JP2808595B2 - Position detecting apparatus and projection exposure apparatus using the same - Google Patents

Description

Description: BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a projection type exposure apparatus used for LSI pattern transfer and the like.

[Conventional technology]

The reduction projection type exposure apparatus has been improved year by year to enable transfer of a pattern with a higher degree of integration. Recently, a stepper in the duplication UV region using an excimer laser as a light source has been released, and has attracted attention as a potential device as a main device of linography in a line width region of 0.5 μm or less. In particular, an apparatus that narrows the oscillation spectrum width of a KVF excimer laser with a wavelength of 248 nm to less than 5 pm (picometer) and uses only quartz as the projection lens as the optical material has a large exposure area on the wafer. , LS
Since the size of I or a plurality of sizes can be secured, it is expected to be put to practical use as a mass production device at an early stage.

[Problems to be solved by the invention]

When using an all-quartz type projection lens as described above,
Only the light of the spectrum used for printing the pattern corrects the chromatic aberration, and if an alignment system of a so-called through-the-lens is provided via a projection lens, the light of the same spectrum must be used for the printing light. Incidentally, many photoresists in the deep UV region have a strong absorption, and there is a problem that the alignment marks on the wafer cannot be observed with the deep UV light due to absorption of the resist. Also, even when a multilayer resist is used and a resist layer having relatively low absorption is provided on the uppermost layer, the lower layer
There is a problem that UV light is almost absorbed. As an attempt to solve such a problem, on the other hand, the resist at only the alignment mark portion is peeled off by an ablation effect using excimer light, but the problems such as throughput and generation of dust are not solved.

Accordingly, it has been considered that the above-mentioned alignment method in the deep UV stepper mainly uses an off-axis alignment method without using a projection lens. However, the off-axis method has the following problems.

(1) There is a relative offset between the position of the reticle image and the alignment measurement position, which drifts.

(2) Changes in projection magnification and distortion caused by device temperature, atmospheric pressure, mechanical vibration, thermal changes due to partial absorption of exposure light, deformation due to elimination of minute mechanical distortion, changes in focus and field curvature. Alignment cannot be performed.

(3) Even if the position of the reticle is precisely measured, an error occurs in the measurement of the image position on the image plane on the wafer side due to effects such as vibration and distortion of the projection lens itself.

(4) Even if measurement using the fiducial Z dial mark provided on the wafer stage cannot be performed during wafer exposure, high-speed monitoring cannot be performed and the throughput decreases.

In view of the above, the present invention has been made in view of the above-described conventional problems, and is an off-axis alignment system for a pattern on a projection object (second object) such as a wafer. It is a main object to obtain a projection exposure apparatus having the same or higher performance. Furthermore, the present invention
It is an object of the present invention to provide a position detection device and a projection exposure device that can improve the adverse effects caused by thermal disturbance caused by a light source or the like.

[Means to solve the problem]

In order to solve the above-mentioned problem, in a position detection device that detects a relative position shift between a first object and a second object, a light source that emits alignment light, and at least one of the first object and the second object are aligned. An illumination optical system for guiding light is provided, and an optical waveguide is included in the illumination optical system. A projection exposure apparatus for projecting a pattern on a first object onto a second object via a projection optical system includes a light source that emits alignment light and an illumination optical system that guides the alignment light to the second object. A position detection device for detecting a relative displacement between the first object and the second object, and a position control device for adjusting a relative position between the first object and the second object based on the displacement. An optical waveguide is included in the optical system.

[Action]

In the present invention, a heat source such as a light source is separated from a projection lens or the like and guided to an illumination optical system for alignment by an optical fiber or the like. A device can be obtained.

〔Example〕

FIG. 1 is a diagram showing the overall configuration of a projection exposure apparatus according to a first embodiment of the present invention, and assumes an excimer stepper using excimer laser as exposure light. Wavelength 248nm
The KrF excimer laser light source 1 that emits a narrow band laser beam is oscillated by an external trigger. The excimer laser light is supplied to a light quantity control device 2 such as a variable attenuator, a variable aperture blind 3 for limiting a projection area (illumination area on the reticle R), a first condenser lens 4, a reflection mirror 5, and a second condenser. Lens 6
, The reticle R is illuminated with a uniform illuminance distribution. On the lower surface of the reticle R such as quartz, a circuit pattern PA of the LSI, a mark pattern RM for various alignments, and the like are formed with a light-shielding material such as chrome or chromium oxide. The reticle R is horizontally fixed on the reticle stage RST by suction.
Reticle stage RST is driven by reticle R
Is moved two-dimensionally in a plane perpendicular to the optical axis AX of the projection lens system PL. The image of the pattern area PA of the reticle R is image-formed and projected on the wafer W via the projection lens system PL and the optically transparent flat reference member RP. The central portion of the reference member RP together with the projection lens system PL during exposure contributes to the image formation of the pattern area PA of the reticle R on the wafer W. In addition, on the water surface around the pattern projection image of the reference member RP, a portion corresponding to the projection image of the mark pattern RM (reference pattern) of the reticle R is used as a reflection surface, and an image surface (reticle R RM near the surface conjugate with the pattern surface of
Are provided. This reference member RP is made of a material having good dimensional stability such as quartz,
The detailed configuration will be described later. The reference member RP
Is mounted such that no mechanical displacement occurs with respect to the projection lens system PL.

An orixis type alignment sensor WA for observing a mark pattern or the like on the wafer W is provided at a position at a fixed distance from the optical axis AX of the projection lens system PL.
In this embodiment, the alignment sensor WA directly observes the wafer surface without passing through the reference member RP. Further, a light transmission system AF1 that irradiates the wafer W obliquely to the wafer W via the reference member RP, and a light reception system AF2 that receives the reflected light of the light beam on the wafer surface again through the reference member RP are projected. A gap sensor that optically detects the distance between the best imaging plane and the wafer surface by the lens system PL.
Output signal represents the position of the wafer surface in the direction of the optical axis AX, so-called defocus, and is used for an automatic focusing system. By the way, the wafer W is suction-fixed on the wafer holder WH, and the wafer holder WH moves in parallel in the x and y directions in a horizontal plane perpendicular to the plane of the paper, rotates slightly in the θ (rotation) direction, and further in the optical axis AX direction. Is provided on the stage WST that moves Z. The stage WST has a hierarchical structure, for example, and is combined from the bottom in the order of Y stage, X stage, Z stage, θ stage, and wafer holder WH. Each stage is driven by a stage driving unit WSD. In particular, coordinate positions in the xy direction are sequentially measured with high resolution by a laser interferometer 9. One of the laser beams for length measurement from the laser interferometer 9 is perpendicularly incident on the fixed (reference) mirror 7, and the other laser beam is perpendicularly on the moving mirror 8 fixed on the Z stage of the stage WST. The laser interferometer 9 enters the stage WST based on the difference between the optical path length to the fixed mirror 7 and the optical path length to the movable mirror 8.
Measure the position of. Although FIG. 1 shows only a laser interferometer in a one-dimensional direction, a laser interferometer is similarly provided in a direction perpendicular to the paper surface. Fixed mirror 7
Is provided as close as possible to the reference member RP or integrally therewith.

In this embodiment, a so-called optical heterodyne type displacement detection system (alignment system) is employed to detect the mark RM on the reticle R and the mark on the reference member RP with higher accuracy and higher resolution. . The configuration will be described below. An argon ion laser light source including a mirror 11, a laser tube 12, and a mirror 14 is used as illumination light for each mark.
A laser beam 17 is used in which a continuous wave laser beam having a wavelength of 488, 496, or 514 nm from the LS is converted into a higher harmonic by an SHG crystal 13 provided inside the resonator. Laser light
The light 17 is incident on the transparent disk 15 rotated at a constant speed by a motor 16 at a substantially right angle from one surface of the transparent disk 15. On the circumference of the rotating disk 15 including the point of incidence of the laser beam 17, a grating pattern in which minute gratings extending in the radial direction are arranged in the circumferential direction is formed. A part of the laser light 17 that has entered the lattice pattern portion of the disk 15 becomes broken light in the reflection direction and is received by the photoelectric element 18. The photoelectric output of the photoelectric element 18 is used for monitoring the rotational position of the disk 15 and determining the operation timing of the position shift detection. Of the broken light transmitted through the lattice pattern of the rotating disk 15, for example, + 1st-order light and -1st-order light are condensed on the end faces of the optical fibers 23 and 24 by the condensing lenses 19 and 20, respectively. On the other hand, the zero-order light transmitted through the rotating disk 15 is
The light is reflected at 22 and again enters the other part of the rotating disk 15.
If there is a lattice pattern in this part, also ± 1
Next-order diffracted light is generated and condensed on each end face of optical fibers 27 and 28 by condensing lenses 25 and 26, respectively. The optical fibers 23, 24, 27, and 28 preferably have low attenuation for wavelengths in the ultraviolet region, and are preferably single-mode storage fibers, but are not necessarily mode-preservation types. Here, the light guided by the optical fibers 23 and 24 is used for measuring the position of each mark in the x direction, and the light guided by the optical fibers 27 and 28 is used for measuring the position in the y direction. The light emitted from the other end surfaces of the optical fibers 23, 24, 27, 28 is
A local area including the mark RM of the reticle R is irradiated via the aperture plate 30, the telecentric objective lens 31, and the mirror 32. The light transmitted through a part of the transparent part of the mark RM passes through the projection lens system PL and the reference member RP, and the reference member RR.
Reaches a reflection surface formed on a part of the lower surface of the light source, where it is reflected and further reaches a reference pattern. The optical information from the mark RM of the reticle R and the optical information from the reference pattern of the reference member RR are again transmitted to the partial mirror 34 via the mirror 32 and the objective lens 31.
Branch from the illumination light path. Partial mirror 34 is an objective lens
31 are provided at the pupil conjugate position of the projection lens system PL. The optical information reflected by the partial mirror 34 is
It is divided into two directions by a partial mirror 36 by a lens system 35. The partial mirror 36 is a mark RM of the reticle R or a reference member RR with respect to the imaging system of the objective lens 31 and the lens system 35.
Mark RM provided at a position almost conjugate with the reference pattern of
Image of the reference pattern and the image of the reference pattern
Guide towards 42 and 41. Optical information from the reference pattern that has passed through the lens system 41 is received by the photoelectric detector 37, and optical information from the mark RM that has passed through the lens system 42 is received by the photoelectric detector.

