, even if the zero-order light of the light beam incident on the positioning mark (14X) is directly incident on the positioning mark (16X), the positioning mark (14X) can be obtained.
X) Other diffracted lights are alignment marks (16X)
Since the light does not become the noise light of the diffracted light from the light source, and vice versa, it is possible to coaxially input the position detection light to the two marks.

At this time, the phase of the diffracted light from the alignment mark (14X) on the mask (R) by the position detection light and the alignment mark (1) on the photosensitive substrate (W).
It is desirable to detect the phase of the diffracted light from 6X) by time division. By the time division, the positions of the two alignment marks (14X, 16X) can be sequentially detected through a common light receiving system.

Further, by comparing the phases of the diffracted lights from the two alignment marks (14X, 16X) by the position detection light with the phases of the predetermined reference signals, the two alignment marks are compared. The position shift amount may be obtained. When position detection is performed by the two-beam interference method and the heterodyne interference method, for example, a reference signal can be generated by mixing a drive signal of an acousto-optical element for generating two light beams having different frequencies. Then, with respect to the alignment mark (16X) on the photosensitive substrate (W), the positional deviation amount thus obtained based on the difference in the projection magnification between the position detection light and the exposure illumination light. Is converted into a displacement amount under the illumination light for exposure, the displacement amount under the illumination light for exposure can be accurately obtained.

Further, the position detecting apparatus according to the present invention has a predetermined wavelength prior to transferring the pattern formed on the mask (R) to the photosensitive substrate (W) under the illumination light (IL) for exposure. Using the position detection light (L) of the area, the amount of misalignment between the grid-shaped alignment mark (14X) on the mask (R) and the grid-shaped alignment mark (16X) on the photosensitive substrate (W) In the position detection device for detecting the position, two light beams (L) having different frequencies from the position detection light beam (L) are used.
1, L2), a driving means (24) for applying ultrasonic waves having a variable frequency to the light modulating means, and two light beams (23) generated by the light modulating means (23). Irradiation optical system (27, 28) that guides L1, L2) together with the projection optical system (PL) to the alignment mark (14X) on the mask (R) and the alignment mark (16X) on the photosensitive substrate (W). And a detector (30) for photoelectrically detecting interference light (LR, LW) due to a plurality of diffracted lights generated in the same direction from the two alignment marks (14X, 16X). The frequency of the ultrasonic wave applied from the driving means (24) to the light modulation means (23) is adjusted in accordance with the pitch of the mark to be detected among the two alignment marks (14X, 16X). is there.

According to the position detecting device of the present invention, 2
When the pitch of the two positioning marks (14X, 16X) is determined so that the diffracted lights from the respective marks do not overlap, the light modulating means (23) in a time-division manner.
The frequency of the ultrasonic wave supplied to the light modulating means (23) is switched so that two light beams (L
1, L2) are sequentially aligned with the alignment mark (14).
X) and the angle according to the pitch of the alignment mark (16X). As a result, the positioning marks (14X, 14X,
16X) are sequentially detected, and the alignment method of the present invention is performed.

[0022]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of the present invention will be described below with reference to the drawings. In this example, a reticle pattern image is projected onto a wafer via a projection optical system while a part of the reticle pattern is projected onto the wafer, and the reticle pattern image is shot on the wafer by synchronous scanning with the projection optical system. The present invention is applied to a step-and-scan type projection exposure apparatus for sequentially transferring an image to an area when performing alignment between a reticle and a wafer.

FIG. 1 shows a schematic configuration of a projection exposure apparatus of this embodiment. In FIG. 1, an exposure light source and an exposure light emitted from an illuminance distribution uniforming optical system 1 including a fly-eye lens and the like at the time of exposure are shown. The illumination light IL includes a first relay lens 2,
The light enters a dichroic mirror 6 via a variable field stop (reticle blind) 3, a second relay lens 4, and a condenser lens 5. Then, the illumination light IL reflected downward by the dichroic mirror 6 illuminates a slit-shaped illumination area on the pattern area of the reticle R with a uniform illuminance distribution. In this example, an excimer laser beam such as a KrF excimer laser (wavelength 248 nm) or an ArF excimer laser (wavelength 193 nm) is used as the illumination light IL. However, as the illumination light IL, a metal vapor laser or YA
G laser harmonics or mercury lamp emission (wavelength 365)
nm i-line etc.) can also be used.

The illumination light IL that has passed through the illumination area on the reticle R at the time of exposure is incident on a telecentric projection optical system PL on both sides (or one side on the wafer side), and the pattern in the illumination area is projected by the projection optical system PL. Magnification M (M is a projection magnification under illumination light for exposure, for example, 1/4,
The image reduced by (1/5 or the like) is projected and exposed on a slit-shaped exposure area on the wafer W having a surface coated with a photoresist. Hereinafter, the optical axis A of the projection optical system PL will be described.
The description will be made by taking the Z axis parallel to X, the X axis parallel to the plane of FIG. 1, and the Y axis perpendicular to the plane of FIG. 1 in a plane perpendicular to the Z axis.

