JPH05114544A - Projection aligner - Google Patents

Abstract

PURPOSE:To prevent the mingling of the alignment light from a mask and the alignment light from a substrate, by making irradiation regions of a first beam and a second beam be mutually isolated from each other on a mask. CONSTITUTION:Two first beams LBr1, LBr2 illuminate a mask mark at a specified cross angle. Two second beams LBw1, LBw2 illuminate a substrate mark at a specified cross angle, via a transparent region RW of a mask. Objective optical systems 3-5 are so installed that the respective irradiation regions of the beams on the mask are isolated in the direction perpendicular to the grating pitch direction of the mark. Thereby irradiation position on the mask of the beams with which the mask mark is irradiated and the transmission position on the mask of the beams with which the substrate mark is irradiated are mutually isolated, so that influence caused by the mingling, the mutual interference, etc., of the mask side beam and the substrate side beam is remarkably reduced.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a projection exposure apparatus having an optical mark detection system for a substrate (reticle, wafer, etc.) on which a diffraction grating-shaped alignment mark is formed.

[0002]

2. Description of the Related Art As an example of a conventional device of this type, for example,
For example, the structure shown in FIG. 1 is considered. This configuration
In the system proposed by the applicant in Japanese Patent Application No. 2-109293
Yes, it is not known. Figure 1 shows the TTR of the projection exposure system.
Shows the schematic configuration of the Luther reticle type alignment system.
Then, the projection optical system below the reticle 7 and a photosensitive substrate such as a wafer.
Illustration of the plate is omitted. However, in FIG. 1 (B)
Only the optical axis AX of the projection optical system is shown. Figure 1 (B)
As shown in, the reticle R is installed at an angle of 45 ° above the reticle R.
The dichroic mirror 6 is arranged and the exposure illumination light I
L (wavelength λ₀) Is from the side to the dichroic mirror 6
It is incident and is reflected vertically here, and the circuit pattern of the reticle 7
Irradiate only the burned area. Dichroic mirror 6 is dew
Wavelength of light IL₀Has a reflectance of 90% or more
And the wavelength λ₀Much longer than λ ₁Alignment
It has a transmittance of 50% or more with respect to the illumination light for use. exposure
The light IL is the g-line (λ₀= 436n
m), i line (λ₀= 365 nm) or KrF excimer
Laser light (λ₀= 248 nm), etc. are used for alignment
The illumination light for the
He-Ne laser light considering photosensitivity and low absorption
(Λ₁= 633 nm) or the like is used.

Now, the light transmitting system 1 which emits the illumination light for alignment generates _two laser beams LB ₁ and LB ₂ which are divided with substantially the same intensity, and these beams LB ₁ and LB are generated.
₂ is reflected by the beam splitter 2, then the lens system 3,
The beam reaches the mark area RM of the reticle 7 via the beam splitter 4, the objective lens 5, and the dichroic mirror 6. In FIG. 1B, the two beams LB ₁ and LB ₂ are positioned in the direction perpendicular to the plane of the paper, and the state is shown in FIG. 1A. FIG. 1A is a diagram of the configuration of FIG. 1B viewed from the left side. Beams LB ₁ and LB ₂ after being reflected by the beam splitter 2 and passing through the lens system 3 are light beams of the objective lens 5. They are located symmetrically parallel to each other with the axis in between. And the objective lens 5
The two beams LB ₁ and LB ₂ that emitted the reticle R
After intersecting with the plane Pw in the space above, the mark area RM of the reticle 7 is irradiated while expanding again.

Here, the two beams LB ₁ and LB ₂ are represented by simple lines in the figure, but in the pupil space (Fourier exchange region) between the lens system 3 and the objective lens 5,
It is a convergent divergent beam, which is a parallel light beam in the image space (inverse Fourier exchange region) between the objective lens 5 and the projection optical system. Further, as shown in FIG. 1B, the optical axis of the objective lens 5 is parallel to the optical axis AX of the projection optical system at the position of the reticle 7.

Now, the two beams LB ₁ and LB ₂ have plane P
It is set to intersect at w because this corresponds to the amount of axial chromatic aberration (longitudinal chromatic aberration) ΔL of the projection optical system.
The wavelength of the illumination light for alignment λ ₁ (eg 633 nm)
Under the condition, the surface Pw and the wafer surface are conjugated with respect to the projection optical system. Therefore, the mark area R of the reticle 7
Two beams LB ₁ and LB that have passed through the transparent part (window) of M
_The beams 2 converge into beam waists near the pupil plane (Fourier exchange plane) in the projection optical system, and then converge to form two parallel light beams that intersect on the wafer.

An example of the planar shape of the mark area RM is shown in FIG. In FIG. 2, two beams LB ₁ and LB
₂ indicates the mark position measuring direction (arrow M) on the reticle 7.
D), and the diffraction grating-shaped marks RMa and RMb having a pitch in the measurement direction within the irradiation area of each beam.
Is provided. Around the marks RMa and RMb is a light-shielding region such as chrome, and a window RW for passing a part of the beams LB ₁ and LB ₂ to the wafer side is provided in the light-shielding region in the direction orthogonal to the mark pitch direction. Is formed in. The diffraction grating mark WM formed on the wafer is
In the state where the alignment between the reticle and the wafer is achieved, it is located immediately below the light-shielding region between the window RW and the marks RMa and RMb as shown in FIG. Therefore, even if exposure is performed in this state, the mark WM on the wafer is protected by the exposure light IL that has not passed through the window RW.

On the other hand, the beam LB for alignment₁, L
B₂Wavelength λ₁On the other hand, the projection optical system has a lateral chromatic aberration (horizontal
A beam that has passed through the window RW due to having a quantity Δβ
LB ₁, LB₂Marks WM so that they intersect on the wafer
Irradiate. The pitch direction of the mark WM is the measurement direction MD
Two of them are aligned and symmetrically inclined in the pitch direction.
Beam LB₁, LB₂Indicates the area Am including the mark WM.
When irradiating at the time, from the mark WM to the vertical direction ± 1 next time
Folding interference light (interference beam) BTW is generated. This dried
The cross beam BTW is the center of the pupil plane of the projection optical system (of the optical axis AX
After focusing on the spot at the passing point), the window R of the reticle 7
It returns to the central portion of W as a substantially parallel light beam.

As shown in FIG. 1, the interference beam BTW from the mark WM travels parallel to the optical axis of the objective lens 5 of the alignment system, and the beam splitter 4, the lens system 3, the beam splitters 2 and 11, and the mirror 15 are caused to travel. Aperture (or slit) 16 is reached via. The aperture 16 is arranged so as to be conjugate with the reticle 7 (mark region RM) with respect to the combined system of the objective lens 5 and the lens system 3, and transmits only the interference beam BTW (parallel light flux at this position). Further, the interference beam BTW is condensed by the lens system 17 and received by the photoelectric detector 18.

On the other hand, from the mark RMa on the reticle 7,
LB₁Beam LB₁Reverse in the same direction as
Propagating first-order diffracted light is generated, and a beam is emitted from the mark RMb.
LB ₂Beam LB₂Reverse in the same direction as
First-order diffracted light is generated. These first-order diffracted lights are
Object lens 5, beam splitter 4, lens system 3, beam
It reaches the beam splitter 11 via the splitter 2 and
It is reflected here and reaches the aperture (or slit) 12.
It The aperture 12 is a combination of the objective lens 5 and the lens system 3.
Consistent with the reticle 7 (mark area RM) regarding the system
Placed, and the first-order diffracted light from each of the marks RMa and RMb
It has an opening for passing only. 2 through aperture 12
One first-order diffracted light is a monitor arranged conjugate with the plane Pw.
Intersect on the lattice plate 13. Transparent to the monitor grid plate 13
-Shaped diffraction grating is formed, where the re-diffracted light is diffracted.
Of these two, two specific orders are coaxial and interfere with each other.
The light is received by the photoelectric detector 14 as a beam BTR. here
To set the pitch of the marks RMa and RMb on the reticle 7 to P
gr, the pitch of the mark WM on the wafer side is Pgw, and
The magnification of the shadow optical system is m (m = 1/5 with a reduction projection lens
5), the relation of each pitch is defined by the following equation.
ing.