Here, the lens systems 41 and 42 are arranged so that the focal plane of the objective lens 32 (that is, the pupil plane of the projection lens system PL) and the light receiving surfaces of the detectors 37 and 38 are substantially conjugate.

Here, when the mark RM of the reticle R and the reference pattern of the reference member RP are diffraction grating patterns, the photoelectric signals of the detectors 37 and 38 are both AC signals having a certain frequency.
These two AC signals have a very important property that the relative displacement between the mark RM and the reference pattern within the grid pitch corresponds to the phase difference of the signals (for example, ± 180 °). Therefore, each detector 37, 38 is measured by the phase meter 39.
The relative positional deviation between the reticle R and the reference member RP is detected by calculating the phase deviation of the output signal from the reticle R. The detected position shift (or phase shift) is input to the main processor MC. The processor MC calculates coordinate values measured by the laser interferometer 9, alignment information from the off-axis type alignment sensor WA, height position information on the wafer surface from the light receiver AF2 of the gap sensor, and information from the photoelectric element 18. Input a timing signal and the like. Furthermore, the processor MC is the drive unit WSD of the stage WST,
The drive information is appropriately output to the drive unit RSD of the reticle stage RST, the excimer laser light source 1, the light amount control device 2, and the like.

As shown by an imaginary line 40 on the right side of the pattern area PA of the reticle R, another set of misalignment detection systems (alignment systems) configured similarly to the members 30 to 39, 41, and 42.
Is provided.

Here, the rotating disk 15 is a so-called radial grating, and the grating pattern moves in a direction traversing the laser light 17 so that the frequency of the positive order light and the negative order light of the diffracted light generated is changed. The difference is given only by quantification.
Therefore, for example, two lights having different frequencies emitted from the optical fibers 23 and 24 and reaching the region of the mark RM of the reticle R via the objective lens 31 interfere with each other to form a mark.
An interference fringe is formed on the RM (or reference pattern), and the interference fringe flows in a direction orthogonal to the fringe at a frequency of a frequency difference (beat). By the way, in the optical heterodyne method, since an interference fringe flows (scans) at a high speed with respect to a mark pattern (diffraction grating pattern), it is not necessary to finely move an object to be detected with respect to an alignment system. There is. This is extremely convenient for detecting a very small displacement error between the reticle R and the projection lens system PL (reference member RP) which does not move its relative position during the exposure operation as shown in FIG. It is.

Further, in the optical heterodyne method, since the phase information of the obtained photoelectric signal corresponds to the positional deviation, only the phase information needs to be detected with high accuracy. Even more advantageously, with current electronic circuit technology, it is very easy to realize a circuit for detecting the phase difference or the like of an AC signal with extremely high resolution.
As a result, 1/100 to 1/100 of the grating pitch of the diffraction grating pattern
A displacement of about 0 is easily measured. A method of detecting the displacement of the lattice pattern on the wafer using the optical heterodyne method is disclosed in detail in, for example, Japanese Patent Application Laid-Open No. 61-215905.

Next, a partial configuration of the apparatus shown in FIG. 1 will be sequentially described in detail.

FIG. 2 is a perspective view showing a configuration from the laser light source LS to the optical fibers 23, 24, 27, and 28, that is, the illumination light generating unit for the displacement detection system shown in FIG. As shown in FIG. 2, lattice pattern portions 51 and 52 are formed on the rotating disk 15 at two locations on the circumference, and the other circumferential portions are transparent portions. A more detailed arrangement of the lattice pattern portions 51 and 52 is shown in the plan view of FIG. Rotating disk
Approximately 90 ° point-symmetrically with respect to the rotation center Og in the direction of the arrow 53 of 15
The grating pattern portions 51 and 52 are formed by the angle of.
The zero-order transmitted light of the laser light 17 becomes the laser light 66 which is again incident on the other part of the rotating disk 15 by the mirrors 21 and 22. Incident point and point of incidence of the laser beam 66 of laser light 17 as shown in FIG. 4 are assumed to be open at both equal distance from the rotational center Og, and the center of rotation Og angle theta _1. Since the angle θ ₁ has two point patterns symmetrically with respect to the grid,
Here, it is set to a constant value between 90 ° and 180 °.
Now, in FIG. 2, the grid pattern portion 51 of the rotating disc 15
52, the transmission unit at a constant pitch d ₁ along a concentric circle around the rotation center Og and the impermeable portion. The rotating disk 15 is rotated by a motor 16 in the direction of arrow 53 in a speed m (rp
When rotated at a constant speed in s), the distance (radius) between the incident point of the laser beam 17 (or 66) and the center Og is defined as r (mm), and the first-order diffracted light + LB and -LB from the grating pattern portions 51 and 52 are converted into Causes a frequency shift expressed by equation (1).

Therefore, the frequency difference Δ∫ ₀ between the +1 order light and the −order light is expressed by the equation (2).

Therefore, the optical fiber 24 (or 28) shown in FIG.
Frequency of the incident light will be higher by Deruta∫ ₀ than the frequency of light incident to the optical fiber 23 (or 27) to.

As described above, the rotating disk 15 has the function of a frequency shifter that varies the frequency between positive and negative first-order diffracted light, and
Grating pattern portions 51 and 52 and transparent portions are provided alternately along the circumference, and the laser light is reflected by mirrors 21 and 22 to form a lens.
It has a switching function of alternately repeating generation of diffracted light incident on lenses 19 and 20 and generation of diffracted light incident on lenses 25 and 26 with time.

By the way, in FIG. 1, four optical fibers 23, 24, 27, 28
Although only one set is shown, if there are a plurality of displacement detection systems, each lens 19, 20, 25, 26
May be distributed to a plurality of optical fibers. Each of the distributors MB1, MB2, MB3, and MB4 shown in FIG. 2 is configured to distribute the condensed light to four optical fibers with substantially the same amount of light.

FIG. 3 is a perspective view showing the internal configuration of distributors MB1 and MB2 as a representative, and distributors MB3 and MB4 are similarly configured. The +1 order light condensed by the lens 19 is split into two by a beam splitter 54, and the light transmitted therethrough is further split into two by a beam splitter 55, and enters the optical fibers 23 and 56, respectively. On the other hand, the light reflected by the beam splitter 54 is further split into two by the beam splitter 57, and enters the optical fibers 58 and 59, respectively. Similarly for the distributor MB2, the beam splitters 60, 62, 63
The primary light is split and made incident on each of the optical fibers 24, 61, 64 and 65. Similarly, as shown in FIG.
The primary light is split into four optical fibers 27, 71, 73 and 74, respectively, and the distributor MB4 splits the primary light into four optical fibers 28 and 7 respectively.
Divide into 7, 79 and 80 respectively. As described above, the optical fiber used here is preferably of a type that can store a single mode, but it is preferable that the polarization plane is not stored. If the polarization plane is preserved, it is necessary to align the polarization plane by inserting a polarizing element in the optical path from the exit end of the optical fiber to the displacement detection system, or by rotating and positioning the optical fiber.

Next, the configuration of the displacement detection system (alignment system) shown in FIG. 1 will be described with reference to FIG. Fifth
The figure is a perspective view schematically showing a system from the emission ends of the optical fibers 23, 24, 27, and 28 shown in FIG. 1 to the reference pattern RF formed on the reference member RP.

The mark RM of the reticle R has an X reference pattern PX having diffraction gratings arranged in the x direction and a Y reference pattern PY having diffraction gratings arranged in the y direction. The marks RM are adjacent to the patterns PX and PY and have a reference pattern RF. Transparent window at corresponding position
PW is provided. The reference patterns PX and PY and the window PW are surrounded by a light-shielding portion having a predetermined size. The mark RM and the reference pattern RF are conjugate to each other under the wavelength of light emitted from the optical fiber. Now, the light from each of the optical fibers 23 and 24 passes through each of the small openings 43 and 44 of the aperture plate 30, and passes through the window filter 47 in which the partial mirror 34 is formed at the center of the transparent plate obliquely provided in the optical path. And enters the objective lens 31. The minute openings 43 and 44 of the aperture plate 30 are provided at positions symmetric in the x direction with respect to the center of the aperture plate 30 (the point where the optical axis of the objective lens 31 passes). Further, the aperture plate 30 is provided with minute openings 45 and 46 for transmitting light from the optical fibers 27 and 28, respectively, and the openings 45 and 46 are formed at positions symmetrical in the y direction with respect to the center of the aperture plate 30. . This aperture plate 30 is preferably provided near the position of the focal point (f) of the objective lens 31. The partial mirror 34 is the objective lens 31
Is formed at a position where the optical axis of the objective lens 31 passes, that is, at the position of the focal point (f) of the objective lens 31, that is, at the position conjugate with the pupil of the projection lens system PL or in the vicinity thereof. Now, of the light that enters the objective lens 31 and travels toward the reticle R, the light (+ 1st-order light from the optical fiber 23) 90a that has passed through the opening 43 of the aperture plate 30
And the light 90b passing through the opening 44 (-1st-order light from the optical fiber 24) is defined by the area of the mark RM along the principal ray inclined at substantially the same angle in the x direction about the optical axis of the objective lens 31. Irradiate. Similarly, the + 1st-order light and the -1st-order light passing through each of the openings 45 and 46 of the aperture plate 30 are also marked RM along the principal ray inclined at an angle in the y-direction about the optical axis of the objective lens 31. Is illuminated. However, due to the switching function of the rotating disk 15 shown in FIGS. 2 and 4, ± primary light passing through the openings 43 and 44 and ± primary light passing through the openings 45 and 46 alternately form the area of the mark RM. Irradiate.