In this embodiment, the reticle R is held on a reticle stage 7, and the reticle stage 7 is
The reticle R is positioned two-dimensionally in a plane perpendicular to the optical axis AX of L, and can scan the reticle R at a predetermined speed in the X direction (scanning direction). The moving mirror 8m on the reticle stage 7 and the two-dimensional coordinates of the reticle stage 7 measured by the laser interferometer 8 are as follows:
The reticle stage is supplied to a main control system 9 that controls the operation of the entire apparatus in a centralized manner, based on the supplied coordinates, and the position of the reticle stage 7 via a reticle stage driving unit 10.
And the scanning speed.

On the other hand, the wafer W is mounted on the wafer stage 11 via a micro-rotatable wafer holder (not shown). The wafer stage 11 positions the wafer W in the X direction and the Y direction, and sets the wafer W in the X direction.
Is configured to scan at a constant speed. In addition, the wafer stage 11 is configured such that the wafer W
Is controlled in the Z direction (focus position). The two-dimensional coordinates of the moving mirror 12m on the wafer stage 11 and the wafer stage 11 measured by the external laser interferometer 12 are supplied to the main control system 9, and the main control system 9 transmits the wafer through the wafer stage driving unit 13. The operation of the stage 11 is controlled. During scanning exposure, in synchronism with being scanned at a velocity V _R to the reticle R is the + X direction (or -X direction) via the reticle stage 7, one shot area on the wafer W via the wafer stage 11 speed β · V _R (β is the projection magnification) is scanned in the -X direction (or + X direction).

Next, the alignment sensor of the projection exposure apparatus of this embodiment will be described in detail. In this example, an alignment sensor of a TTR system and a two-beam interference system (LIA system) is used. Then, the reticle R and the wafer W
Each of the shot areas has a diffraction grating alignment mark (reticle mark and wafer mark) formed thereon.

FIG. 2A shows an arrangement of reticle marks on the reticle R. In FIG.
Formed by arranging light-transmitting portions and light-shielding portions at a predetermined pitch in the X-direction in a + Y-direction light-shielding band surrounding the pattern area PA.
A reticle mark 14X in the form of an axial diffraction grating and a reticle mark 14Y in the form of a Y-axis diffraction grating in which light transmitting portions and light shielding portions are arranged at a predetermined pitch in the Y direction are formed.
Also, symmetrically with the pair of reticle marks, the X-axis reticle mark 15 in the form of a diffraction grating is located within the light-shielding band in the −Y direction.
X and Y axis reticle marks 15Y are formed.

Next, FIG. 2B shows an arrangement of wafer marks attached to one shot area SA on the wafer W. In FIG.
X-axis diffraction grating wafer mark 1 in which convex portions and concave portions are arranged at a predetermined pitch in the X direction so as to be adjacent to each other in the X direction.
6X, and a Y-axis diffraction grating-shaped wafer mark 16Y formed by arranging convex portions and concave portions at a predetermined pitch in the Y direction is formed. Further, an X-axis wafer mark 17X and a Y-axis wafer mark 17Y in the form of a diffraction grating are formed symmetrically to the pair of wafer marks and adjacent to the shot area SA in the + Y direction. When performing alignment by the TTR method, the wafer W corresponding to the four reticle marks 14X, 14Y, 15X, 15Y on the reticle R
Wafer marks 16X, 16Y, 1 in shot area SA
The amount of displacement in the measurement direction with respect to 7X and 17Y is detected by the corresponding alignment sensor, and exposure is performed in a state where the results of these detections fall within predetermined allowable ranges.

Furthermore, in this embodiment, the alignment light for detecting the positions of the reticle mark and the wafer mark is different from the light flux in a wavelength range that does not expose the photoresist on the wafer W, that is, the illumination light IL for exposure. Light beams in the wavelength range are used. Therefore, in the projection optical system PL in which aberration correction is performed on the exposure illumination light IL, chromatic aberration (magnification chromatic aberration,
Axial chromatic aberration) may remain. Therefore, in this example, the projection magnification β from the reticle R to the wafer W in the projection optical system PL under the alignment light is measured in advance. When the pitch of the reticle mark 14X in the X direction is Pr and the pitch of the wafer mark 16X in the X direction is Pw, the following two relationships are established using any integer m or n of 1 or more. The pitches Pr and Pw are set as described above.