Pgr = m · Pgw / 2 That is, the reticle side and the wafer side have a double pitch relationship, and the marks RM of the _two beams LB ₁ and LB ₂
Letting θr be the incident angle with respect to a and RMb, the incident angle θw with respect to the wafer mark WM is θw = m · θr, and the pitch Pgw is set so that the interference beam BTW of ± 1st order diffracted light is generated in the vertical direction from the wafer mark WM. Then, the first-order diffracted light from each of the marks RMa and RMb of the reticle naturally goes backward in the respective optical paths of the beams LB ₁ and LB ₂ .

Here, the two beams LB from the light transmitting system 1
Assuming that ₁ and LB ₂ are coherent (same polarization state), two beams LB ₁ and LB are formed on the wafer mark WM.
Fixed pitch by the intersection of ₂ (actually the PGW / 2) interference fringes appear in. When a wafer stage (not shown) is finely moved at a constant speed so that the wafer mark WM moves in the pitch direction with respect to this interference fringe, the photoelectric detector 18 obtains a sinusoidal photoelectric signal Sw as shown in FIG. Be done. This signal Sw
1 cycle corresponds to the amount of movement of one pitch of the interference fringes of the mark WM, that is, Pgw / 2. In FIG. 3, the horizontal axis represents the moving position of the mark WM, and the vertical axis represents the intensity of the photoelectric signal. In alignment, the mark WM (wafer stage) is pre-scanned so that the peak value and the bottom value of the signal Sw are obtained, the slice level Vs is set between the peak and the bottom, and the slice level Vs and the signal are set. The position of the intersection (zero cross point) with Sw may be obtained, or the wafer stage may be positioned so that the signal Sw matches the slice level Vs. Therefore, a laser interference type length measuring device is provided on the wafer stage, and the level of the signal Sw is measured by an up / down pulse (for example, 0.0) from the length measuring device.
It is sufficient to use a method in which digital sampling is performed in response to every 1 μm) and the data is stored in the memory. in this case,
Assuming that a half cycle of the signal Sw is used for alignment, it is necessary to find the coordinate position of the wafer mark WM in advance within a range of ± Pgw / 4. Therefore, as shown in FIG. 1B, another alignment system 59 for observing the mark WM on the wafer via the reticle 7 and the projection optical system is provided, and the position of the mark WM is ± Pgw / After roughly detecting within the range of 4, the process shifts to a precise alignment using the signal Sw under the TTR alignment system via the objective lens 5. The alignment system 59 may be a method of detecting the wafer mark only through the projection optical system, or a method of independently detecting the mark by an off-axis system provided separately from the projection optical system.

On the other hand, the alignment of the reticle 7 is performed in exactly the same manner, and the marks RMa, R of the reticle are aligned.
When the reticle 7 is finely moved so that Mb moves in the pitch direction, the photoelectric detector 14 outputs a signal Sm similar to that shown in FIG.
Is output. However, on the reticle side, interference fringes are not formed on the marks RMa and RMb, but interference fringes are formed on the monitor grating plate 13 due to the interference of the ± 1st-order diffracted light. Along with this, it moves in the pitch direction. Therefore, the signal Sm from the photoelectric detector 14
The reticle alignment is performed by positioning the reticle stage (not shown) so that is at a predetermined level (for example, Vs).

The above TTR alignment system is a two-beam system.
Of the Wataru-type alignment methods, it is called homodyne
And another method is to emit light from the light transmission system 1.
Two beams LB₁, LB₂A constant frequency difference Δf
The heterodyne method with the above can also be applied. Heterodyne
The application of the law also applies to Japanese Patent Application No. 2-109293 mentioned above.
Although disclosed in detail, it is already known
Japanese Unexamined Patent Publication No. H2-1231504 and Japanese Unexamined Patent Publication No. 62-56818.
The same thing is disclosed in gazettes, so here
The description of the detailed configuration is omitted. Two bees
LB₁, LB ₂The frequency difference Δf between the photoelectric detector 1
Sufficient response to changes in sine wave intensity of 4 and 18
Is set to a value such as 100 KHz or less.
In addition, because of the heterodyne method, as shown in FIG.
The position of the beam splitter 4 in the optical path of the alignment system.
Beam LB from the light transmission system 1₁, LB₂Split a part of
And lead it to the lens system (inverse Fourier transform) 8
The two beams are transmitted by the system 8 on the transmission type reference grating plate 9.
The ± 1st-order diffracted light generated from the reference grating plate 9 is crossed and dried.
Structure for receiving the cross beam BTS by the photoelectric detector 10
To do. This photoelectric detector 10 has a reference signal S of frequency Δf._R
Is output.

When the heterodyne method is used, the interference beam BTW and the reticle mark R obtained from the wafer mark WM are obtained.
Interference beam BTR formed by Ma, RMb and monitor grating plate 13, and interference beam BTS from reference grating plate 9
In both cases, the intensity changes sinusoidally at the frequency Δf even when each mark is stationary. Therefore, as shown in FIG. 4, the phase difference Δφ _M of the signal Sm (broken line) from the photoelectric detector 14 with respect to the reference signal S _R and the signal S _W from the photoelectric detector 18
The phase difference Δφ _W with respect to the reference signal S _R is calculated, and the difference (Δφ _M / 2−Δφ _W ) is calculated to obtain the positional deviation amount in the measurement direction MD between the reticle and the wafer. In this case, the phase difference Δφ _M and Δφ _W are ± 180
Measured in the range of °, but the range is the wafer mark W
This corresponds to the range of ± 1/4 of the pitch of each of M or the reticle marks RMa and RMb. Therefore, if the pitch Pgw of the wafer mark WM is set to 4 μm and the phase measurement resolution with stable reproducibility is set to 0.5 °, the detection and resolution of the positional deviation amount reaches about 0.0028 μm (2.8 nm).

[0015]

In the above-described technique on which the present invention is based, the two beams LB ₁ and LB ₂ irradiate the marks RMa and RMb on the reticle 7, respectively, and a part of each beam is Since the beam diameter is relatively large so as to pass through the window RW, it is difficult to completely separate and receive light even if only the necessary return light is extracted by the apertures 12 and 16 on the light receiving side. , Window R
Interfering beam BTW from wafer mark WM passing W
There is a problem that a part of the light is diffracted by the edge portion of the window RW or a part of the reticle marks RMa and RMb and leaks into the light receiving system (photoelectric detector 14, monitor grid plate 13, etc.) on the reticle side. .. Further, in the mark arrangement as shown in FIG. 2, when a new mark is replaced next to the mark WM on the wafer, the replacement mark on the reticle is
The position is within the irradiation region of the beam LB ₁ or LB _{2 in} FIG. 2 and is aligned with the mark WM in the measurement direction MD. At this time, diffracted light is also generated as stray light from the mark on the reticle for overwriting. This stray light is mixed into the light receiving system (12, 13, 14) on the reticle side and the light receiving system (16, 17, 18) on the wafer side, which causes erroneous detection during alignment (positional deviation detection). There may be a problem such as The present invention has been made in view of such a problem, and it is necessary to prevent the alignment light from the reticle (mask) and the alignment light from the wafer (substrate) from essentially mixing or avoid mixing them. It is an object of the present invention to obtain an exposure apparatus having an alignment apparatus having a configuration that can be sufficiently separated in the step of signal processing even if it is not possible.

[0016]

Therefore, in the present invention, a region (RW) which is transparent to the alignment beam is provided on the mask (reticle) side, and a grid-like mark (RMa,
A first beam transmission system which is provided near RMb) and creates _two first beams (LBr ₁ and LBr ₂ ) for irradiating a mask mark at a predetermined crossing angle as an irradiation means of an alignment beam. (19A; 25A, 26A, 27
A, 27B; 37; 40, 42, 25A, 26A, 27
A, 27B; 40, 33R to 36R, 38R, 39R,
26A, 27A) and a second beam for creating _two second beams (LBw ₁ , LBw ₂ ) for irradiating the substrate mark (WM) at a predetermined crossing angle through the transparent region (RW) of the mask. Light transmission system (19A; 25B, 26B, 27B; 3
7; 40, 41, 25B, 26B, 27B; 40, 33
W-36W, 38W, 39W, 26B, 27B).