The minute apertures 43, 44, 45, and 46 of the aperture plate 30 have a function of improving the directivity and coherence of light from the optical fiber and allowing the light to enter the objective lens 31, that is, a function of increasing spatial coherency. . So, below, mark
The measurement of the displacement of the RM and the reference pattern RF in the x direction will be described.

The two lights 90a and 90b from the objective lens 31 are
Are very small, so that they are condensed into almost parallel light beams, respectively, and since the frequencies of the lights 90a and 90b are different by Δf ₀ , x is applied to the irradiation area including the mark RM of the reticle R.
A bright and dark interference fringe flowing in the direction (the fringe itself extends in the y direction) is generated. At this time, the pitch of the interference fringes on the mark RM is set to be exactly 1/2 of the grating pitch of the X reference pattern PX. The pitch of the interference fringes is determined by the incident angles and wavelengths of the light 90a and 90b. Here, assuming that the grid pitch of the X reference pattern PX is 2d ₂ (mm), the moving speed of the interference fringes in the x direction is represented by d ₂ Δf ₀ (mm / Sec). Then, assuming that the incident angles of the light beams 90a and 90b are both θr and the wavelength is λ, the incident angle θr is expressed by Expression (3).

When the relationship of equation (3) is established, the X reference pattern
Objective 31 in the direction of the normal line (optical axis of objective 31) from PX
The interference light 90c generated toward the optical path includes a light beam signal having a frequency Δf ₀ as long as the relative positional relationship between the reticle R and another optical system does not change.

On the other hand, the interference fringes generated in the window portion PW in the mark RM are reduced (or equal in size) at a reduction ratio 1 / M of the projection lens system PL, and become interference fringes with a pitch d ₂ / M on the reference pattern RF. The image is projected. Therefore, here, the reference pattern RF has a pitch of 2.
In the case of a lattice pattern in which lattice elements are arranged at d ₂ / M, interference light from the reference pattern RF also occurs parallel to the projection lens system PL optical axis, passes through the window PW of the reticle R, and forms interference light 90c. It goes to the objective lens 31 in the same optical path. It is assumed that the wafer W (reference pattern RF) side of the projection lens system PL is a telecentric system. When the reticle R side of the projection lens system PL is also a telecentric system, the optical axis of the objective lens 31 is It is determined parallel to the optical axis AX of the projection lens system PL. When the reticle R side is a non-telecentric system, the case where the optical axis of the objective lens 31 is parallel to the optical axis AX and the case where the optical axis is aligned with the off-axis inclined principal ray of the projection lens system PL can be considered.

As described above, the interference light 90c returning from the X reference pattern PX and the interference light returning from the reference pattern RF both reach the partial mirror 34 of the spatial filter 47 via the objective lens 31. Since the objective lens 31 is a telecentric system and the wafer W side of the projection lens system PL is a telecentric system, each interference light generated in parallel with the optical axis is condensed in the partial mirror 34, that is, in the center of the pupil. Are reflected here and enter the second objective lens 35.

Each image of the X reference pattern PX and the reference pattern RF is formed on the partial mirror 36 by the interference light passing through the second objective lens 35. The partial mirror 36 has a reflecting portion 48 for reflecting only the image of the reference pattern RF on a transparent plate obliquely provided in the optical path, and is configured to transmit the images of the X reference pattern PX and the Y reference pattern PY. You. The interference light from the reference pattern RF reflected by the reflection unit 48 is received by the detector 37 via the lens 41 (see FIG. 1), and the interference light from the reference patterns PX and PY transmitted around the reflection unit 48. Is received by the detector 38 via the lens 42 (see FIG. 1). An L-shaped band-shaped absorbing portion 39 is formed around the reflecting portion 48 of the partial mirror 36 shown in FIG. 5 to completely separate the image of the reference pattern RF from the image of the reference patterns PX and PY. It is assumed.

As described above, the X reference pattern PX and the reference pattern
When detecting the relative positional shift of the RF in the x direction, the phase shift of the output signal of the detector 37 with respect to the output signal of the detector 38 is measured by the phase meter 39. For example, when the phase shift obtained by the phase meter 39 is φx, when this phase shift φx is converted into a positional shift amount Δx on the wafer surface, It is considered that the reference pattern RF is shifted in the x direction by Δx.

 The same is true for the y direction.

Next, an example of the detailed shape of the reticle mark RM and the reference pattern RF will be described with reference to FIGS. 6, 7A and 7B. FIG. 6 shows a reticle mark RM whose periphery is surrounded by a light-shielding portion (shaded portion), and the X reference pattern PX is y
Slit-shaped transparent grid elements 1041, 1042, extending in the direction
1043,1044,1045 are arranged in the x direction at a pitch 2d _2.
In the Y reference pattern PY, slit-like transparent lattice elements 1031, 1032, 1033, 1034... Extending in the x direction are arranged in the y direction at a pitch 2d ₂ . The grid duty of the reference patterns PX and PY is
50%. The window portion PW is formed adjacent to the upper part (X direction) of the X reference pattern PX and adjacent to the right side (X direction) of the Y reference pattern PY. FIG. 7 (a) shows the shape of a reference pattern RF formed to have a size that just fits into the window PW. Here, a square opaque grid element 1051, in the x and y directions at a pitch of 2d ₂ / M, 1052 ... in which are arranged in a matrix at 50% duty, and acts as a two-dimensional diffraction grating, Figure a. 7 (b) is a modification of the reference pattern RF shown in Figure No. 7 (a), the pitch 2d ₂ / M opaque grid element 1
. Are arranged in a checkered pattern in the x and y directions, and similarly function as a two-dimensional diffraction grating.

The interference fringes formed by the light emitted from the objective lens 31 with respect to the reticle mark RM shown in FIG.
It is formed so as to cover the entire RM. However, rotating disk 15
With the switching function described above, the interference fringes extending in the x direction and the interference fringes extending in the y direction irradiate the mark RM without temporally overlapping.

Next, the configuration of the reference member RP shown in FIG. 1 will be described with reference to FIG.

The reference member RP has a structure in which two circular quartz plates RPa and RPb are overlapped, and is disposed so that the optical axis AX of the projection lens system PL passes through the center thereof. In FIG. 8, the center point Os
And two lines lx and ly orthogonal to the optical axis AX are x in the coordinate system xy, respectively.
This is a line parallel to the axis and the y-axis and included on the upper surface of the quartz plate RPa. A circular area IF concentric with the optical axis AX is an imaging field (image field) of the projection lens system PL on the upper surface of the quartz plate RPa, and is a rectangular area circularly tangent to the area IF.
PAm is a portion through which the image light flux of the pattern area PA of the reticle R passes.

Here, a case where reference patterns RFa and RFb are provided at two positions symmetrically separated in the x direction with respect to the optical axis AX of the reference member RP will be described. The reference patterns RFa and RFb are formed on the surface between the quartz plates RPa and RPb, and the perpendicular lines La and Lb passing through each of the reference patterns RFa and RFb and parallel to the optical axis AX are:
It coincides with the optical axis of the objective lens 31 shown in FIG. 1 and FIG. Each of the reference patterns RFa and RFb is provided outside the rectangular pattern image area PAm and inside the image field IF where the aberration has been well corrected. Vertical lines La, Lb
The upper points Oa and Ob are points on the line lx, that is, points included in the upper surface of the quartz plate RPa, and the distance in the x direction from the center Os to each of the points Oa and Ob is equal in the present embodiment. I do.

Now, on the lower surface of the quartz plate RPb (the surface facing the wafer W) below the reference patterns RFa and RFb, reflection portions Ma and Mb are formed by vapor deposition or the like. The reflection portions Ma and Mb are formed with an area larger than the size of the reference patterns RFa and RFb so that the perpendicular lines La and La pass through the centers thereof. However, the pattern image area
The setting is made so as not to block the imaging light flux at the time of exposure through PAm.

As is clear from the above configuration, the reference pattern RF
a, RFb and the reflection parts Ma, Mb are separated by the thickness of the quartz plate RPb in the optical axis AX direction.

FIG. 9 shows a projection lens system PL, a reference member RP, and a wafer W.
Of a cross-section showing the arrangement of the optical axis AX direction, the thickness of the quartz plate RPb and G _2, the distance between the lower surface and the wafer W of the quartz plate RPb with G _1,
The light beam Lex along the principal ray lm parallel to the optical axis AX is the reticle R
Represents an imaging light beam from one outermost point of the pattern area PA, and a light beam Lfr along the principal ray La (the optical axis of the objective lens 31).
Represents an image-forming ray from one point (on the optical axis of the objective lens 31) within the mark RM of the reticle R. If there is no reference member RP, the conjugate plane of the reticle R with respect to the projection lens system PL due to the illumination of the excimer laser light can be formed as the plane FP in FIG. However, a conjugate plane is formed a predetermined amount below the plane FP from the overall thickness of the reference member RP, and the optical refractive index n,
Here, the surface of the wafer W is fitted.

By the way, if the wavelength of the light illuminating the mark RM of the reticle R is very close (or coincident) with the wavelength of the excimer laser light, the image point of the light beam Lfr is assumed to have no reference member RP and the surface FP
Can be on the same side as However, since the reflection part Ma exists on the lower surface of the reference member RP, the light beam Lfr is reflected upward here and forms an image on the surface of the reference pattern RFa. The same applies to the operation of the other reference pattern RFb and the reflection portion Mb.

Here, considering the image forming relationship between the reference pattern RFa and the reticle mark RM, the depth of focus of the projection lens system PL is generally very shallow at ± 2 μm or less. Therefore, the reference pattern RFa
Is observed directly from the reticle R side only through the quartz plate RPa, the image of the reference pattern RFa is extremely largely defocused. For this reason, at the position of the partial mirror 36 shown in FIGS. 1 and 5, only the image of the reference pattern RFa via the reflecting portion Ma of the reference member RP is formed.