Pw ≠ m · β · Pr (3) n · Pw ≠ β · Pr (4) That is, the pitch β · Pr of the image of the reticle mark 14X on the wafer W is equal to the pitch Pw of the wafer mark 16X. Set so that they do not have an integral multiple relationship. FIG.
An image 14XW of the reticle mark 14X on the wafer W;
This shows the relationship with the wafer mark 16X. In FIG. 5, the pitch β · P of the image 14XW of the reticle mark 14X
r and the pitch Pw of the wafer mark 16X are expressed by the following equation (3).
If the relationship of the formula (4) is satisfied, the image 1 will be generated when the light beam 32 is virtually incident on these two marks vertically.
The diffraction angle θ _m of the m-th order (m is an integer of 1 or more) diffracted light 34 from 4XW and the first-order diffracted light 3 from the wafer mark 16X
3 does not coincide with the diffraction angle α. That is, since the m-th order diffracted light 34 and the first order diffracted light 33 do not overlap,
As described later, only the ± 1st-order diffracted light from the wafer mark 16X can be detected without being affected by the reticle mark 14X. Similarly, only the ± 1st-order diffracted light from the reticle mark 14X can be detected without being affected by the wafer mark 16X. The pitches between the other reticle marks and the wafer marks are set in a similar relationship, and the principle of detection is the same. Therefore, the configuration of one alignment sensor for detecting the amount of displacement in the X direction between the X-axis reticle mark 14X and the wafer mark 16X will be described below.

Returning to FIG. 1, near the optical Fourier transform plane (pupil plane) with respect to the pattern surface of the reticle R of the projection optical system PL of this embodiment under the alignment light, the projection optics for the alignment light A chromatic aberration control plate 17 for controlling chromatic aberration of the system PL is provided. The chromatic aberration control plate 17 is composed of a plurality of chromatic aberration control elements (17a, 17a,
17b), and the chromatic aberration of the alignment light is controlled by these chromatic aberration control elements, so that the reticle mark 14X and the wafer mark 16 are formed.
X is conjugated to the projection optical system PL under the alignment light.

On the dichroic mirror 6 above the reticle R, an alignment optical system of an X-axis two-beam interference type alignment sensor 21 is arranged. In this alignment optical system, a laser beam L as alignment light (position detection light) emitted from a laser light source 22 composed of, for example, a He-Ne laser light source enters a frequency-variable two-beam generating system 23. This frequency variable 2
The light beam generating system 23 includes two acousto-optical elements, and applies high-frequency signals of variable frequencies f ₁ and f ₂ having a predetermined frequency difference Δf from an external AOM driving system 24 to these two acousto-optical elements. Thus, two light beams (heterodyne beams) L1 and L2 having a predetermined frequency difference Δf and being coherent with each other are generated at a variable emission angle.

FIG. 4 shows the frequency-variable two-beam generating system 23.
In FIG. 4, two acousto-optic elements (hereinafter, abbreviated as “AOM”) 31A and 31B are arranged close to each other. The first AOM 31A shown in FIG.
, The electrode plate 4 is provided on one side surface of the acousto-optic medium 41A.
2A, a transducer 43A for generating ultrasonic waves, and an electrode plate 44A are sequentially fixed, and a sound absorbing material 46A is provided on the other side surface.
Is fixed, and a high-frequency signal having a variable frequency f ₁ is supplied from the AOM drive system 24 between the electrode plates 42A and 44A,
As a result, a traveling wave 47A is formed. Also, the second A
In the OM 31B, an electrode plate 42B and a transducer 43 for generating ultrasonic waves are provided on the other side surface of the acousto-optic medium 41B.
B, and the electrode plates 44B are sequentially fixed, the sound absorbing material 46B on one side is fixed, a variable high-frequency signal of frequency f ₂ is supplied from AOM drive system 24 between the electrode plates 42B and 44B, thereby first Traveling wave 4 in 1 AOM31A
A traveling wave 47B traveling in the opposite direction to 7A is formed.

As the acousto-optic media 41A, 41B,
In addition to ordinary glass, tellurium dioxide (TeO)_Two) Single crystal,
Quartz, quartz, lead molybdate single crystal and the like can be used. Ma
The sound absorbing materials 46A and 46B are acoustic impedances.
(The product of density and sound velocity) is acousto-optic media 41A, 41B
It is a material close to that of and easy to absorb sound waves, for example
A metal film such as lead or aluminum can be used. Also,
As transducers 43A and 43B, lithium
Iovate (LiNbO_Three) Single crystal, LiIO _Three Simple connection
Crystal or Ba_ThreeNaNb_FiveO_FifteenSingle crystal etc. can be used
You.

In this case, a region sandwiched between the transducer 43A and the sound absorbing material 46A forms a first ultrasonic working region, and a region sandwiched between the transducer 43B and the sound absorbing material 46B forms a second ultrasonic working region. Is formed. Then, in the incident laser beam L, AOM3
+ 1st-order diffracted light L ₀ (1) due to traveling wave 47A in 1A and + 1st-order diffracted light L due to traveling wave 47B in AOM 31B
_A The mixed light of (1) becomes a light flux L1 and is-
The mixed light of the first-order diffracted light L ₀ (-1) and the first-order diffracted light L _A (-1) due to the traveling wave 47B forms a light beam L2. The change in the frequency of the light beam L1 is substantially (f ₁ -f ₂ ) /
2 and the change in the frequency of the light beam L2 is substantially-(f
₁ is a -f ₂₎ / _2. Further, the distance s converted into the air length between the center of the ultrasonic action area in the AOM 31A and the center of the ultrasonic action area in the AOM 31B is a photoelectric conversion signal (light beat signal) of the interference light of the two light beams L1 and L2. Are set to maximize the contrast of the image.