Further, the two first beams (LBr ₁ , L
4 beams are masked so that each irradiation area on the mask of Br ₂ ) and two second beams (LBw ₁ , LBw ₂ ) is separated in a direction orthogonal to the grating pitch direction of the mark. The objective optical system (3 to 5) that emits light toward is provided.

[0018]

In the present invention, the mask marks (RMa, R
Since the irradiation position of the beam for irradiating Mb) on the mask and the transmission position of the beam for irradiating the substrate mark (WM) on the mask are separated from each other, the beam on the mask side is different from the conventional one. The influence of mixing with the beam on the substrate side, mutual interference, etc. is significantly reduced. Further, if a frequency difference (for example, 1 MHz or more) higher than the response frequency of the alignment signal processing system including the photoelectric detector is given between the light transmitting beam to the mask mark and the light transmitting beam to the substrate mark, Even if the beam or stray light from the mask side slightly interferes with the beam stray light or the like from the substrate side, the influence thereof does not cause an error in mark position detection.

[0019]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An example of the overall construction of a projection exposure apparatus to which the present invention is applied will be described with reference to FIG. In FIG.
Reticle 7 with circuit pattern and alignment mark
Are held on a reticle stage RST that is two-dimensionally movable in the X, Y, and θ directions. Reticle stage RST is driven by reticle stage control system 200 including a motor or the like for positioning reticle 7 in the apparatus or for alignment between reticle 7 and wafer W. The XY plane of the reticle stage RST (the optical axis A of the projection optical system
A reticle interferometer 202 is provided to detect a movement position within a plane (perpendicular to X). Reticle interferometer 202
Is the X of reticle stage RST (ie reticle 7)
Direction, the movement position in the Y direction, and the displacement amount in the θ direction (rotation amount in the XY plane) are measured.

On the other hand, the wafer W is held on a two-dimensionally movable wafer stage WST and moves on an XY plane perpendicular to the optical axis AX of the reduction projection lens PL as a projection optical system. Wafer stage WST is driven and controlled by wafer stage control system 204 including a motor and the position thereof is measured by wafer interferometer 206. Here, the projection lens PL has a wavelength λ of the illumination light IL from the exposure illumination system 208.
A two-sided telecentric system in which chromatic aberration is favorably corrected with respect to ₀ (a system in which both the exit pupil and the entrance pupil are located at approximately infinity). Therefore, the chief ray passing through the center point (the point through which the optical axis AX passes) on the pupil plane (Fourier transform plane) FP of the projection lens PL has the optical axis A on the reticle 7 side and the wafer W side.
It becomes almost parallel to X. Then, under the exposure wavelength λ ₀ , the pattern surface of the reticle 7 and the surface of the wafer W are arranged conjugate with each other with respect to the projection lens PL.

Regarding the alignment system, the objective lens 5 and the beam splitter 2 are provided above the dichroic mirror 6.
A plurality of alignment optical systems including the above are arranged so that the alignment beam from the alignment beam transmitting system 210 is guided to the objective lens 5 of each alignment optical system. The alignment light receiving system 212 is a system for photoelectrically detecting diffracted light (interference light) from the reticle mark or the wafer mark through the objective lens 5, and includes a monitor grating plate, a photoelectric element, and the like. Alignment signal processing system 2
Reference numeral 14 processes various signals photoelectrically detected to calculate the respective positions of the reticle mark and the wafer mark, or the positional shift amount, and sends the result to the central control system 216. The central control system 216 has a microcomputer or a minicomputer, and centrally manages communication with various control systems, sequences, parameter settings, error processing, and the like.

Next, the configuration of the alignment system according to the first embodiment of the present invention will be described with reference to FIGS. 6, 7 and 8.
In the first embodiment, based on the mark arrangement on the reticle shown in FIG. 2, two beams irradiating each of the reticle marks RMa and RMb and two beams irradiating the wafer mark WM through the window RW. The beam and the beam are separated from each other in the image space. Therefore, in FIG. 6 (A), the alignment beam light transmitting system 210 and the two beams LBr _1, LBr ₂ for the reticle mark, two beams for the wafer mark LBW _1, emits a LBW _2. The two beams LBr ₁ and LBr ₂ are reflected by the upper half of the beam splitter 2 as shown in FIG. 6 (A) and once intersected at the plane Pw via the lens system 3, the beam splitter 4, and the objective lens 5. After that, the mark RM on the reticle 7
Irradiate each of a and RMb.

As shown in FIG. 6B, two beams LB
r ₁ and LBr ₂ pass through an inclined optical path on the front focal plane (a plane conjugate with the pupil plane FP of the projection lens PL) Ep of the objective lens 5 and, when exiting from the objective lens 5, from the optical axis of the objective lens 5. The eccentric position is set to have a center line (bisector of the two beams LBr ₁ and LBr ₂ ) parallel to the optical axis.

On the other hand, the two beams L from the light transmitting system 210
Bw ₁ and LBw ₂ are reflected by the lower half of the beam splitter 2 as shown in FIG. 6 (A), and similarly, they cross the surface Pw once through the lens system 3, the beam splitter 4, and the objective lens 5, and then the reticle. The mark WM on the wafer W is reached through the upper window RW. Two beams LBw ₁ and LBw
_When viewed from the direction of FIG. 6 (B), ₂ passes through the focal plane Ep with an inclination, and when exiting from the objective lens 5, two beams L
Proceed in parallel with Br ₁ and LBr ₂ at regular intervals.

Here, the geometrical optical arrangement of each beam centered on the beam splitter 2 will be described in detail with reference to FIG. In FIG. 7, the splitter 19A is provided in the light transmitting system 210, and the two coherent beams LB ₁ and LB ₂ are incident while being inclined symmetrically with respect to the optical axis AXa of the alignment optical system (objective lens 5). .. The two beams LB ₁ and LB ₂ are the same as the beams LB ₁ and LB ₂ shown in FIG.
₁ is 633 nm of He-Ne laser. This beam L
B ₁ and LB ₂ cross the plane PF ₁ in the state of parallel light flux and then enter the splitter 19A, and the beam LB ₁ to the beam L
Two beams Br ₁ and LBw ₁ are produced, and _two beams LBr ₂ and LBw ₂ are produced from the beam LB ₂ . Further, the tilted state between the reticle beams LBr ₁ and LBr ₂ and the tilted state between the wafer beams LBw ₁ and LBw ₂ are equal to each other, and the tilted state between the original beams LB ₁ and LB ₂ is preserved. There is. When the four beams are Fourier transformed by the lens system 3, they are symmetrically positioned on the focal plane Ep with the optical axis AXa in the X direction (measurement direction).
There are two spots SP ₁ and SP ₂ . The spot SP ₁ is generated when both of the beams LBr ₁ and LBw ₁ intersect at a beam waist position.
_The beam ₂ is generated when the beams LBr ₂ and LBw ₁ both intersect at a beam waist position. Here, the plane PF where the original two beams LB ₁ and LB ₂ intersect
₁ is conjugate with the surface Pw in FIG. 6 and the wafer surface.

Now, the four beams emitted from the objective lens 5 are irradiated onto the reticle as shown in FIG. Figure 8
, Reticle marks RMa, RMb, and window RW
Is the same as in FIG. 2, and the measurement direction is the X direction.
The reticle beams LBr ₁ and LBr ₂ are irradiated as parallel light beams having a cross section (rectangle) of a size that covers only the reticle marks RMa and RMb, respectively, and the wafer beams LBw ₁ and LBw ₂ are only inside the window RW. Is emitted as a parallel light beam having a cross section (rectangle) of a size that passes through. In this embodiment, the original two beams LB ₁ and LB ₂
If the sectional shapes (rectangles) of (1) and (4) have the same size, the sectional shapes 56a and 56b of the four beams on the reticle all have the same size.