Next, the operation of this embodiment will be described with reference to FIG. 10 and FIG. FIG. 10 shows a state of a mark portion of the reticle R. Here, an X reference pattern PX and a window portion PW are shown. As shown in FIG. 5, the reticle mark RM is simultaneously irradiated with light beams 90a and 90b having an incident angle θR from two directions, and a part thereof is reflected and diffracted by the X reference pattern PX and interfering light 90 perpendicular to the reticle R
Become C. The luminous flux 90a, 90b transmitted through the window PW as it is
Is a reference pattern as shown in FIG. 11 via the projection lens system PL.
Irradiate RFa from the surface. Reference pattern R for luminous flux 90a, 90b
The incident angle θW to Fa is n, the optical refractive index of the reference member RP,
If the reduction ratio of the projection lens system PL is 1 / M, it is given by the following equation.

When this condition is satisfied, the diffracted light (for example, +1 light and −1 order light) generated by each of the light beams 90a and 90b diffracted by the reference pattern RFa becomes interference light 90d traveling in the perpendicular direction to the projection lens system PL. Then, the light enters the objective lens 31 through the window PW in FIG. Therefore, at the position of the partial mirror 36 shown in FIG. 5, the reference pattern image (dark field image) formed by the interference light 90C from the reference pattern PX (PY) and the interference light 90c from the reference pattern RFa. The formed reference pattern image (dark field image) forms an image. However, the interference light beams 90c and 90d have the luminous fluxes 90a and 90b as described above.
Brightness and darkness are repeated at a beat frequency having a frequency difference Δf ₀ . Then, the interference lights 90c and 9 separated by the partial mirror 36
The phase difference of 0d is measured by the phase meter 39. The phase difference detected by the phase meter 39 is sent to the processor MC shown in FIG. The processor MC inputs the photoelectric output from the photoelectric element 18 as a synchronization signal at the time of measurement, and discriminates whether the phase difference measured by the phase meter 39 is the x direction or the y direction. 2 and FIG. 5 and the mark shapes of FIGS. 6 and 7, the photoelectric element 18 generates an output signal when the optical fibers 23 and 24 emit laser light. Yes, this is when measuring the displacement in the x direction. So processor MC
Calculates the phase difference when the photoelectric element 18 is generating an output signal as the shift amount in the x direction, and calculates the phase difference when there is no output signal as the shift amount in the y direction.

Here, assuming that the reticle R is accurately positioned with respect to the optical axis of the projection lens system PL by a reticle alignment system (not shown), two marks RM on the reticle R measured by the phase meter 39 are , 2 reference patterns
The displacement of each of RFa and RFb in the x and y directions becomes almost zero.

However, since a slight positioning error or the like remains in practice, the amount of displacement is measured at the position of each mark.

For example, marks RMa and RMb are provided on the left and right sides of the reticle R as in the reticle mark RM shown in FIG. 12, and the amount of displacement at each mark position is (ΔXa, ΔYa), (Δ
Xb, ΔYb). Therefore, the deviation amount Δ
Xa and ΔXb are compared, and the direction of the shift is the same (positive x,
Or one of the negative directions), the processor MC finely moves the reticle stage RST in the opposite direction. When the signs of the deviations ΔXa and ΔXb are opposite to each other, reticle stage RS is set so that | ΔXa | = | ΔXb |
Drive in the position of T. The same applies to the positioning in the y direction, but the reticle R
Reticle stage RST if the rotation amount is larger than the allowable value.
Is driven to rotate so that the rotational deviation becomes zero.
As described above, the reference patterns RFa, RFb of the reticle R
Is constantly detected and the reticle R is misaligned due to some cause (vibration during stepping, etc.), the servo control of the reticle stage RST is activated, and the reticle R and the projection lens system PL are integrated as if they were integrated. Behave like.

At the time of actual exposure, the distance between the pattern center (or reticle mark RM) of the reticle R and the detection center of the off-axis type alignment sensor WA is accurately obtained.
After storing the interval value as a reference, the wafer W is sequentially stepped in a step-and-repeat manner based on the position where the alignment sensor WA detects the mark on the wafer W and the stored interval value. During this stepping exposure, the marks RMa and RMb of the reticle R are in the pattern area.
Since it is slightly away from the periphery of the PA, the mirror 32 at the tip of the displacement measuring system shown in FIG. 1 does not block the exposure light. Therefore, the shift between the reticle R and the reference patterns RFa and RFb is constantly monitored even during the exposure operation.

As described above, in the present embodiment, the phase difference detected by the phase meter 39 is
It corresponds to the difference between the reticle mark RM and the reference pattern RF, but the mark RM is based on the grid pattern of the rotating disk 15.
And the reference pattern RF may be obtained. In this case, the grid pattern 5 of the rotating disk 15 shown in FIG.
A grating pattern (radial grating) having the same pitch is provided on the entire circumference, such as inside 1, 52, and this grating pattern is also irradiated with a partial beam branched from the laser beam 17. Then, the + 1st-order light and the -1st-order light generated from this lattice pattern interfere with each other to optically generate a reference beat signal, or after the + 1st-order light and the -1st-order light are separately photoelectrically detected, the reference beat signal is electrically detected. Get a signal or create a reference AC signal. The phase difference between the reference AC signal and the output signal from each of the detectors 37 and 38 is determined. In this case, two phase meters 39 are required for the reticle mark RM and the reference pattern RF, and a circuit for comparing the obtained two phase differences is also required.

Although the reference member RP is formed by laminating two quartz plates RPa and RPb, the reference member RP may include only one quartz plate RFb in which the reference pattern RF and the reflection portion M are formed on each of the front and back sides.

Also when previously the difference is known about the wavelength of the illumination light for positional displacement measurement and excimer laser light for exposure (90a, 90b, etc.), the gap G between the thickness G ₂ and the wafer W of the quartz plate RPb
By appropriately adjusting _{1 and the} like, the conjugate relationship between the reticle mark RM and the reference pattern RF and the conjugate relationship between the reticle R and the wafer W can be satisfied at the same time. If the reference member RP is moved up and down precisely in the direction of the optical axis AX only, the focal position of the projected image of the pattern PA of the reticle R is not changed at all, and only the reticle mark RM and the reference pattern RF are used. Focusing is possible.

Next, a second embodiment of the present invention will be described with reference to FIGS. 13, 14, and 15. FIG. FIG. 13 is a plan view of the reference member RP, and FIG. 14 is a plan view of the reticle R corresponding thereto. As shown in FIG. 13, the reference member RP has
Similarly to the drawing, reference patterns RFa and RFb are provided, and below them, reflection portions Ma and Mb are formed. Further, in this embodiment,
Reference patterns RFc and RFd of the same shape and dimensions also on the y-axis,
Reflection parts Mc and Md are formed. These reference patterns RFa, R
Fb, RFc, and RFd are all provided outside the maximum pattern transfer area PAm in the image field IF and in a portion where aberration is well corrected. In FIG. 14, marks (reference patterns) corresponding to the four reference patterns are respectively provided.
RMa, RMb, RMc, and RMd are formed in the light shielding portions 132, 134, 136, and 138. Further, on the outer peripheral portion of the pattern area PA, a light-shielding band 140 having a width determined by the blur of the opening edge of the reticle blind 3 and the like is set. Further, reticle R
At two corners outside the image field IF, cross-shaped reticle pre-alignment marks 131a and 131b are provided. These pre-alignment marks 131a and 131b are used for positioning with reference marks provided in the apparatus.For example, the marks 131a and 131b and the reference marks are detected using a photoelectric microscope or the like to position the reticle stage RST. Then, pre-alignment can be performed at about ± 1 μm. After this pre-alignment, fine positioning of the reference pattern RF and the reticle mark RM is performed.

FIG. 15 is a perspective view showing an arrangement of a displacement detection system suitable for performing the fine positioning and the displacement measurement. As shown in FIG. 15, in the present embodiment, the displacement detection systems DMSa, DMSb, and DM as shown in FIG.
Sc and DMSd are provided, and a two-dimensional displacement between the reference pattern RF and the mark RM is measured at each position. In FIG. 15, PA 'is a projection image of the pattern area PA of the reticle R, and Ep represents a pupil of the projection lens system PL. By providing the same positional deviation detection system at four locations, it is possible to monitor the imaging characteristics from the reticle R to the reference pattern RF.
Representative of the imaging characteristics are a change in projection magnification, a distortion characteristic, a change in inclination of the reticle R from a horizontal plane, and the like.

Conventionally, many attempts have been made to measure this type of imaging characteristic, but none of these methods can be performed in real time during the exposure operation. However, according to the present embodiment, even during the exposure operation, since the image forming characteristics can be monitored in addition to the sequential displacement of the reticle R, a more accurate overlay exposure can be performed.

Generally, a change in magnification and a change in distortion of the projection lens system PL are extremely small, and a measurement accuracy of submicron is required. In the second embodiment, as in the first embodiment, if each of the four positional shifts is measured by employing the double diffraction method based on the optical heterodyne method, the measurement resolution at each of the four locations can be easily determined. Since about 0.01 μm can be obtained, the change in magnification and the change in distortion can be determined very accurately.

Now, when a change in magnification or the like is detected in real time by such a measurement system, it can be applied as a control signal of a correction device that drives the change in magnification or the like to a predetermined value. FIG. 16 shows a schematic configuration of an optical characteristic correcting device according to a third embodiment of the present invention. In FIG. 16, two-dimensional displacement information from the four displacement detection systems DMSa, DMSb, DMSc, and DMSd is input to a characteristic variation calculation unit 200, where a desired characteristic change amount is calculated. When the magnification error is calculated as the characteristic change amount, the magnification error information 201b is sent to the pressure control unit 202 or the field lens driving unit 206. The pressure control unit 202 controls the pressure of the sealed air chamber (single or plural) in the projection lens system PL by a predetermined amount, and performs feedback control so that the magnification error becomes zero (or a constant offset value). In the case of the drive unit 206, the magnification error is corrected by moving the field lens unit 212 on the reticle R side of the projection lens system PL up and down by a predetermined amount.