In the configuration example of FIG. 4, AO
By changing the frequencies f ₁ and f ₂ of the high-frequency signals supplied from the M drive system 24 to the AOMs 31A and 31B, the emission angles (intersection angles) of the emitted light beams L1 and L2 can be controlled. Accordingly, as described later, the detection target is changed from the wafer mark 16X to the reticle mark 14X in FIG.
When the switching is sequentially performed, the intersection angle of the light beams L1 and L2 can be adjusted according to the pitch of the mark to be detected. However, even when the frequencies f ₁ and f ₂ are changed in such a manner, the frequency difference Δf (= f ₁ −f ₂ ) is kept constant, so that the interference light obtained from the wafer mark 16X or the reticle mark 14X is not changed. The beat frequency is a constant Δf, and the configuration of the signal processing system is simple. The frequencies f ₁ and f ₂ are set to about several tens of MHz in order to drive the acousto-optic element stably, and the frequency difference Δf is set to, for example, about several tens of kHz to facilitate signal processing.

It is desirable that the interval s between the two AOMs 31A and 31B be 0 if possible, but the interval s cannot be 0 when the two AOMs 31A and 31B are arranged close to each other as shown in FIG. Therefore, to make the interval s substantially zero, two A
What is necessary is just to arrange a relay lens system between OM31A, 31B. In order to increase the diffraction efficiency, AOM31A, 3
An acousto-optic device that performs anisotropic Bragg diffraction may be used as 1B.

Returning to FIG. 1, the frequency-variable two-beam generating system 23
The coherent light beams L1 and L2 having a predetermined frequency difference Δf emitted from the light source are condensed by the objective lens 27 and reach the beam splitter 28, and the light beams L1 and L2 reflected by the beam splitter 28 are dichroic. The light passes through the mirror 6 and enters the reticle R so as to intersect the reticle mark 14X on the X-axis on the pattern surface of the reticle R. At this time, if the crossing angle of the light beams L1 and L2 matches the pitch Pr of the reticle mark 14X, interference light composed of ± 1st-order diffracted light is generated vertically upward from the reticle mark 14X, and this interference light is dichroic. Mirror 6 and beam splitter 2
After that, the light enters a photoelectric detector 30 composed of a photodiode or the like. Also, the beam splitter 28 and the photoelectric detector 3
A spatial filter 29 that allows only a light beam generated vertically upward from the reticle mark 14X or the wafer mark 16X to pass therethrough is arranged between the reticle mark 14X and the wafer mark 16X so as to block light beams other than interference light to be detected. .

On the other hand, a part of the light beams L1 and L2 incident on the reticle mark 14X (also called light beams L1 and L2).
Is transmitted through the reticle mark 14X and enters the projection optical system PL. The luminous fluxes L1 and L2 are deflected by the chromatic aberration control elements 17a and 17b on the chromatic aberration control plate 17 installed in the projection optical system PL, and emitted from the projection optical system PL. The light is incident so as to intersect on the axis wafer mark 16X. And
As in the case of the reticle mark 14X, the light flux L1,
If the intersection angle of L2 matches the pitch Pw of the wafer mark 16X, interference light LW composed of ± 1st-order diffracted light is generated vertically upward from the wafer mark 16X, and this interference light LW is a chromatic aberration in the projection optical system PL. Returning to the reticle mark 14X via the corresponding chromatic aberration control element on the control plate 17, the interference light LW transmitted through the reticle mark 14X enters the photoelectric detector 30 via the dichroic mirror 6, the beam splitter 28, and the spatial filter 29. . Photoelectric detector 30
The optical beat signal SX having a frequency Δf obtained by photoelectrically converting the interference light incident on the optical disc is supplied to the signal processing system 26.

As shown in FIG. 4, the distance s between the AOMs 31A and 31B in the frequency-variable two-beam generating system 23 of this embodiment is represented by s.
Does not become 0, the components modulated at the frequencies f ₁ and f ₂ are superimposed on the emitted light fluxes L 1 and L ₂ , and the high frequency component is sometimes superimposed on the obtained optical beat signal SX. . Therefore, the frequency f _1/2 from the optical beat signal SX is applied between the photoelectric detector 30 and the signal processing system 26 in FIG.
A low-pass filter circuit for removing a high frequency component of a degree or more may be added.

The AOM drive system 24 supplies the mixer circuit 2
5 is supplied with high-frequency signals of frequencies f ₁ and f ₂ separated from the high-frequency signal supplied to the frequency-variable two-beam generating system 23, and the mixer circuit 25 mixes these high-frequency signals to obtain a frequency Δf ( = F ₁ −f ₂ ), and supplies the reference beat signal SR to the signal processing system 26. The signal processing system 26 compares the phases of the reference beat signal SR and the optical beat signal SX to determine the phase difference φ of the optical beat signal SX with respect to the reference beat signal SR.