Two beams LBw ₁ , L for the wafer are also provided.
Bw ₂ is set so as not to reach the edge of the window RW as much as possible, and these two beams LBw ₁ and LBw ₂ intersect each other on the mark WM of the wafer W via the projection lens PL to generate an interference fringe. .. The arrangement of the mark WM is set so as to utilize the chromatic aberration of magnification as in the case of FIG. The interference beam (interference of ± first-order diffracted light) BTW vertically generated from the mark WM passes through the projection lens PL and passes through the central portion of the window RW of the reticle 7 as a parallel light flux having a substantially rectangular cross-sectional shape 56C. .. The size of the cross-sectional shape 56C is determined by the size and shape of the overlapping area between the shape of the irradiation area on the wafer and the outer shape of the wafer mark WM by the _two beams LBw ₁ and LBw ₂ . Here, the straight line La in FIG. 8 is the reticle mark R.
The center lines of the Ma, RMb and the window RW in the measurement direction (X direction) are represented, and the straight line La is arranged so that the direction of the arrow is substantially toward the center of the circuit pattern region. Further, a broken line Lb extending in the X direction between the window RW and the reticle marks RMa and RMb represents an edge image of an imaging reticle blind (illumination field stop) provided in the exposure illumination system 208. Exposure of exposure light to RMa and RMb is prevented. The shaded area in FIG. 8 is a light-shielding band made of a chromium layer or the like.

The interference beam BTW from the wafer mark WM passes through the objective lens 5 through the center point (the point where the optical axis AXa passes) in the focal plane Ep in FIG. 7, and further the lens system 3 and the beam splitter 2 are used. Point PC in the plane PF ₂ through the
It passes through w as a parallel light beam. This surface PF ₂ is a lens system 3
It is at a position conjugate with the plane Pw with respect to the combined system of the objective lens 5 and. Also, two beams LBw ₁ and LBw for the wafer
_Since both ₂ irradiate the wafer mark WM with a symmetrical incident angle, the 0th-order diffracted lights DF ₁ and DF ₂ by the respective beams travel backward in the respective optical paths of the _two beams LBw ₁ and LBw _2. .. These two 0th-order diffracted lights are DF ₁ , D
F ₂ is transmitted through the beam splitter 2 and the point P in the plane PF ₂
It becomes a parallel light flux that intersects at Cw.

On the other hand, the first-order diffracted light generated from each of the reticle marks RMa and RMb travels backward in the optical path of each of the light-transmitting beams LBr ₁ and LBr ₂ as described with reference to FIGS. Through the inside objective lens 5, the positions of the spots SP ₁ and SP ₂ in the focal plane Ep, the lens system 3, and the beam splitter 2, parallel light beams intersect at a point PCr in the surface PF ₂ . Since mutually intersect each other in the plane PF ₂ interfering beams BTW and 0-order diffracted light DF _1, DF _1, surface PF ₂
The cross-sectional shape of the intersection at is the light-transmitting beams LBw ₁ and LBw ₂
Reticle mark RM
a, the surface PF _{2 of} the portion where the first-order diffracted light from RMb intersects
The cross-sectional shape inside is substantially similar to the outer shapes of the marks RMa and RMb. Further, at or near the surface PF ₂ , the interference beam BTW from the wafer mark WM and the 0th-order diffracted light D
F ₁ and DF ₂ and the two first-order diffracted lights from the reticle mark RM are separated in the Y direction between the reticle mark RM and the window RW so that they are sufficiently separated in the Y direction (non-measurement direction) in FIG. 7. Increase the space.

From the wafer mark WM thus obtained,
Interference beam BTW, 0th-order diffracted light DF ₁, DF₂Is shown in FIG.
As shown in (B), the transparent portion 23B of the partial reflecting mirror 23,
Via a lens 15 and a lens system (Fourier transform lens) 17
The light is received by the photoelectric detector 18. This photoelectric detector 18
The light receiving surface is on the Fourier transform plane which is conjugate with the focal plane Ep. one
However, the two first-order diffracted lights from the reticle mark RM are partially
The light is reflected by the total reflection portion 23A of the reflecting mirror 23, and the surface PF₂Distributed to
Interference fringes cross on the placed monitor grid plate 13
Is generated. Re-diffraction from the monitor grid plate 13
The ± 1st-order diffracted light that has been transmitted travels coaxially and becomes an interference beam BTR.
Then, the light is received by the photoelectric detector 14.

The partial reflecting mirror 23 has a wavelength in this embodiment.
λ₁Pattern light of the reticle 7 (mark R
The bottom surface of the reticle on which M is formed)
Then, the beam LBr for the reticle on the reticle 7₁, LB
r₂And wafer beam LBw ₁, LBw₂Is separated
Therefore, even in the partial reflecting mirror 23,
The return light is separate and can be easily split.

In the above structure, when the reticle 7 is slightly moved in the X direction, the interference fringes generated on the monitor grating plate 13 by the _first-order diffracted lights DF ₁ and DF ₂ from the reticle marks RMa and RMb respectively have a pitch. In the direction (X direction), the phase relationship between the monitor grating and the interference fringes changes, and the photoelectric detector 14 obtains the signal Sm as shown in FIG. Similarly, when the wafer W is slightly moved in the X direction, the intensity of the interference beam BTW vertically generated from the wafer mark WM changes in a sine wave shape, and the signal Sw as shown in FIG. 3 is obtained. Therefore, in the alignment signal processing system 214 shown in FIG. 5, the reticle interferometer 202 and the wafer interferometer 206 detect the position of the reticle 7 and the position of the wafer W at which the levels of these signals Sw and Sm become predetermined values. Then, the reticle stage control system 200 and the wafer stage control system 204 control so that the relative position between the reticle 7 and the wafer W is driven into a desired relationship. After that, the illumination light IL is emitted from the exposure illumination system 208, and the image of the circuit pattern area of the reticle 7 is printed on a predetermined position on the wafer W.

As described above, in this embodiment, the two beams LBr ₁ and Lr which are irradiated on the marks RMa and RMa on the reticle 7 are used.
Since Br ₂ and the two beams LBw ₁ and LBw ₂ with which the mark WM on the wafer W is irradiated are separated so as not to mix with each other on the reticle 7, this is a detection signal of each mark. The waveforms of the photoelectric signals Sm and Sw become close to an ideal sine wave, and waveform distortion and noise components are eliminated by stray light mixing, so that the mark position measurement accuracy based on the levels of the signals Sm and Sw is improved. Is obtained. Further, when the reticle 7 and the wafer W are scanned by the homodyne method to generate the signals Sm and Sw as described above, the times at which the reticle mark RM and the wafer mark WM pass the respective predicted alignment end positions are aligned as much as possible. When the reticle stage RST and the wafer stage WST are simultaneously scanned, the error can be canceled by the vibration of the objective lens 5 and the like, and stable position detection can be performed.

Here, the beam arrangement on the reticle in this embodiment will be described again with reference to FIG. 9 is basically the same as that shown in FIG.
A mark pattern 52 for replacing M is provided. In FIG. 9, the distance KY between the window RW and the reticle mark RM in the Y direction is different from the mark arrangement in FIG. 2 described above.
Can be extended. That is, as shown in FIG. 2, if the interval KY is widened while keeping the beam cross-sectional shape that simultaneously covers the reticle mark RM and the window RW, the beam cross-sectional shape becomes extremely large. Further, since the replacement mark 52 enters the beam irradiation area, there is a possibility that diffracted light generated from the mark 52 may be mixed into the light receiving system as stray light. However, as in this embodiment shown in FIG. 9, by separating the beam into four beams on the reticle, all of these problems can be solved.
In the Y direction with respect to the replacement mark 52 window RW, Δ
They are formed apart by β, but this Δβ is the projection lens PL
Is the amount of lateral chromatic aberration on the reticle side. Mark 52
The portion of the distance ΔDB between the reticle mark RM and the reticle mark RM is a light-shielding band, and the edge image Lb of the reticle blind is located there, as shown in FIG. Therefore, the interval ΔDB
Is determined by the setting accuracy of the variable grade edge of the reticle blind and the Y-direction fine movement range during alignment of the reticle 7.

Here, the two beams LBr ₁ and LBr ₂
(Or LBw ₁ , LBw ₂ ) the center distance in the X direction is K
X, the pitch of the reticle marks RMa and RMb is Pgr, and the wavelength of the beam is λ ₁ , the two beams LBr ₁ and LBr
Incident angle of Br ₂ on reticle (alignment system optical axis A
The inclination from Xa) θr is expressed by the following equation from the condition (first-order diffraction angle is 2θr) that the first-order diffracted light from the mark RM travels backward along the transmitted beam.