In the case caused focus changes with the correction of magnification error, by introducing the offset information 201a corresponding to the focus change in the focus driving unit 204 for vertically moving the holder WH of the wafer W, the focus signal from the light receiving system AF ₂ Can be given an offset. Further, when the reticle R side is a non-telecentric projection lens system, magnification correction can be performed by moving the reticle R up and down according to the error information 201b. Further, the inclination of the reticle R can also be obtained by selecting the tendency of the displacement of the four displacement measurement points in the x direction or the y direction. In this case, a leveling mechanism may be provided on reticle stage RST, and the leveling mechanism may be driven so that the inclination becomes zero. Alternatively, when the wafer stage WST is provided with a leveling mechanism, the inclination of the projection image plane is predicted according to the measured direction and amount of the inclination of the reticle R, and the surface of the wafer W is parallel to the projection image plane. May be tilted by a leveling mechanism so that

Next, a fourth embodiment of the present invention will be described with reference to FIGS. In each of the embodiments described so far, the relative alignment between the reticle R and the projection lens system PL1 (reference member RP) is performed, but the alignment with the wafer W is impossible. Therefore, in the present embodiment, a part of the reference member RP having the reference pattern RF serving as a reference for the alignment of the reticle mark RM is provided with a second reference pattern serving as a reference for the alignment of the alignment mark formed on the wafer W. Was provided. FIG. 17 shows a cross section of the reference member RP, FIG. 18 shows a plane, and the quartz plates RPa, RPb, the reference patterns RFa, RFc and the reflection portions Ma, Mc, etc. described in each of the above embodiments are the same. It is provided exactly as in the case of FIG. However, the reflection portions Ma and Mc (same for Mb and Md) are formed to extend outside the image field IF. An opening 178 is formed at a position outside the outer diameter of the lens barrel of the projection lens system PL on the x-axis of the reflection part Ma (Mb), and the reflection part Mc (Md) is formed.
An opening 184 is formed on the y-axis at a position outside the outer diameter of the lens barrel. Further, a second reference pattern (second reference lattice pattern) RGa is formed at a position at a fixed distance from the reference pattern RFa in the x direction so as to sandwich the opening 178, and at a position at a fixed distance from the reference pattern RFc in the y direction. Second reference lattice pattern RGc is formed so as to sandwich opening 184.
The reference grid patterns RGa and RGc are provided on the same plane as the reference patterns RFa and RFc, and the grid pattern RGa is
, And the grid pattern RGc is used for positioning the wafer W in the y direction. Although the lattice patterns RGa and RGc are divided into two parts with the openings 178 and 184 interposed therebetween, only the center part of the lattice arrangement having the same pitch is originally made transparent so as to match the openings 178 and 184.

Now, on the wafer W, the alignment mark WMa
It is formed as shown in the figure. The mark WMa is provided for alignment of the wafer W in the x direction, and a mark for alignment in the y direction is provided similarly. Mark WMa
Is composed of a diffraction grating pattern in which grating elements are arranged at a constant pitch in the x direction. When aligning (aligning) wafer W with reference member RP, wafer stage WST is positioned such that mark WMa is located immediately below opening 178 of reflection portion Ma. A system for detecting the mark WMa includes a He-Ne laser light source 161 for outputting a laser beam having a wavelength insensitive to the resist on the wafer surface, and a lens 16 for spot-focusing the laser beam on the wafer W with a predetermined diameter.
2. Beam splitter 1 for splitting laser beam in two directions
63, a prism 166 having two reflecting surfaces 166a and 166b, a mirror 175 for extracting optical information from the mark WMa, a lens 176 for condensing the optical information to a detector 177, and the like.
Incidentally, since the laser light source 161 and the like may cause thermal disturbance, the laser light source 161 and the like are separated from the reference member RP and the projection lens system PL and the like as shown in FIGS. It is desirable to guide the He-Ne laser light to the lens 162.

In the above configuration, one laser beam 167 reflected by the beam splitter 163 travels in the reference member RP in parallel with the optical axis AX, and vertically irradiates one of the two grating portions of the reference grating pattern RGa. Further, the other laser beam 168 transmitted through the beam splitter 163 is reflected by the reflecting surface 166a of the prism 166,
After being reflected in the order of 166b, the light travels in the reference member RP in parallel with the optical axis AX, and irradiates the other grating portion of the grating pattern RGa vertically. Here, from the beam splitter 163, the grating pattern RGa
Must be shorter than the coherence length of the He-Ne laser beam.

When the two beams 167 and 168 irradiate the grating pattern RGa in this manner, ± first-order diffracted lights 171 and 172 and a zero-order light that pass through the grating pattern RGa are generated. The 0-order light is shielded by the reflector Ma, and the ± first-order lights 171 and 172 pass through the opening 178 and reach the wafer W, where they interfere with each other to form interference fringes in which bright and dark fringes are arranged at a constant pitch in the x direction. Occurs. In this embodiment, 2
Since the frequency difference between the beams 167 and 168 is zero, the interference fringes generated on the wafer W are stationary. As in the previous embodiment, the pitch of the interference fringes is determined to be exactly 1/2 of the grating pitch of the mark WMa of the wafer W.

When the grating pattern of the mark WMa and the interference fringe overlap, the interference light 174 is generated almost vertically with an intensity corresponding to the positional relationship between the grating of the mark WMa and the interference fringe in the x direction. This interference light 174
When the mark WMa is moved in the x-direction relatively to the interference fringes radiated on the wafer, a substantially sinusoidal intensity change is produced. Then, a change in light intensity is converted into a photoelectric signal by the detector 177, and the wafer W is positioned in the x direction so that the signal has a certain level.
Is achieved in the x direction with respect to the reference grating pattern RGa. Also in the y direction, the reference grid pattern RG
Using c and a mark on the wafer W, alignment is performed using a similar mark detection system.

Alternatively, the level change of the photoelectric signal from the detector 177 is
Laser interferometer 9 for measuring the position of wafer stage WST 9
Sampling is performed in synchronization with the measurement pulse (or position information) and the like, and the sampled waveform is analyzed, so that only the scanning of the wafer stage WST in the x direction, that is, without stopping the stage WST, the reference of the mark WMa The position (x direction) with respect to the lattice pattern RGa is obtained.

In FIG. 18, the areas indicated by broken lines above and below the reference pattern RFc and included in the reflection part Mc are the parts where the projected images of the reference patterns PX and PY of the reticle R are located. Lights 90a and 90b for detecting the reticle mark RM and the reference pattern RF are close to the wavelength of the exposure light, and the resists are exposed to light, so that the reflection portions Ma, Mc, Mb, Md, etc.
It is necessary to provide a sufficient area to shield the light so that a and 90b do not reach the wafer W. Also in this embodiment,
Optical heterodyne detection may be performed by giving a certain frequency difference to the two divided beams 167 and 168. In this case, the +1 order light 171 and the −1 order light 17 from the lattice pattern RGa
2 and the interference fringes flow at high speed at the beat frequency (Δf ₀ ) on the wafer W. When the mark WMa is detected by the optical heterodyne method as described above, since the interference fringes flow, the mark WMa is detected at a predetermined accuracy, for example, ± 10%.
The position shift amount ΔX is obtained from the phase difference obtained when the wafer stage is stopped after being driven to about 1 μm, and the wafer stage WS when stopped is obtained.
The coordinate value of T is obtained from the laser interferometer 9 as a reference position,
From both information, the position of the mark WMa on the xy coordinate system is defined. In contrast, the two beams 167, 1
In the so-called optical homodyne system, which does not have a frequency difference to 68, the wafer stage WST does not have to be kept still,
Mark position detection with high throughput is possible.

As described above, in the present embodiment, the reference pattern RF and the reference grid pattern are
In order to form RG integrally with a predetermined positional relationship, a detector
The off-axis method can monitor the relative position change (drift, etc.) between the detection center position of the off-axis type alignment mark detection system including the 177 and the projected position of the image of the reticle R with high accuracy. Can be corrected.

Further, as in the prism 166 shown in this embodiment, if the number of reflections in the measurement direction of the two beams 167 and 168 to be caused to interfere with each other has an odd number and an even number (including 0), a laser beam (beam splitter 163) is used. In the case of the optical homodyne method, the generation position of the interference fringe with respect to the reference grating pattern RG does not change with respect to the measurement direction even if the beam incident on the reference beam pattern drifts.

Further, as is clear from the arrangement in FIG. 18, the reference grating pattern RGa is located on the length measurement axis in the x direction of the laser interferometer 9, and the reference grating pattern RGc is located on the length measurement axis in the y direction. Therefore, the Abbe error when measuring the mark WMa or the like can be regarded as zero.

Further, the reference grating pattern RG and the mark detection system may be configured so that two-dimensional (x and y directions) alignment (position measurement) can be performed at one place.

Next, a fifth embodiment of the present invention will be described with reference to FIGS. The optical system shown in FIG. 19 represents only the light transmission system from the optical fibers 23 and 24 to the reticle R in the position shift detection system shown in FIG. In this embodiment, four slit openings 4 are formed in the aperture plate 30 'as shown in FIG.
The difference is that 3 ', 44', 45 'and 46' are provided. The optical fibers 23 and 24 emit laser beams when detecting the displacement of the reticle mark RM and the reference pattern RF in the x direction, and these laser beams illuminate the slit openings 43 'and 44'. The slit openings 43 'and 44' are provided at positions symmetrical in the measurement direction (x direction) from the center of the aperture plate 30 ', and extend elongated in a direction orthogonal to the measurement direction (x direction).
Similarly, the slit openings 45 'and 46' are arranged symmetrically with respect to the measurement direction (y direction), and extend in the x direction in an elongated manner.
In this embodiment, the reticle mark RM or the reference pattern RM
90a, 90b illuminating the light from two directions at an incident angle θR or θW
Each of them aims at increasing the amount of irradiation light by the slit opening by using an advantage that it is sufficient to approach the parallel light beam only at least in the measurement direction.

In FIG. 19, the aperture plate 30 'is shown as being on the pupil plane of the telecentric objective lens 31, that is, on the front focal plane. Therefore, the slit apertures 43 ', 44', etc. are used as new line light sources (actually, surface light sources because of the slit width), and the light beams 90a, 90b appropriately narrowed in the width direction of the slit are applied to the objective lens 31. The luminous flux enters the reticle R, and becomes two substantially parallel light beams inclined symmetrically in the x direction. Angle of incidence θ of these two light beams on reticle R
R is a pair of slit apertures 43 'and 44', or a pair of slit apertures, assuming that the magnification and the like of the objective lens 31 are constant.
The distance is determined by the distance between the center of the aperture plate 30 '(the point where the optical axis La passes) of 45' and 46 '.