Note that a pulse signal from a simple reference clock generator may be used instead of the reference beat signal SR. In FIG. 1, a laser light source 22 to a photoelectric detector 30
The alignment sensor 21 is composed of the members described above. Further, a laser light source 22, a frequency-variable two-beam generating system 2
3. An alignment optical system is constituted by the objective lens 27 to the photoelectric detector 30.

Next, a description will be given of an operation for detecting the amount of displacement in the X direction between the reticle mark 14X on the reticle R and the wafer mark 16X on the wafer W using the alignment sensor 21 of this embodiment. . First, in order to detect the position of the wafer mark 16X, by adjusting the frequencies f ₁ and f ₂ of the high-frequency signals supplied to the frequency variable two light beam generation system 23, the light beam emitted from the frequency variable two light beam generation system 23 is adjusted. The incident angles θ _{W of} L1 and L2 on the wafer mark 16X are set to satisfy the following expression. Here, λ is the wavelength (center wavelength) of the light fluxes L1 and L2, and Pw is the pitch of the wafer mark 16X. When Expression (5) is satisfied, the intersection angle of the light fluxes L1 and L2 is the pitch of the wafer mark 16X. It is said to conform to Pw.

Sin θ _W = λ / Pw (5) FIG. 3 (a) shows a state in which the crossing angle of the light fluxes L1 and L2 matches the pitch Pw of the wafer mark 16X. 2), when the two light beams L1 and L2 are incident on the reticle mark 14X, a part of the light beams becomes zero-order light beams L10 and L20 and is transmitted to the projection optical system PL. At this time, diffraction at the reticle mark 14X also occurs,
For example, considering the light flux L1, the reticle mark 14
At X, a first-order diffracted light L11, a second-order diffracted light L12, and the like are generated. However, in this example, since the equations (3) and (4) hold, there is no adverse effect of these first-order or higher-order diffracted light.

If the inequality expression (3) or (4) does not hold, that is, if the expression (3) or (4) does not hold, the first-order diffracted light L11 or the second-order diffracted light L1
2 or the like travels in the same direction as the zero-order light L20, and the reticle mark 14X is added to the interference light LW from the wafer mark 16X.
And the first or higher order diffracted light from the optical signal is mixed, and the SN ratio of the detected optical beat signal SX deteriorates. In other words, (3)
As long as the expressions and the expression (4) hold, zero-order light (transmitted light) L10, L2 from the reticle mark 14X of the light beams L1, L2.
0 is a light beam that does not include extra diffracted light, that is, does not include position information of the reticle R. Then, the zero-order light L10, L
When 20 is incident on the wafer mark 16X, ± 1st-order diffracted light is generated in parallel from the wafer mark 16X as interference light LW based on the relationship of the expression (5), and the interference light LW is received by the photoelectric detector 30 in FIG. . On the other hand, for the light LR11, LR21, etc., reflected and diffracted by the reticle mark 14X,
Since there is a relationship between the expressions (3) and (4), there is no combination in the same direction. Therefore, the wafer mark 16X
, The position of the wafer W is accurately detected based on the interference light LW.

Next, when detecting the position of the reticle mark 14X, the frequencies f ₁ and f ₂ of the high-frequency signal supplied to the frequency-variable two-beam generating system 23 in FIG. The incident angles θ _R on the reticle mark 14X of the light beams L1 and L2 emitted from 23 are set to satisfy the following expression. Here, Pr is the pitch of the reticle mark 14X, and when the expression (6) is satisfied, it is said that the intersection angle of the light beams L1 and L2 matches the pitch Pr of the reticle mark 14X.

Sin θ _R = λ / Pr (6) FIG. 3B shows a state in which the crossing angle of the light beams L1 and L2 conforms to the pitch Pr of the reticle mark 14X. ), When the two light beams L1 and L2 enter the reticle mark 14X, the reticle mark 1
Interference light LR consisting of ± 1st-order diffracted light from 4X is generated almost vertically above reticle mark 14X and received by photoelectric detector 30 in FIG. At this time, reticle mark 1
The first-order diffracted lights LW11 and LW21 are also generated from the wafer mark 16X by the 0th-order lights L10 and L20 from the 4X, but since the equations (3) and (4) are satisfied, these first-order diffracted lights LW11 and LW21. Are not parallel to the interference light LR. Therefore, only the interference light LR can be detected with a high SN ratio, and the position of the reticle mark 14X can be detected with high accuracy.

After the detection of the relative position between the reticle R and the wafer W in the X, Y, and θ directions based on the reticle mark and the wafer mark, the main control system 9 determines the positions of the reticle stage 7 and the wafer stage 11. Adjust to complete alignment. After completing the alignment, the main control system 9 outputs a signal for opening a shutter unit (not shown) provided inside the illumination system to open the shutter unit, and relatively scans the reticle R and the wafer W. Then, the process proceeds to an exposure operation for transferring the pattern of the reticle R onto the wafer W.