Tan θr = KX / 2 · ΔL, and sin2θr = λ / Pgr As an example, the pitch Pgr is 10 μm and the wavelength λ is 633n.
When m and the axial chromatic aberration amount ΔL are 20 mm, the incident angle θr is about 1.81 ° and the interval KX is about 1.27 mm. By the way, if the number of grids of the reticle marks RMa and RMb is 20, the size of each mark in the X direction is about 200 μm (0.
2 mm), which is sufficiently smaller than the interval KX. Therefore, under the above conditions, the length of the window RW in the X direction may be set to about 1.6 mm, which is slightly larger than the value obtained by adding the distance KX to the length of the mark RM in the X direction.

In general, a beam from a coherent laser light source has a circular cross section in many cases, and a rectangular cross section as shown in FIG. 9 cannot be obtained as it is. Therefore, for example, two beams LB shown in FIG. _The four beams LBr ₁ , LBr ₂ , LBw ₁ and L are provided by forming an aperture plate having a rectangular opening on the plane PF ₁ where ₁ and LB ₂ intersect.
The cross-sectional shape of Bw ₂ may be shaped.

In actual alignment,
A reference mark (diffraction pattern) fixed on the e-stage WST
(Child-like) FM through projection lens PL and reticle 7
6 (B) to detect with the observation microscope 59,
Based on the mark on the chicle 7 (the window RW may be used instead)
With the wafer interferometer 206, the coordinate values of the reference mark FM
Detect and store. The microscope 59 is the same as the exposure light IL
Each mark shall be detected under the illumination light of the wavelength.
After that, the reference mark FM has two beams LBw. ₁, LB
w₂Wafer stage so that it is directly under the irradiation area Am
The signal Sw of the photoelectric detector 18 is set to a predetermined value by moving WST.
The wafer stage WST coordinate value at the point
If you memorize it, the reticle
2 beams LBw with reference to the projection point at the center of rule 7
₁, LBw₂The coordinate position of the predetermined phase point of the interference fringe is obtained by
Maru This coordinate position measurement is the so-called baseline measurement
It is called, and is created on the wafer as in this embodiment.
Wafer mer based on the static interference fringes formed
The method of detecting the position of the wafer
Prior to the peat exposure, for example, every wafer exchange, or
It needs to be executed for each lot.

By the way, beam transmission in the first embodiment
An example of a specific configuration of the system 210 is shown in FIG.
The configuration is suitable. Emits a laser beam with a wavelength of 633 nm
The parallel beam from the He-Ne laser tube 19M is a beam
Beam passes through the beam expander 19L that expands the diameter
It is divided into two by the splitter 19K. One side divided
Beam LB₁Is reflected by the mirror 19J and the lens 19
After being focused on the surface ES by G, it diverges. Another minute
The other beam LB that was split₂Is lens 19H, mirror 1
After condensing on the surface ES via 9F and 19E, it is diverging
Ku. The plane ES is a position conjugate with the focal plane Ep in FIG.
Two beams LB passing through the surface ES₁, LB ₂Each center line of
They are parallel to each other and have regular intervals. Two
Beam LB₁, LB₂Via the mirror 19D
Incident on the lens system (inverse Fourier transform lens) 19C,
Two from this lens system 19C that intersect at the rear focal position
Is emitted as a parallel light flux of. Rear focus of lens system 19C
Surface, that is, the surface PF in FIG.₁Shape the beam cross section to
Aperture plate 19B is placed, and the two that pass through here
Beam LB₁, LB₂Enters the divider 19A. Minute
The divider 19A is formed by laminating triangular prisms back to back.
The beam splitter (half mirror)
Beam, which is an ordinary rectangular parallelepiped beam.
It may be a splitter. In this embodiment, the divider 19A is joined.
The optical axis AXa of the alignment system is set to match the surface
And two beams LB₁, LB₂Is eccentric from the joint surface
It is set so that it will be incident on the slope at a different position. By this
The splitter 19A has four beams LBr separated from each other.
 ₁, LBr₂, LBw₁, LBw₂Inject.

In the above structure, the aperture plate 19B
Also has the function of making the intensity distribution uniform within the beam cross section. Usually, the He-Ne laser light has a Gaussian distribution intensity in its cross section, so the enlargement ratio of the expander 19L is set so that the rectangular aperture area of the aperture plate 19B is near the center (peak) of the Gaussian distribution. Then, both unnecessary sides of the Gaussian distribution can be cut.

Next, the structure of the beam transmitting system according to the second embodiment of the present invention will be described with reference to FIG. In this embodiment, the configuration of the first embodiment is used except for the beam transmitting system. This embodiment is different from the first embodiment (FIG. 10) in that the two beams for the reticle and the two beams for the wafer are made from different laser light sources.

FIGS. 11A and 11B show four beams LBr incident on the beam splitter 2 shown in FIG.
₁ , how to make LBr ₂ , LBw ₁ , LBw ₂ ,
The X, Y, and Z coordinate systems are aligned with FIG. See also FIG.
1 (B) is a view on arrow B-B in FIG. 11 (A). Beam LB emitted from laser light source (He-Ne, etc.) 24A
The light splitter 25A splits w into two parallel beams LBr ₁ and LBr ₂ . Beam LBr ₁ , L
In FIG. 11A, Br ₂ are arranged in a direction perpendicular to the paper surface, and further, are deflected by a certain amount with respect to the optical axis AXa by the D prism (wedge prism) 26A, and the mirror 2
After being reflected by 7A, it enters the combining prism 27B.
Beams LBr ₁ , LB reflected by the combining prism 27B
As shown in FIG. 11B, r ₂ is converted by the lens system 28 into parallel light beams that intersect at the surface PF ₁ . However,
As shown in FIG. 11A, the two beams LBr ₁ and LBr
Br ₂ is decentered from the optical axis AXa by a predetermined amount in the YZ plane.

On the other hand, the beam LBw emitted from the laser light source 24B is split into two beams LBw ₁ and LBw ₂ by the light splitter 25B having exactly the same structure as the light splitter 25A, and these two beams are D prisms. After being simultaneously deflected by 26B, they are incident on the combining prism 27B, and are converted by the lens system 28 so as to intersect with the plane PF ₁ in the XY plane.

In FIGS. 11A and 11B, the D pre
The position of the beam deflection point by the prisms 26A and 26B.
It is substantially aligned with the front focal plane of the system 28. Lens again
All four beams emitted from the system 28 are parallel light beams.
Therefore, in practice, the D prism 26A, 26B and the laser
Insert a lens system between each of the light sources 24A and 24B
At the position of the D prism (front focal plane of lens system 28)
To create a focal point (beam waist) for each beam
ing. Further face PF₁The position is shown in Figure 10.
An unaperture plate 19B is arranged. However, the beam
LBr₁, LBr ₂Intersection area and beam LBw₁, LB
w₂PF with the intersection area of₁The position on the optical axis AXa
Since it is sandwiched and separated in the Z direction, the aperture plate 19B
Two rectangular openings (illumination field stop)
Is formed. Therefore, two beads for the reticle
Mu LBr₁, LBr₂Cross-sectional shape and dimensions on the reticle
And two beams LBw for the wafer₁, LBw₂Reticles
Independent adjustment of cross-sectional shape and dimensions on the vehicle
You can

In this embodiment described above, the two beams for the reticle and the two beams for the wafer are produced from different laser light sources, so that the same He-Ne having a wavelength of 633 nm is used.
Even if a laser is used, the signals do not interfere with each other, and among stray light components that enter the light receiving system, the signals caused by mutual interference between stray light generated by the reticle beam and stray light generated by the wafer beam. Inconvenient phase distortion on the waveform can be reduced. of course,
The polarization state of each beam from the two laser light sources 24A and 24B may be different from each other, and the signal light from each mark may be polarized and separated also in the light receiving system.