FIG. 21 shows a light transmission system of a displacement detection system according to a sixth embodiment of the present invention, and optical fibers 23, 24, etc. are used instead of the aperture plates 30, 30 'shown in FIG. 5 or FIG. Condensing lenses 300a, 30 such as a spherical (or aspheric) lens for receiving a laser beam from the exit end, or a cylindrical lens
A light beam having an appropriate NA (numerical aperture) is narrowed by the condenser lenses 300a and 300b, and is once focused in the air as circular spots or slit-shaped spots SPa and SPb. The focal points of the spots SPa and SPb substantially coincide with the pupil plane of the objective lens 31. Therefore, there is an advantage that the irradiation light amount of the light 90a, 90b or the like is increased as compared with the case where the aperture plates 30, 30 'are provided. Also, openings 43 to 46 such as pinholes and slit openings 4
Since there is no 3 'to 46', unnecessary diffraction phenomenon does not occur at the opening edge, and the reticle R or the reference pattern
The area to irradiate RF is relatively cleanly divided and the reference member
There is also an advantage that irradiation light does not inadvertently protrude from the reflection portions Ma, Mb, Mc, Md, etc. on the lower surface of the RP.

FIG. 22 shows a light transmission system of a displacement detection system according to a seventh embodiment of the present invention. In the modified embodiment, the mark on the reticle R
A field stop 311 for limiting an irradiation area is provided according to the shape of the RM. 22, the lens system 310 corresponds to the objective lens 31 shown in FIGS. 5 and 19, and irradiates the field stop 311 with substantially parallel beams emitted in two directions. Therefore, interference fringes are formed at the position of the field stop 311. Two light beams 90a and 90 transmitted through the stop 311 with a predetermined cross-sectional shape
b is incident on the second objective lens 312 and the objective lens 31
The pupil plane e ₁ spot SPa, was condensed becomes SPb, and emitted from the objective lens 31 substantially parallel light flux 90a, the two directions as 90b illuminates the reticle R. Here, the second objective lens 31
The back focal plane of the 2 coincides with the pupil plane e _1, the front focal plane aperture 311
Are arranged so as to match. Therefore, the aperture image of the stop 311 is image-formed and projected so as to include the reticle RM, and the area of the reticle mark RM or the interference fringe illuminating the reference pattern RF is extremely clearly defined. Partial mirror as a spatial filter since there is pupil plane e ₁ in the space, here between in this embodiment and the first objective lens 31 and the second objective lens 312
A configuration in which interference light is drawn by inserting 34 can be realized very easily.

FIG. 23 is a view showing a light transmitting system according to an eighth embodiment of the present invention, which is shown in FIG. 17, FIG. 19, and FIG. It can be used for both the wafer misalignment detection system. First
As shown in FIG. 2 and FIG. 2, a frequency shifter can be used as a method of giving a difference between the frequencies of two beams incident from two directions, instead of using a rotating radial grating. In FIG. 23, a linearly polarized laser beam LB is split into two beams of P-polarized light and S-polarized light by a polarization beam splitter 320, and is incident on frequency shifters 322 and 324, respectively. The frequency shifters 322 and 324 respond to the drive signals S ₁ and S ₂ from the drive circuit 328 to change the frequencies of the emitted beams LB ₁ and LB ₂ to f ₁ and f ₂ , respectively.
Only change. Drive frequency f _1, f ₂ is set to a predetermined frequency difference Delta] f _0, the drive circuit 328 outputs a signal of frequency Delta] f ₀ of the difference as the reference signal S _3. Reference signal S ₃ is used to determine the phase difference between the photoelectric signals from the detectors 37 and 38, or 177. After the beams LB ₁ and LB ₂ are spaced at a predetermined interval by the corner mirror 326, they are incident on, for example, the optical fibers 23, 24 or 27, 28, etc., and are guided to the respective detection systems.

In this embodiment, two beams 90a, 90b or beam 16
The flow speed of the interference fringes formed by 7, 168 is
The frequency band of the high-frequency range and the change in the light and dark of the received interference light (beat frequency) also become the high-frequency range, so that the circuit scale of the detection system increases. In addition, there is an advantage that the displacement detection resolution is also increased.

As described above, in each embodiment of the present invention, the optical heterodyne detection method using a double diffraction grating, or to detect the relative displacement by the optical homodyne detection method, the present invention is not necessarily required, For example, Japanese Patent Publication No. 56-4625
As disclosed in Japanese Patent Laid-Open No. 5 (1994), a combination of laser spot (or sheet beam) scanning and detection of scattered light from a pattern (mark) edge may be used. In this case, the reticle mark RM is simply a rectangular transparent window, and a rectangular mark or a cross mark is provided as a reference pattern RF at a position of the reference member RP that matches the inside of the transparent window. The NA of the projection lens system PL (numerical aperture: usually 0.35 to
A laser beam focused to an NA smaller than 0.5) is formed as a spot on both the window of the reticle mark and the reference pattern, and the spot is scanned one-dimensionally or two-dimensionally. At this time, a largely defocused laser beam is radiated from above the reticle (the reticle side), and the beam reflected upward by the reflection portions Ma, Mb, etc. on the lower surface forms an image at the position of the reference pattern to form a spot. I do. The state at this time is the same as the state of the imaging light beam Lfr shown in FIG. When the edge of the reference pattern is scanned by the spot,
Of the scattered light from the edge, the light generated below is the reflection part Ma,
The light is reflected by Mb or the like and is detected again through the transparent window of the reticle. Therefore, if the difference in the scanning direction between the edge position of the transparent window of the reticle and the edge position of the reference pattern is monitored, it is possible to detect positional deviation in the same manner. In the case of a method of scanning a laser spot, if the arrangement is such that the scattered light from the edge of the reticle mark or the edge of the reference pattern is condensed using a relatively large numerical aperture (NA), the laser spot de- grees. Since the focus can be detected by the change in the contrast of the photoelectric signal (the rise or fall of the waveform), the focus change over time (change in the imaging optical path between the reticle surface and the reference pattern) and the position change and inclination of the image surface Etc. can be measured simultaneously. However, in each of the embodiments using the double diffraction grating, the NA of the illuminating light beam for forming interference fringes is almost zero (parallel light beam). There is a feature that the sensitivity can be maintained.

Therefore, both a system for scanning the formed laser spot via the objective lens 31 and a system for irradiating the parallel beam for the double diffraction grating from the two directions via the objective lens 31 are incorporated. When used alternately for position shift detection, xy
The shift in the direction and the shift in the z direction can be determined with high accuracy.

Therefore, a configuration of a displacement measurement system using a reference pattern (reticle mark RM) and a reference pattern (RF) having the same shape as a conventionally used alignment mark is shown in FIG. 24 as a ninth embodiment of the present invention. This will be described with reference to FIG.

FIG. 24 shows only a displacement measurement system for one mark RM on the reticle R. The mark RM is formed by a light shielding portion having a rectangular window PW. Now, a laser beam having a spectrum close to the exposure wavelength emitted from the laser light source LS is converted into a predetermined light beam diameter (parallel light beam) by a beam expander 400, and then passes through a beam splitter 402, a cylindrical lens 404, and a rotating polygon. It is incident on one reflecting surface of a mirror 406. The rotating polygon mirror 406 rotates at a constant speed in the direction of the arrow 405, and the reflected beam is deflected in the drawing in the drawing. After being reflected by the mirror 408, the beam enters the scanning lens 401, and is converged to a small spot (slit-like) light at the position of the split mirror 412,
It is linearly scanned in the direction of the arrow A _1. The split mirror 412 has a reflecting surface at 45 ° to the optical axis of the system, and is provided such that the upper end of the reflecting surface coincides with the optical axis. Therefore, when the spot light is located above the optical axis, it is directly incident on the second objective lens 414, and when the spot light is located below the optical axis, it is bent upward at a right angle by the split mirror 412. After being reflected by the mirror 416, the light enters the second objective lens 418. The beam emitted from the second objective lens 414 becomes a substantially parallel light beam and is horizontally incident on one surface 416a of a trapezoidal mirror 416 having two reflecting surfaces 416a and 416b. Reflected on the other side 416b
And exits horizontally again. The trapezoidal mirror 416 and the mirror 417 are integrally provided so as to be rotated around the center axis 416c coincident with the horizontal optical axis by 45 ° with respect to the paper surface, and function as an image rotator. Now, after the beam reflected by the surface 416b has passed through the beam splitter 418,
The light enters the first objective lens 422, is reflected vertically downward by the mirror 424, and is condensed as spot light (slit shape) SPY on the surface of the mark RM on the lower surface of the reticle R. On the other hand, the beam reflected by the split mirror 412 is the second objective lens 418,
The light enters the beam splitter 418 via the mirror 420, is reflected there, is similarly incident on the first objective lens 422, and is condensed on the lower surface of the reticle R as a spot light (slit shape) SPX. Linear scanning is performed in _two directions. Here, the spot light SPY formed by the beam passing through the image rotator (416, 417) extends in a slit shape to the left and right in the plane of the paper, is scanned in a direction perpendicular to the plane of the paper, and is spotted by the beam that bypasses the image rotator. Optical SPX
Extends in a slit shape in a direction perpendicular to the paper surface, and is scanned in the left-right direction in the paper surface. Therefore, the rotation of the rotating polygon mirror 406 causes the spot lights SPY and SPX to change the reticle mark RM.
Are alternately scanned in the y direction and the x direction, respectively. If the light-shielding portion around the window PW of the reticle mark RM is formed of a reflective metallic thin film such as chrome, reflected light is generated when the spot light SPY or SPX irradiates the light-shielding portion. This reflected light travels in the laser beam transmission path in the opposite direction, and enters the beam splitter 4 after the cylindrical lens 404.
The light is reflected at 02, and is branched to a light receiving unit including a condenser lens 430 and a photoelectric detector 432. The beam transmitted through the window PW and passing through the projection lens system PL to the reference member RP is the reference pattern RF.
Irradiates light information such as reflected light and transmitted light from the reference pattern RF. This optical information is similarly received by the photoelectric detector 432 after moving backward through the window PW.