Further, a specific pitch Pr of the reticle mark 14X for satisfying the relations of the equations (3) and (4),
An example of the pitch Pw of the wafer mark 16X is as follows. Pw = 0.93β · Pr (7) That is, the pitch Pw of the wafer mark 16X is set to 93% of the pitch Pr of the reticle mark 14X corrected by the projection magnification β under the alignment light. Then
In the case of FIG. 3A and the case of FIG.
The difference in the opening angle of L2 almost disappears, and the reticle mark 14
When detecting the positions of X and the wafer mark 16X, the optical paths of the two light beams L1 and L2 can be regarded as substantially the same, and the influence of air fluctuation in the optical path above the reticle R can be minimized.

As described above, in this embodiment, the frequencies f ₁ and f _{1 of the} high-frequency signal supplied to the frequency-variable two-beam generating system 23 shown in FIG.
By switching ₂ , the positions of the wafer mark 16X and the reticle mark 14X are sequentially detected in a time division manner. In this case, the position detection speed of the heterodyne interference method is high, and the period of air fluctuation or mechanical vibration (100 H
(approximately z) can be detected at a period sufficiently faster than that, so that the time required for position detection is short.

Next, in FIG. 1, the reticle mark 14X and the wafer mark 16X are determined based on the optical beat signal SX and the reference beat signal SR output from the photoelectric detector 30.
The method for obtaining the positional deviation amount of the above will be described in detail. As described above, the signal processing system 26 obtains the phase difference φ between the optical beat signal SX and the reference beat signal SR. In this case, assuming that a phase difference φ of the optical beat signal SX detected at the time of detecting the position of the reticle mark 14X is φr, a positional deviation amount ΔXr in the X direction from the reference position of the reticle mark 14X (reticle R) on the reticle stage. Is as follows:

ΔXr = φr · Pr / (2π) (8) Similarly, if the phase difference φ of the optical beat signal SX detected at the time of detecting the position of the wafer mark 16X is φw, the wafer of the wafer mark 16X (wafer W) The displacement ΔXw in the X direction from the reference position on the stage is as follows. ΔXw = φw · Pw / (2π) (9) From the equations (8) and (9), the signal processing system 26 calculates the relative displacement between the reticle R and the wafer W in the X direction. That is, using the projection magnification M under the illumination light for exposure and the projection magnification β under the alignment light,
Since the equation (9) is the displacement amount under the alignment light, the equation (9) is first converted into the displacement amount ΔXw ′ under the illumination light for exposure as follows.

ΔXw ′ = M · ΔXw / β (10) Next, the amount of positional deviation ΔX in the X direction between the reticle mark 14X and the wafer mark 16X on the wafer stage under illumination light for exposure. Is obtained from the following equation. ΔX = −M · ΔXr−ΔXw ′ (11) In Expression (11), it is assumed that the inverted image is transferred by the projection optical system PL. This displacement ΔX is determined by the main control system 9.
Supplied to The misregistration amounts for other marks are also supplied to the main control system 9, and based on these misregistration amounts, the main control system 9 controls the reticle R and the wafer W via the reticle stage drive system 10 and the wafer stage drive system 13. Is corrected.

In the above embodiment, since the beat frequency of the optical beat signal SX when detecting the reticle mark 14X and the wafer mark 16X is the same, signal processing is facilitated. However, the beat frequency of the optical beat signal SX when detecting these two marks does not necessarily need to be the same. That is, equation (8),
Since the beat frequency is not included in the equation (9), the detected positional shift amount is the same even if the beat frequency changes.

In the above-described embodiment, as shown in FIG. 2, the reticle marks for the X-axis and the Y-axis are separated. However, the reticle mark may be a two-dimensional mark in which X-axis and Y-axis marks are mixed. FIG.
FIG. 6A shows a reticle mark 35 composed of a two-dimensional mark. In FIG. 6A, the reticle mark 35 is configured by arranging a square light-shielding portion 35a and a square light transmitting portion 35b in a grid pattern. ing. Light-shielding part 35 in this case
The ratio of the width of a to the width of the light transmitting portion 35b is 1: 1. In this case, the corresponding wafer mark is also composed of a two-dimensional mark.

In the above embodiment, reticle R
The alignment mark 14X on the reticle R and the alignment mark 16X on the wafer W in the projection exposure apparatus that transfers the upper pattern onto the wafer W via the projection optical system PL.
However, as shown in FIG. 7, the present invention employs a proximity method or the like in which a pattern on a reticle R and a wafer W are exposed in close or close contact with each other without using a projection optical system PL. The present invention can be applied to an exposure apparatus.

FIG. 7 shows an enlarged view of an exposure section for exposing the pattern of the reticle R on the wafer W in the proximity type exposure apparatus. In the exposure apparatus shown in FIG. The point different from the above is that there is no projection optical system PL for projecting the pattern on the reticle R onto the wafer W, and the reticle stage 7 and the wafer stage 11 do not move at the time of exposure, and only when the reticle and the wafer are aligned. It is controlled by the main control system 9 so as to move, and other than these points, it is the same as that shown in FIG. In the case of the proximity type exposure apparatus shown in FIG. 7, the projection optical system PL does not exist, but since the reticle R and the wafer W are arranged close to each other, the projection magnification β of the projection optical system PL is reduced. Equations (1) to (4) are used as 1,
Equations (7), (10), and (11) can be applied, and even in this case, the relationships of the above equations (1) to (11) hold.