The beams from the two laser light sources 24A and 24B may be slightly (for example, 1% or less) different in wavelength. However, when the amount of axial chromatic aberration and the amount of lateral chromatic aberration of the projection lens PL differ due to a slight difference in the wavelength, the surface PF _{1 in} FIG.
You can change the state of each beam in. Next, a modification of the beam transmitting system according to the third embodiment of the present invention will be described with reference to FIG. The system of FIG. 12 is equivalent to the system described in the previous example of FIG. 1, and the beam from the laser light source 24 is split into two beams by the beam splitter 33, and one beam LB ₁ thereof is split by the mirror 35. After being reflected, it enters the combining prism 36. The other split beam LB ₂ is reflected by the mirror 34 and then enters the combining prism 36. The lens system 28 combines the _two beams LB ₁ and LB ₂ that are combined in parallel with each other by the combining prism 36 into the plane PF ₂
Cross on top. In this embodiment, a light splitter 37 (same type as the splitters 25A and 25B in FIG. 11) is provided at the position of the plane PF ₂ , where two beams LBr ₁ and LBr for the reticle are provided.
₂ and two beams LBw ₁ and LBw ₂ for the wafer. The light splitter 37 corresponds to 19A in FIG.

Next, a beam according to a fourth embodiment of the present invention
A modified example of the light transmission system will be described with reference to FIG. 13 and FIG. 6 described above.
To do. In the present embodiment, the homoda of the first to third embodiments described above is used.
The heterodyne method is applied to the in method
Of. First, the system shown in FIG. 13 is exactly the same as the system shown in FIG.
The difference is that the frequency shift for heterodyne is
Covers 38 and 39 are divided into two beams LB₁, L
B₂It is to provide for each of. This frequency shift
The data 38 and 39 are sounds driven by a predetermined high frequency signal.
Acoustic optical modulator (AOM) and 1 modulated by this AOM
It is composed of a slit or the like for extracting only the next-order diffracted light. So
And the frequency shifter 38₁Is driven by
Shooting beam LB₁Frequency f₀Against f₀+ F₁Only
Frequency f₁The first-order diffracted light of
Force Similarly, the frequency shifter 38 uses the frequency F₂Dry in
Incident beam LB₂Frequency f₀Against f₀+
F ₂Frequency f shifted by₂Transmits the first-order diffracted light of
Output as Therefore, it is emitted from the light splitter 37.
Beam LBr₁, LBw₁Is the frequency f₁Become the beam
LBr₂, LBw₂Is the frequency f₂become.

By the way, since the reference signal is required to apply the heterodyne method, the four beams are partially branched at the position of the beam splitter 4 in the optical path of the alignment system shown in FIG. It is made incident on the system 8. As a result, the two beams LBr ₁ and LBr ₂ become parallel light beams and intersect on the grating formed in the upper half of the reference grating plate 20, and the two beams LBw ₁ and LBw ₂ also become parallel light beams and become the reference grating. They intersect on a grid formed in the lower half of the plate 20. From the reference grating plate 20, an interference beam BTSr made of two beams for a reticle and an interference beam BTSw made of two beams for a wafer are separately generated, and these are generated by a photoelectric detector. The light is individually received by 21 and 22. Then, in the hetero-die measurement, the phase difference Δφ _m between the reference signal SRm obtained from the photoelectric detector 21 and the signal Sm obtained from the photoelectric detector 14 is obtained as shown in FIG. 4 and obtained from the photoelectric detector 22. Reference signal SR
Phase difference Δ between w and the signal Sw obtained from the photoelectric detector 18
Find φ _w . These signals Sm, Sw, SRm, SR
Each of w is a beat frequency having a frequency Δf (= f ₁ −f ₂ ). Further, the phase difference Δφ _m obtained here changes by 2π (one cycle) with respect to the positional deviation of 1/2 of the pitch Pgr of the reticle marks RMa and RMb, and the phase difference Δφ _w
Also changes by 2π with respect to the positional deviation of 1/2 of the pitch Pgw of the wafer mark WM.

Next, the alignment sequence of this embodiment will be briefly described. First, the reticle 7 and the mark FM of the reference plate on the wafer stage WST are simultaneously observed by the microscope 59 (FIG. 6) having the same wavelength λ ₀ as the exposure light as the illumination light.
Either or both of reticle stage RST and wafer stage WST are driven so that the mark on reticle 7 and reference mark FM are not displaced relative to each other. At this time, the same lattice mark as the wafer mark WM is juxtaposed on the reference plate, and it is arranged so as to be located immediately below the window RW of the reticle 7. Then, when the reticle 7 and the reference mark FM are aligned by the microscope 59, the phase difference Δ between the measurement signal Sm and the reference signal SRm.
φ _mo , the measurement signal Sw, and the phase difference Δφ _wo between the reference signal SRw (detecting the grid of the reference mark) and the reference signal SRw are obtained and stored. After that, when the wafer W is actually aligned, it is obtained by detecting the reticle marks RMa and RMb while the wafer stage WST is positioned at the target position. The phase difference Δφ _m and the phase difference Δφ _w obtained by detecting the wafer mark WM are the values Δφ stored in advance.
Reticle stage RST or wafer stage WST is controlled so that _mo and Δφ _wo are equal. Alternatively, since the phase difference change amount and the stage position change amount uniquely correspond to each other, assuming that the reticle mark and the wafer mark have a double pitch relationship, (Δφ _m −Δφ _mo ) / 2
Either or both of the reticle stage and the wafer stage may be controlled so as to satisfy the relationship of = Δφ _w -Δφ _mo . Also in the present embodiment, the reticle interferometer 202 and the wafer interferometer 206 are used to determine the reticle mark RM and the wafer mark WM by their grating pitch Pgr,
It goes without saying that the phase difference measurement is performed after positioning within ± 1/4 of Pgw.

Next, a modification of the beam transmitting system according to the fifth embodiment of the present invention will be described with reference to FIG.
The system of No. 4 is basically similar to the system shown in FIG. However, in this embodiment, the number of laser light sources 24 is one, and a predetermined frequency shift is provided between the beam for the reticle and the beam for the wafer, so that stray light from the reticle side and stray light from the wafer side are not generated. The coherence control of is reduced.

FIG. 14A shows the arrangement in the Z-Y plane, and FIG. 14B shows the arrangement in the XY plane.
It is the figure seen from the lower side. In FIG. 14, the same members as those used in FIG. 11 are given the same reference numerals. The beam from the laser light source 24 is transmitted by the beam splitter 40 to the beam LBr for the reticle and the beam L for the wafer.
And Bw. Of these, the beam LBw is the mirror 4
Two beams LBw reflected by the prism 25B
₁ and LBw ₂ , and then intersects at the plane PF ₁ through the D prism 26B, the compound prism, and the lens system 28. On the other hand, the beam LB transmitted through the beam splitter 40
The r is incident on the frequency shifter 42, where a beam of the first-order diffracted light shifted by a certain frequency is formed, and the light splitter 25
The beam A is further divided into _two beams LBr ₁ and LBr ₂ . The two beams pass through the D prism 26A, the mirror 27A, the combining prism 27A, and the lens system 28, and intersect at the surface PF ₁ .

In this embodiment, the frequency shifter 42 is placed on the reticle beam LBr side, but it may be placed on the wafer beam LBw side instead. In addition, the frequency shifter 42
May be driven with an appropriate high frequency signal for the AOM, and the frequency Fd may be any value. As described above, in this embodiment, the two beams LBr ₁ and LBr on the reticle side are provided.
₂ and the two beams LBw ₁ and LBw ₂ on the wafer side, a frequency difference Fd is given, but the frequency difference between the two beams intersecting in the image space is zero. The method is applied as the homodyne method as in the above example. Therefore, in alignment, it is necessary to scan the reticle mark RM and the wafer mark WM in the pitch direction.

Next, a modification of the beam transmitting system according to the sixth embodiment of the present invention will be described with reference to FIG. This Figure 1
The system of No. 5 is a combination of the systems of FIGS. 13 and 14 described above for heterodyneization. Further, FIG. 15B is a view of the arrangement of FIG. 15A viewed from the lower side. The beam from the laser light source 24 is split into a reticle beam LBr and a wafer beam LBw by the beam splitter 40 as shown in FIG. The wafer beam LBw is reflected by the mirror 41, then enters the beam splitter 33W, and is further divided into _two beams LBw ₁ and LBw ₂ as shown in FIG. 15B. Of these, the beam LBw ₁ passes through the frequency shifter 38W and the mirror 35W and is incident on the decentering synthesis prism 36W. The other beam L
Bw ₂ is reflected by the mirror 34W, and the frequency shifter 39
The light passes through W and enters the decentered synthesis prism 36W. The two beams LBw ₁ and LBw ₂ from the prism 36W are deflected in the same direction by the D prism 26B, pass through the combining prism 27B and the lens system 28, and intersect at the surface PF ₁ .