As described above, in the optical system shown in FIG. 24, the first objective lens 422 is defined as a telecentric system, the rear focal plane coincides with the plane of the mark RM of the reticle R, and the front focal plane (pupil plane) is the second objective lens. It is determined so as to coincide with the respective rear focal planes of 414 and 418. The front focal points of the second objective lenses 414 and 418 coincide with the vertex of the split mirror 412, and also coincide with the rear focal points of the scanning lens 410. Scanning lens 410
Is determined so as to coincide with the reflection surface of the rotating polygon mirror 406.

The front focal plane of the first objective lens 422 is
It is determined to be conjugate with the pupil plane of L. For this reason, rotating polygon mirror 4
Due to the beam deflection by the reticle R, the reticle R is always scanned while being irradiated with the laser beam vertically, and so-called telecentric illumination is performed.

FIG. 25 is a partial cross-section showing a preferred arrangement of the reference member RP and the wafer W in the present embodiment, and a small rectangular window at the center is located at a position facing the window PW of the reticle mark RM on the upper surface of the reference member RP. The reference pattern RF is formed by a rectangular (or circular) light-shielding layer having the following. A reflection portion Mb is formed in a predetermined area including the reference pattern RF on the lower surface of the reference member RP. Here, the reference member RP is a piece of quartz glass, and the projection lens system PL is located immediately above the reference pattern RF. The projection image of the window PW of the reticle mark RM is formed on the same plane as the reference pattern RF via the reflection part Mb under the exposure wavelength.
Here, when the window PW is viewed from the reticle R side, FIG. 26 (A)
As shown in the figure, it is smaller than the external dimensions of the reference pattern RF,
In addition, it is determined to be larger than the outer dimensions of the transparent window in the reference pattern RF. Now, in Fig. 25, the window
The laser beam that has passed through the PW is largely defocused on the upper surface of the reference member RP, passes through the periphery of the reference pattern RF, etc., reaches the reflection portion Mb, reflects there, and spots on the back surface of the reference pattern RF. An image is formed as light (SPX '). For the spot light SPX that scans along the arrow A ₂ in the x-direction as the upper reticle R FIG. 26 (A), the spot light SPX irradiating the reference pattern RF from the lower surface 'arrow A _3' along the Scanning is performed in the x direction. Due to the irradiation of the spot light SPX ′, the reflected light reflected by the reference pattern RF and traveling toward the reflecting portion Mb travels in the same optical path as that of the incident light in the opposite direction, and the spot is reflected within the window PW of the mark RM on the lower surface of the reticle R. The image is re-formed and received by the photoelectric detector 432 in the measurement optical system. On the other hand, the spot light SPX ′ transmitted through the transparent window of the reference pattern RF upward has a relatively large numerical aperture and is re-imaged on the reticle R side by the projection lens system PL. Is not in a conjugate relationship with the lower surface of the reticle R, so that the position of the mark RM defocuses more than the size of the window PW. For this reason, most of the beam of the spot light SPX 'that has passed through the reference pattern RF from below to above is shielded around the window PW of the reticle mark RM, and the amount of light traveling to the upper part of the reticle R, that is, the photoelectric detector
The amount of light reaching 432 decreases.

FIG. 26A shows a state in which the reticle mark RM and the reference pattern RF are aligned, and the x-direction scanning of the spot light SPX and the y-direction scanning of the spot light SPY are alternately performed in time. Now, the spot light SPX (SPX ′) is the edge Ea of the reticle mark RM and the edges Eb and E of the reference pattern RF.
When scanning is performed in the order of c and the edge Ed of the mark RM, the level of the photoelectric signal SS of the photoelectric detector 432 changes as shown in FIG. 26 (B). First, spot light SPX is reticle mark RM
When scanning outside the window PW, the light reception amount becomes extremely large. When the spot light SPX is a light-shielding layer (reflective) portion of the reference pattern RF in the window PW, that is, between the edge Ea and Wb,
When scanning between the edges Ec and Wd, the spot light
Since the light-shielding layer portion of the reference pattern RF is irradiated with SPX ′ from below, the light once reflected by the reflection portion Mb is received by the photoelectric detector 432, and the light reception amount during this period decreases. . In addition, the spot light SPX '
When scanning within the transparent window of the reticle mark RM, that is, when scanning between the edges Eb and Ec, as described above,
Since the amount of light passing through the PW is only a small part of the largely defocused beam, the amount of light received during this period is extremely reduced.

Therefore, two appropriate slice levels VR ₁ and VR ₂ are set and the positions Xa, Xb and X of the rising and falling portions on the photoelectric signal SS are set.
c, Xd is detected. Here, when the spot light is scanned linearly at a constant speed, the positions Xa, Xb, Xc, and Xd are measured as edge detection times. Thus, based on the detected position, the relative positional deviation ΔX ₀ between the reticle mark RM and the reference pattern RF in the x direction is obtained. The displacement ΔX ₀ is Required by The same applies to the measurement in the y direction. However, since the photoelectric detector 432 is shared for measurement in the x direction and the y direction, a signal waveform as shown in FIG. 26 (B) repeatedly occurs in time. Therefore, it is necessary to perform the calculation by alternately dividing the displacement measurement value in the x direction and the displacement measurement value in the y direction for each signal waveform process by appropriate synchronization processing. Even in such a case, the slice levels VR ₁ and VR ₂ may be common for the x direction and the y direction.

In this embodiment, the reference pattern RF as in the first embodiment is used.
Since it is not necessary to provide a diffraction grating pattern as the reticle mark RM, there is an advantage that the reticle mark RM can be made smaller.

In the present embodiment, the amount of reflected light is received. However, only the edge scattered light generated at the edges Ea, Eb, Ec, and Ed of each pattern may be photoelectrically detected.

In each embodiment of the present invention, the best effect can be obtained when the spectrum of the excimer laser light is narrowed and applied to a stepper mainly using quartz as an optical material of the projection lens system. However, the same effect can be obtained even when applied to other types of steppers that do not use excimer laser light. For example, for a stepper using g-line exposure light at a wavelength of 436 nm, a He-Cd laser light source that outputs a laser beam having a wavelength of 441.6 nm as a laser beam LS for position shift detection, and a beam having a wavelength of 488 and 514 nm are output. An argon ion laser light source, a He-Ne laser light source that outputs a beam with a wavelength of 544 nm, an SHC harmonic of a semiconductor laser, or the like can be used. For a stepper using i-line having a wavelength of 365 nm as exposure light, a He-Cd laser light source outputting a beam of 325 nm or an argon ion laser light source outputting a beam of 365 nm is suitable as the laser light source LS. As described above, when the wavelength of the exposure light and the wavelength of the laser light source for alignment are different, the image forming position may be different due to chromatic aberration under the light of both wavelengths.
May be adjusted so that the position in the optical axis direction of the reference pattern RF provided in the step (b) coincides with the image forming position with the reticle under the light of the wavelength for alignment.

Further, as shown in FIG. 1, the laser light source LS used in each embodiment is an argon ion laser containing an SHG crystal inside, and therefore has a problem that the cost of the laser light source itself is high. However, since there is enough output power, the laser beam from one laser light source LS can be branched and sent to a plurality of exposure apparatuses using an optical fiber or the like. This reduces the laser cost per exposure apparatus. However, in this method, failure (down) of one laser light source means stoppage of operation of a plurality of exposure apparatuses, which is extremely inconvenient. Therefore, it is better to prepare a spare laser light source LS and immediately switch the laser light source to be used by moving a mirror or the like. In this case, when compared with the laser cost of providing one laser light source for one exposure apparatus, combining three or more exposure apparatuses with two laser light sources (one spare) is costly. The effect of reduction is obtained.

By the way, as described in each embodiment, the position of the reticle R is monitored with reference to the reference pattern RF of the reference member RP. In particular, in the case of a stepper, the criterion for arranging the shot (chip) areas in a matrix on the wafer W is a coordinate system defined by the feed coordinate system of the wafer stage WST, that is, the orthogonal x, y measurement axes of the laser interferometer 9. System. For this reason, it is assumed that a line segment connecting the reference patterns RFa and RFb provided in at least two places of the reference member RP is arranged so as to be exactly parallel to the x-axis as shown in FIG. 8, for example. , At least two marks RM of reticle R and reference patterns RFa, RFb of reference member RP
When the reticle R is found to be slightly rotating in the xy plane by measuring the positional deviation from the reticle R, the rotation (θ) of the reticle R itself may be corrected, but the wafer stage WST
The upper wafer holder WH may be rotated in the same direction by the measured rotation amount.

However, if the arrangement of the reference patterns RFa, RFb, etc., has caused slight rotation (or parallel displacement in the x and y directions) with respect to the xy coordinate system defined by the laser interferometer 9, some countermeasure is required. is there. One countermeasure is, for example, as shown in FIGS. 17 and 18, a second reference pattern RG provided in at least two places of the reference member RP, and a position in the projection image plane based on the second reference pattern RG. Mark (grid pattern
WM, etc.) to detect an off-axis mark. At this time, as a mark located in the projection image plane, a reference mark (fiducial mark) fixed on wafer stage WST at substantially the same height position as the wafer surface may be used. Figure 27 shows the reference member RP
The arrangement of a reference pattern RG suitable for detecting a rotational displacement with respect to the xy coordinate system is shown, and will be described as a tenth embodiment.

In FIG. 27, it is assumed that the reference member RP is rotated by a small angle Δθ with respect to the xy coordinate system. The reference grid patterns RGa and RGb as the second reference patterns are both provided so as to be suitable for detecting a displacement in the x and y directions. The distances between the lattice patterns RGa and RGb from the center CC of the reference member RP are assumed to be equal, and are provided in a point-symmetric relationship. A fiducial mark FM is provided on wafer stage WST, and diffraction grating marks Fmx and Fmy equivalent to wafer mark WM are formed on the upper surface thereof. Mark Fm
x is used for position measurement in the x direction, and mark Fmy is used for position measurement in the y direction.