As the alignment sensor of the example of FIG. 7, the alignment sensor 21 of FIG. 1 can be used as it is, and the reticle mark 14X on the reticle R and the wafer mark 16X on the wafer W can be used as in the example of FIG.
Is detected. Specifically, first, in order to detect the wafer mark 16X, the high-frequency signals f ₁ and f ₂ supplied to the frequency-variable two-beam generating system 23 shown in FIG. Angles θ of the light beams L1 and L2 emitted from the wafer mark 16X on the wafer mark 16X
_W is set so as to satisfy the above equation (5). In this case, as shown in FIG. 7B, two marks L1 and L2 are irradiated on each mark, and interference light LW for detection is generated in the vertical direction from the wafer mark 16X. However, also in this example, since the relationship between the above equations (3) and (4) is established, for example, unnecessary diffracted lights L11 and L12 transmitted and diffracted from the reticle mark 14X with respect to the light beam L1 and reflected from the reticle mark 14X. Unwanted unnecessary diffracted light LR1
1, LR21 is generated in a direction different from the interference light LW for detection, and only the interference light LW for detection reaches the photoelectric detector 30 in the alignment sensor 21 as an alignment system, so that high accuracy is achieved. The position of the appropriate wafer W is detected.

Next, the reticle mark 14X is detected.
To be supplied to the frequency-variable two-beam generating system 23 shown in FIG.
High frequency signal f₁, f_Two And adjust its frequency variable 2 light
Reticle of light fluxes L1 and L2 emitted from flux generation system 23
Incident angle θ on mark 14X _R Satisfies the above equation (6)
It is set as follows. In this case, as shown in FIG.
Each mark is irradiated with two light beams L1 and L2, and a reticle
The interference light LW for detection in the vertical direction from the mark 14X is
Occur. However, also in this example, the above equation (3) is satisfied.
Since the relationship with equation (4) holds, the wafer mark 1
Unnecessary first-order diffracted lights LW11 and LW reflected and diffracted from 6X
21 is generated in a different direction from the interference light LW for detection
Only the interference light LR for detection is an alignment sensor
High-precision reticle to reach the photoelectric detector 30 in 21
The position of the rule R is detected.

After the detection of the relative position between the reticle R and the wafer W in the X, Y, and θ directions is completed by the reticle mark and the wafer mark, the main control system adjusts the positions of the reticle stage 7 and the wafer stage 11. To complete the alignment. After completing the alignment, the main control system 9 outputs a signal for opening a shutter unit (not shown) provided inside the illumination system to open the shutter unit, and transfers the pattern of the reticle R onto the wafer W. To an exposure operation.

However, in this embodiment, since the light beam transmitted through the reticle mark is used for detecting the position of the wafer mark, in order to increase the position detection light at the wafer mark, the light transmitting portion must be larger than the width of the light shielding portion. Should be widened. FIG. 6 (c)
Shows a reticle mark 37 in which the width of the light transmitting portion is wider than the width of the light shielding portion 37a. FIG.
Instead of the reticle mark 14X shown in FIG.
As shown in (b), a one-dimensional grid-like reticle mark 36 in which the width of the light transmitting portion 36b is wider than the width of the light shielding portion 36a may be used.

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.

[0064]

According to the position detecting method of the present invention, the alignment marks (reticle marks) on the mask and the diffracted lights from the alignment marks (wafer marks) on the photosensitive substrate are not overlapped. Since the pitch of the marks is determined, for example, when detecting the positional shift amount between the two marks via the projection optical system using position detection light (alignment light) having a different wavelength from the illumination light for exposure. Also,
There is an advantage that a light beam traveling along substantially the same (coaxial) optical path can be used. Therefore, the displacement of the two marks can be detected through a common simple optical system, and the influence of the fluctuation of the gas above the mask can be minimized, so that the displacement can be accurately detected. be able to.

Further, a projection optical system is used, the pitch of the mark is Pr, the pitch of the alignment mark on the photosensitive substrate is Pw, and the mask of the projection optical system under the position detection light is moved from the photosensitive substrate to the photosensitive substrate. Where β is the projection magnification to
When the pitches Pr and Pw are set using the above-mentioned arbitrary integers m and n so as to satisfy the relations of the equations (1) and (2), the diffracted light from the two marks can be obtained by a simple method. Will not overlap.

Since the present invention can be applied to a proximity type exposure apparatus that does not use a projection optical system, the above advantages can be obtained with a proximity type exposure apparatus. When detecting the phase of the diffracted light from the alignment mark on the mask by the position detection light and the phase of the diffracted light from the alignment mark on the photosensitive substrate in a time-division manner, There is an advantage that the displacement of two marks can be detected by one light receiving unit.