On the other hand, the beam LBr for the reticle is shown in FIG.
5 (B), that is, the beam splitter 33
The two beams L pass through the R, the mirror 34R, the frequency shifters 38R and 39R, and the decentering synthesis prism 36R.
After being converted into Br ₁ and LBr _2, they are simultaneously deflected by a certain amount by the D prism 26A, enter the combining prism 27B via the mirror 27A, and intersect at the surface PF ₁ .

As described above, in this embodiment, the two frequency shifters 38R and 39R driven by the high frequency drive circuit are provided between the two reticle beams LBr ₁ and LBr _2.
Is used to provide a predetermined frequency difference Δfr, and two frequency shifters 38W and 39W provide a predetermined frequency difference Δfw between the two wafer beams LBw ₁ and LBw ₂ .
Here, four frequency shifters 38R, 39R, 38W, 3
Each drive frequency of 9W is Fr ₁ , Fr ₂ ,
When Fw ₁ and Fw ₂ are set, the following relationship is established with the original frequency of the beam from the Rege light source 24 being f ₀ .

Δfr = (Fr ₁ + f ₀ ) − (Fr ₂ + f ₀ ) = Fr ₁ −Fr ₂ Δfw = (Fw ₁ + f ₀ ) − (Fw ₂ + f ₀ ) = Fw ₁ −Fw ₂ Then, for each drive signal If you set the frequency appropriately,
The frequency difference Δfr and Δfw can be made different, and as a result, the interference beam BTR obtained from the reticle mark RM can be obtained.
Can be sufficiently separated from the beat frequency (Δfr) of the interference beam BTW obtained from the wafer mark WM, and even if they are mixed with each other, they can be discriminated on the signal processing circuit. It will be possible. Of course, according to the configuration shown in FIG. 6, the frequencies of the reference signals SRr and SRw are naturally different accordingly.

Here, the frequencies Δfr and Δfw are of course set within the range of the responsivity of the photoelectric detectors 14, 18, 21, and 22, and the ratio Δfr / Δfr is 1 time or the harmonics. It is desirable to set so that the relationship (1: 2, 1: 3, etc.) does not occur. Further, the center frequencies Fr ₁ and Fr ₂ of the two drive signals for the reticle beam are Frc.
[(Fr ₁ + Fr ₂ ) / 2] and the center frequencies F of the _two drive signals Fw ₁ and Fw ₂ for the wafer beam.
It is desirable to set the difference Δfc from wc [(Fw ₁ + Fw ₂ ) / 2] to a frequency sufficiently higher than the responsivity of the photoelectric detectors 14, 18, 21, and 22, that is, a value that is too high to respond. .. As an example, the frequencies Fw ₁ and Fw ₂ are 80
MHz range, frequencies Fr ₁ and Fr ₂ are 90 MHz (or 70 M
Hz), Fw ₁ = 80.000 MHz, Fw ₂ = 80.020 MHz (Δ
fw = 20 KHz) Fr ₁ = 90.000 MHz, Fr ₂ = 90.050 MHz (Δ
fr = 50 KHz).

In this case, the ratio Δfr / Δfw is 2.5, the period of the frequency Δfr (50 KHz) is 20 μSec., And the period of the frequency Δfw (20 KHz) is 50 μSec. Therefore, their common multiple is N. When set to an integer of 1 or more, 100
・ N (μSec.). Therefore, it is advisable to set the digital sampling time of each signal Sm, Sw, SRm, SRw for measuring the phase difference in the signal processing system to 100.N (μSec.). Since this is also related to the signal processing method, the processing method will be described below.

FIG. 16 shows a circuit configuration for capturing the waveform of each signal, and here, for simplification of description, only the capturing of the signals Sw and SRw is shown. Figure 1
6, high-speed analog-to-digital converter (ADC)
100 converts the magnitude of the signal Sw into a digital value and sends the value to a memory (RAM) 102. The conversion timing of the ADC 100 is performed in response to the clock pulse CK from the arithmetic processing circuit (CPU) 110. This pulse CK
The frequency is determined according to the conversion speed of the ADC 100, but is determined to be twice or more the frequency Δfw of the signal Sw. Here, the ADC 100 that can be A / D converted in response to the 200 KHz clock pulse CK is used. Further, the counter (CNT) 108 has a clock pulse C
K is sequentially counted, and the count value is output as the address value AD for the RAM 102.

Similarly, the magnitude of the reference signal SRw is ADC.
RAM 10 digitally sampled by 104
6 is stored. Both the conversion timing of the ADC 104 and the address change of the RAM 106 respond to the clock pulse CK. CPU110 is RAM102,
The signal waveform stored in 106 is arithmetically processed to obtain a phase difference Δφ _w between the two signals Sw and SRw, and this phase difference Δ
The positional deviation amount ΔX _W or ΔY _W ) corresponding to φ _W is output to the central control system 216 shown in FIG.

Now, the CPU 110 causes the RAMs 102 and 10 to operate.
When processing the waveform data of 6, the common multiple 100 previously defined
The waveform data is read by the number of sampling points on the waveform to be used for the calculation based on N (μSec.) And the frequency Fck (for example, 200 KHz) of the clock pulse CK. First,
The period Tc of the frequency Fck is 10 ⁶ / Fck (μSec.), And when Fck = 200 KHz, Tc = 5 (μSec.). Therefore, it is only necessary to read the waveform data of at least 20 points (20 pulses of the clock pulse CK) from the ratio (100 / Tc) to the minimum value 100 where N = 1 among the common multiples. However, since the data of 20 points corresponds to only two cycles of each waveform of the signals Sw and SRw, the averaging effect in the calculation is reduced. Therefore, if 20 cycles of the waveforms of the signals Sw and SRw are used in consideration of the averaging effect, N = 10. At this time, the common multiple is 1000 (μSec.) And the number of points is 200 (1000 / Tc).

Therefore, the CPU 110 has the RAMs 102, 10
It is assumed that the same sampling point is set as a start point on each of the six waveforms, data of 200 points is read from this point, and the following calculation is performed. In this embodiment, Fourier integration is performed in order to obtain the phase difference between fundamental frequencies (sinusoidal waves) having defined sections. First, when the reference signal SRw is represented by a mathematical expression, the following expression is obtained with the amplitude being E _o .

Signal SRw: E _o sin (ωt + Δθ _o ) ... (1) Here, the angular velocity ω is uniquely known from Δfw. The phase component Δθ _o is a phase shift when the start point of the waveform portion of the signal SRw to be read is used as a reference. Then C
The PU 110 has a sine wave Asin ωt and a cosine wave Acos ωt with the same angular velocity ω and an amplitude A with zero phase shift at the start point.
And a waveform data table (same sampling interval as the signal SRw) are prepared, and the number of sampling points k of the waveform is set to 1 to
When n is set, the following calculation is executed

[0064]

[Equation 1]

In the present embodiment, the integration interval of these equations (2) and (3) is 2 from the start point address in the RAM 106.
00 points (n = 200. Expressions (2) and (3))
The values Dr and Di obtained in step 1 are components in each coordinate axis direction when the signal SRw is vectorized on the polar coordinate system with the start point as a reference. Alternatively, they are the real part (Dr) and the imaginary part (Di) on the complex plane.

Therefore, the CPU 110 uses the phase component Δθ _o with the start point of the sampled signal SRw as a reference.
Is calculated by the following formula. Δθ _o = tan ⁻¹ (Di / Dr) (4) Similarly, for the sampling signal Sw, the CPU 110 also has a phase component Δθ _{w based} on the start point of itself.
Is calculated according to the same method as the above equations (2), (3) and (4). Finally, the CPU 110 determines the signal Sw for the signal SRw from the difference between the two phase components Δθ _o and Δθ _w.
Find the phase difference Δφ _w of.