In such a configuration, first, after positioning the wafer stage WST so that each of the marks Fmx and Fmy overlaps the lattice pattern RGa, the lattice pattern RGa is formed using the off-axis type mark detection system shown in FIG. The wafer stage WST is finely moved in the y direction so that the lattice arrangement in the y direction and the mark Fmy are aligned in the y direction in a predetermined alignment state. Then, when this alignment is completed, the coordinate value Ya of the wafer stage WST in the y direction is set to the laser interferometer 9.
Read from. Next, while keeping the coordinate value Ya, the wafer stage WST is moved in the x direction by a fixed distance SL (the distance between the patterns RGa and RGb), and the positional deviation ΔY between the grid line in the y direction of the grid pattern RGb and the mark Fmy is calculated there. And a mark detection system as shown in FIG. Since the angle Δθ is usually a very small amount of 1 ° or less, it can be obtained by Δθ ≒ ΔY / SL.
Alternatively, after the wafer stage WST is sent in the x direction, the wafer stage WST is finely moved in the y direction such that the displacement of the grid pattern RGb and the mark Fmy detected by the mark detection system in the y direction becomes zero, Is read as the coordinate value Yb. Therefore, in this case, it is obtained by Δθ ≒ (Ya−Yb) / SL. When the second reference pattern RG and the fiducial mark FM are aligned using the diffraction grating as in the present embodiment, the relative displacement in the y direction at two places of the patterns RGa and RGb is equal to the grid pitch of the mark Fmy. Is preferably within the range. Since the pitch of the grid pattern or mark Fmy formed on the wafer is usually several microns or more, it is relatively easy to manufacture in advance by controlling the angle Δθ within such accuracy.

Alternatively, a dedicated slit mark or the like may be formed instead of the reference pattern RG of the diffraction grating or at another position, and these slit marks may be simultaneously observed from above the reference member RP with a microscope or the like. At this time, the slit mark and the fiducial mark formed on the reference member RP are separated at a certain interval in the direction of the optical axis AX, so that the objective lens and the eyepiece (or the relay system) of the mark observation microscope are connected. It is preferable to provide a bifocal element (such as a quartz lens) using a birefringent element between them, and detect slit marks separated in the optical axis direction on the same image plane. Alternatively, the fiducial mark is formed as an opening slit, and the opening slit is illuminated from the inside of the wafer stage WST, and the matching state with the reference pattern RG is photoelectrically detected above the reference member RP, thereby obtaining at least two reference patterns. The coordinate position of RG may be measured.

In each case, the second reference patterns RGa, RGb
Since the positional relationship between the reference patterns RFa and RFb relative to the reference patterns RFa and RFb is precisely obtained in advance, the positions of the reference patterns RFa and RFa in the xy coordinate system can also be obtained accurately.

Further, the positional relationship between each of the reference patterns RFa, RFb, RFc, and RFd and each of the reticle marks RMa, FMb, RMc, and RMd, or the arrangement of the reference patterns RGa, RGb, RGc, and RGd is determined by the exposure apparatus and the reticle R. Since it differs depending on the type, the mark arrangement of the reticle R, and the like, it is necessary to store the predetermined positional relationship in the exposure apparatus.

In each of the above embodiments, the reference mirror (fixed mirror) of the laser interferometer 9 for measuring the position of the wafer stage WST is:
Since the reference member RP is fixed almost integrally with the vicinity of the reference pattern RF, an effect that a change in the position of the projected image of the reticle pattern can be accurately monitored is obtained.

The optical transmission system of the optical heterodyne system shown in FIGS. 2, 4 and 5 can be independently applied to other position detecting devices in addition to the use method shown in each embodiment of the present invention. . That is, if a two-dimensional diffraction grating pattern is formed on an object for which a two-dimensional (x, y direction) displacement is to be detected, an extremely simple light transmission system as shown in FIG. Optical heterodyne detection can be performed alternately.

〔The invention's effect〕

As described above, in the present invention, a heat source such as a light source is separated from a projection lens or the like, and guided to an illumination optical system for alignment by an optical fiber or the like. An apparatus and a projection exposure apparatus can be obtained.

[Brief description of the drawings]

FIG. 1 is a diagram showing a configuration of a projection exposure apparatus according to a first embodiment of the present invention, and FIG. 2 is a perspective view showing a configuration of a part of a light transmission system for generating illumination light for a position shift detection system. FIG. 3 is a perspective view showing the structure of a distributor for dividing illumination light into a plurality of parts, and FIG.
The figure shows a rotating disk (radial grating) in the light transmission system
5, FIG. 5 is a perspective view showing the configuration of a misregistration detection system, FIG. 6 is a plan view showing the shape of a mark on a reticle, and FIGS. 7 (a) and 7 (b) are reference members. FIG. 8 is a perspective view showing the structure of the reference member, FIG. 9 is a cross-sectional view showing the imaging relationship in the arrangement of the projection lens, the reference member and the wafer, and FIG. FIG. 11 is a cross-sectional view showing illumination light passing through the reticle mark and interference light generated from the reticle mark. FIG. 11 is a cross-sectional view showing interference light generated in the reference pattern. FIG. FIG. 13 is a plan view showing a mark arrangement, FIG. 13 is a plan view showing a configuration of a reference member according to a second embodiment of the present invention, and FIG.
FIG. 15 is a plan view showing a mark arrangement of a reticle suitable for the second embodiment, FIG. 15 is a perspective view showing an arrangement of a displacement detection system in the second embodiment, and FIG. 16 is a third embodiment of the present invention. FIG. 17 is a diagram showing a mechanism for adjusting optical characteristics according to an embodiment, FIG. 17 is a partial sectional view showing the structure of a reference member and the configuration of a mark detection system according to a fourth embodiment of the present invention, and FIG. FIG. 19 is a partial plan view showing the arrangement of each pattern of the reference member, and FIG.
A diagram showing a partial configuration of a position shift detection system according to the embodiment of the present invention,
FIG. 20 is a plan view showing the structure of the aperture plate shown in FIG. 19, FIG. 21 is a diagram showing a part of a displacement detection system according to a sixth embodiment of the present invention, and FIG. FIG. 23 is a diagram showing a configuration of a part of a displacement detection system according to a seventh embodiment, FIG. 23 is a diagram showing a configuration of a light transmission system according to an eighth embodiment of the present invention, and FIG. FIG. 25 is a partial cross-sectional view showing a configuration of a reference member suitable for the apparatus shown in FIG. 24, and FIG. 26 is a partial sectional view showing a configuration of a reference member suitable for the apparatus shown in FIG. 24. FIG. 27 is a plan view showing the appearance, and FIG. 27 is a plan view showing the arrangement of a reference member and fiducial marks according to a tenth embodiment of the present invention. [Description of Signs of Main Parts] 1 ... Excimer laser light source, 15 ... Rotating disk, 23, 24,
27, 28 ... optical fiber, 30, 30 '... aperture plate,
31 Objective lens, 34 Partial mirror, 37, 38 Photodetector, 39 Phase meter, LS Laser light source, R Reticle, PL Projection lens system, RP Reference member , W… wafer, RM, RMa, RMb, RMc, RMd… reticle mark (reference pattern), RF, RFa, RFb, RFc, RFd… reference pattern, Ma, Mb, Mc, Md… reflection part, RPa, RPb: quartz plate

──────────────────────────────────────────────────続 き Continued on front page (58) Field surveyed (Int.Cl. ⁶ , DB name) H01L 21/027

Claims (12)

(57) [Claims] 1. A position detecting device for detecting a relative positional deviation between a first object and a second object, comprising: a light source for emitting alignment light; a split optical member for splitting the alignment light into two; An optical waveguide for transmitting the two light beams; and an illumination optical system for irradiating the two light beams to diffraction marks provided on at least one of the first object and the second object from directions different from each other; A position detecting device comprising a light receiving element for receiving optical information from a diffraction mark. 2. The position detecting device according to claim 1, wherein said optical waveguide includes two optical waveguides for independently transmitting said two light beams. 3. The position detecting device according to claim 1, wherein said optical waveguide is an optical fiber. 4. The position detecting device according to claim 3, wherein said optical fiber is a single mode storage optical fiber. 5. The position detecting device according to claim 1, wherein said light receiving element receives interference light of a plurality of diffracted lights traveling in the same direction from said diffraction mark. . 6. The position detecting device according to claim 1, wherein the split optical member includes a modulation optical member that gives a different frequency difference between the two light beams. 7. The position detecting device according to claim 1, wherein the optical waveguide has an aperture plate on the exit end side for improving coherency of each of the two light beams. 8. A projection exposure apparatus for exposing a pattern on a first object onto a second object via a projection system, comprising: a light source for emitting alignment light; a division optical member for dividing the alignment light into two parts; 2 for independently transmitting the two light beams from the split optical member
Two optical waveguides; an illumination optical system for irradiating the two light fluxes to diffraction marks provided on at least one of the first object and the second object from directions different from each other; and optical information from the diffraction marks And a position control device for adjusting a relative position between the first object and the second object based on a signal from the light receiving element. 9. The illumination optical system irradiates the two light beams to the second object via the projection system.
Projection exposure apparatus. 10. The projection exposure apparatus according to claim 8, wherein the illumination optical system irradiates the second light beam to the second object without passing through the projection system. 11. A substrate position detecting device for irradiating a mark provided on a substrate with alignment light and detecting light information from the mark, comprising: a light source for emitting the alignment light; an alignment light from the light source; A splitting optical element for splitting; at least two frequency modulators for giving a frequency difference to the two alignment lights split by the splitting optical element; an alignment light emitted from the two frequency modulators being incident on the substrate; A position detecting device comprising: an optical element that emits light toward the substrate; and a light receiving element that receives light information from the substrate. 12. The position detecting apparatus according to claim 11, further comprising a combining optical system for combining the two divided alignment lights so as to be substantially parallel at a predetermined interval.