Further, by comparing the phase of the diffracted light from the two alignment marks by the position detection light with the phase of a predetermined reference signal, the amount of displacement between the two alignment marks is obtained. in case of,
The amount of displacement between these two marks from the reference position is detected independently. Therefore, the projection magnification under the illumination light for exposure and the projection magnification under the position detection light (alignment light) in the projection optical system are determined in advance, and the position of the alignment mark on the photosensitive substrate is determined. By converting the amount of displacement into the amount of displacement under the illumination light for exposure, there is an advantage that the amount of displacement under the illumination light for exposure can be accurately detected.

Next, according to the position detecting device of the present invention, the above-described position detecting method of the present invention can be implemented.

[Brief description of the drawings]

FIG. 1 is a schematic configuration diagram illustrating a projection exposure apparatus according to an example of an embodiment of the present invention.

2A is a plan view showing an arrangement of a reticle mark on a reticle R in FIG. 1, and FIG. 2B is an enlarged view showing an arrangement of a wafer mark attached to one shot area on a wafer W in FIG. It is a top view.

FIG. 3 shows a reticle mark 14X and a wafer mark 16X.
FIG. 4 is an explanatory diagram when detecting the position of an image by a time division method.

FIG. 4 is a configuration diagram showing an example of a variable frequency two-beam generating system 23 in FIG. 1;

FIG. 5 is a diagram for explaining a relationship between pitches of a reticle mark and a wafer mark according to an example of an embodiment of the present invention;

FIG. 6 is a diagram showing a modified example of a reticle mark used in an example of an embodiment of the present invention.

FIG. 7 is a diagram showing another example of the embodiment of the present invention.

[Explanation of symbols]

 R Reticle PL Projection optical system W Wafer 9 Main control system 11 Wafer stage 14X, 14Y Reticle mark 16X, 16Y Wafer mark 17 Chromatic aberration control plate 21 Alignment sensor 22 Laser light source 23 Frequency variable two light beam generation system 24 AOM drive system 25 Mixer circuit 26 Signal processing system 29 Spatial filter 30 Photoelectric detector

Claims (6)

[Claims] 1. Prior to transferring a pattern formed on a mask to a photosensitive substrate under illumination light for exposure, a grid-like alignment on the mask is performed using position detection light in a predetermined wavelength range. A position deviation amount between the alignment mark on the photosensitive substrate and the pitch of the alignment mark on the photosensitive substrate. Are set so that the diffracted lights from the two alignment marks by the position detection light do not overlap, and the two alignments obtained when the position detection light is applied to the two alignment marks at the same time. A position shift amount of the two alignment marks is detected based on a phase of the diffracted light from the use mark. 2. The position detecting method according to claim 1, wherein the pattern of the mask is transferred onto the photosensitive substrate via a projection optical system, and the pitch of the alignment mark on the mask is Pr. When the pitch of the alignment marks on the photosensitive substrate is Pw, and the projection magnification from the mask of the projection optical system to the photosensitive substrate under the position detection light is β, an arbitrary integer m of 1 or more. , N, the pitches Pr and Pw are set so that the following two relationships are established. Pw ≠ m · β · Pr and n · Pw ≠ β · Pr 3. The position detecting method according to claim 1, wherein the pitch of the alignment marks on the mask is Pr, and the pitch of the alignment marks on the photosensitive substrate is Pw. A position detection method characterized by setting pitches Pr and Pw using arbitrary integers m and n such that the following two relationships are satisfied. Pw ≠ m · Pr and n · Pw ≠ Pr 4. The position detecting method according to claim 1, 2 or 3, wherein a phase of a diffracted light from the positioning mark on the mask by the position detecting light and a position on the photosensitive substrate. 1. A position detecting method comprising: detecting a phase of a diffracted light from a use mark by time division. 5. The position detection method according to claim 1, 2, 3, or 4, wherein a phase of a diffracted light from the two alignment marks by the position detection light is set to a phase of a predetermined reference signal. A position shift amount of the two alignment marks is obtained by comparing the position shift amount with the position adjustment mark. 6. A grid-like alignment on the mask using position detection light in a predetermined wavelength range prior to transferring a pattern formed on the mask onto a photosensitive substrate under illumination light for exposure. A position detecting device for detecting an amount of misalignment between the mark for positioning and the grid-shaped alignment mark on the photosensitive substrate; a light modulating means for generating two light beams having different frequencies from the light beam for position detection; Driving means for applying ultrasonic waves having a variable frequency to the light modulating means; and positioning the two light beams generated by the light modulating means together with the projection optical system on the mask and the photosensitive substrate. An irradiating optical system for leading to the use mark, and a detector for photoelectrically detecting interference light due to a plurality of diffracted lights generated in the same direction from the two alignment marks, respectively. Position detecting device and adjusts the frequency of the ultrasonic added to the light modulating means from said drive means in accordance with the pitch of the mark to be detected of the positioning marks.