Δφ_w= Δθ_o−Δθ_w(5) This phase difference Δφ_wIs the lattice pitch P of the wafer mark WM
Immediately calculate the position shift amount (ΔXw, ΔYw) based on gw
Can be converted. However, the reticle and wafer are relatively aligned.
, The phase difference Δφ on the wafer side_wAnd retik
Phase difference Δφ _mOnly need to monitor relationship with
Therefore, the positional deviation amount (ΔXw, ΔY
There is no need to convert up to w).

As described above, in each embodiment of the present invention, the two beams LB for irradiating the reticle marks RMa and RMb are used.
The two r ₁ , LBr ₁ or the two beams LBw ₁ , LBw ₂ for irradiating the wafer mark WM are said to be coherent to each other, but they have complementary polarization states without coherence. You can use the same beam. In this case, interference fringes due to the intersection of the two beams are not formed on the wafer mark WM, the reference grating plate 20, or the monitor grating plate 13, and the interference beat light BTW from the wafer mark does not become an interference beat. , Only two first-order diffracted lights having complementary polarization states travel coaxially. However, an analyzer (λ / 2 plate, polarization beam splitter, etc.) is provided immediately before each photoelectric detector 14, 18, 21, 22 and
If the polarization states of two complementary diffracted lights are converted so as to include components that interfere with each other, an AC signal whose amplitude changes at the same beat frequency can be obtained.

In addition to the method of reducing the coherence between the reticle side and the wafer side by frequency shift using the four AOMs shown in FIG. 15, in addition to the beam light source on the reticle side and the wafer side as shown in FIG. The beam light source may be separately provided.

[0070]

According to the present invention, since the two first beams for aligning the mask and the two second beams for aligning the substrate are separated on the mask, it is difficult to mix the respective signal lights. There is. Also, the first of the two
Since the beam and the two second beams are emitted from the same objective lens, even if the objective lens moves, the behaviors of both beams match and the positional deviation due to the movement of the beams is offset. In this respect, the same advantages as those obtained without separation are obtained. Further, according to the second embodiment of the present invention, since the first beam and the second beam are emitted from different lasers, the beams do not interfere with each other.

In the other embodiments, the first beam and the second beam are produced from the beams emitted from the same laser, so that even if the laser beam emission angle changes, they are canceled. Furthermore, in the fourth and sixth embodiments, because of the heterodyne system,
Alignment is possible while the object is stationary, and alignment continues during exposure. In the fifth embodiment, the frequencies of the first beam and the second beam are different, and by setting this difference to be equal to or higher than the response speed of the signal detection circuit including the photoelectric converter, stray light generated by the light transmission of both beams is generated. It can be disregarded even if they interfere with each other and interfere with each other.

In the sixth embodiment, in addition to the effect of the heterodyne system, the beat frequencies of the two alignment signals are made different so that even if signals are mixed, they can be separated at the time of phase detection. This is because, if the phase detection time (waveform sampling time) is set to a common multiple of both signal periods as described above, both signals mathematically have an orthogonal relationship in that section.

Further, in this embodiment, the transmissive portion is formed on the mask, but this may be used as a reflective portion to irradiate the alignment beam from the back surface of the mask. Also, not a projection lens,
It may be a reflection type projection system, a contact, a proximity, or the like.

[Brief description of drawings]

FIG. 1 is a diagram showing a configuration of a conventional two-beam interference type alignment system.

FIG. 2 is a diagram showing a relationship between a conventional reticle mark arrangement and an irradiation area of two beams.

FIG. 3 is a diagram showing an example of a signal waveform obtained by photoelectrically detecting a wafer mark or a reticle mark.

FIG. 4 is a reference signal obtained by the heterodyne method (S
R) and mark signals (S _M , S _W )

FIG. 5 is a diagram showing an overall configuration of a projection exposure apparatus to which each embodiment of the present invention is applied.

FIG. 6 is a diagram showing the overall configuration of an alignment system according to the first embodiment.

FIG. 7 is a diagram showing a beam backlit state in the vicinity of a beam splitter 2 in FIG.

FIG. 8 is a diagram showing an arrangement of marks and windows on a reticle applied to each embodiment of the present invention.

9A and 9B are views for further enlarging and explaining the marks in FIG. 8 and the arrangement of windows and beam irradiation regions.

10 is a diagram showing a specific configuration of the beam transmission system shown in FIG.

FIG. 11 is a diagram showing a configuration of a beam transmission system according to a second embodiment.

FIG. 12 is a diagram showing a configuration of a beam transmission system according to a third embodiment.

FIG. 13 is a diagram showing a configuration of a beam transmission system according to a fourth embodiment.

FIG. 14 is a diagram showing a configuration of a beam transmission system according to a fifth embodiment.

FIG. 15 is a diagram showing a configuration of a beam transmission system according to a sixth embodiment.

FIG. 16 is a diagram showing an example of a circuit for sampling a signal waveform in the heterodyne method.

[Explanation of symbols for main parts]

5 Objective lens 7 Reticle 14, 18, 21, 22 Photoelectric element 19A Light splitting prism 19M, 24, 24A, 24B Laser light source PL Projection lens W Wafer WM Wafer mark RMa, RMb Reticle mark RW Reticle window LBr ₁ , LBr _{2 For} reticle Transmitted beam LBw ₁ , LBw ₂ Transmitted beam for wafer

Claims (5)

[Claims] 1. A projection optical system for imaging and projecting a pattern of a mask onto a photosensitive substrate under exposure light, a diffraction grating-shaped mask mark formed on the mask, and diffraction formed on the photosensitive substrate. Beam irradiation means for irradiating each of the lattice-like substrate marks with two illumination beams for alignment so as to be incident at symmetrical angles, diffracted light from the mask mark, and the beam generated from the substrate mark. A light receiving unit that receives the diffracted light that passes through the projection optical system and outputs a photoelectric signal according to the intensity of each diffracted light, and a relative positional deviation between the mask and the photosensitive substrate based on the photoelectric signal. In the projection exposure apparatus including a means for detecting, the mask has a region transparent to the illumination beam at a position near or adjacent to a direction perpendicular to the grating period direction of the mask mark; Beam irradiation means,
A first beam transmission system that creates two first beams that intersect at a predetermined angle to illuminate the mask mark; intersect at a predetermined angle to illuminate the substrate mark through the transparent region of the mask A second beam transmitting system for producing two second beams for controlling the two beams; the two first beams and the second beam for two years
And an objective optical system for injecting the first beam and the second beam so that the irradiation regions of the first beam and the second beam are separated from each other on the mask. 2. The beam irradiation means, a laser light source for outputting the beam, a first splitter for splitting the beam from the laser light source into two, and a light-sending condition for the two split beams. And a second splitter for splitting the two beams into two while maintaining the above, and the two beams split by the second splitter are transmitted to the first and second beams respectively. Device according to claim 1, characterized in that it leads to a system. 3. A substrate, which is provided with a frequency shifter for providing a predetermined frequency difference between the two beams in an optical path between the first divider and the second divider, and which is received by the light receiving means. 3. The amplitude modulation is applied to each of the interference light intensity of the diffracted light from the mark and the interference light intensity of the diffracted light from the substrate mark at a frequency according to the predetermined frequency difference. apparatus. 4. The beam irradiation means divides a laser light source for outputting the beam, and a beam from the laser light source into a first beam transmission system and a second beam transmission system. And a frequency shifter that gives a frequency difference between the two divided beams is provided in at least one of the first and second beam transmitting systems, and the diffracted light from the mask mark and the diffracted light from the substrate mark are The device according to claim 1, characterized in that a frequency difference is provided between and. 5. The beam irradiating means divides a laser light source that outputs the beam into a beam from the laser light source into a first beam sending system and a second beam sending system. A splitter; the first beam delivery system includes a second splitter for obtaining the two first beams, and
A first frequency shifter for providing a frequency difference Δfr between the two first beams; the second beam transmission system includes a third splitter for obtaining the two second beams, and the two beam splitters. Second
The second frequency difference Δfw different from the aforementioned Δfr is given between the beams.
The frequency shifter of the first beam so as to provide a frequency difference ΔFc different from both Δfr and Δfw between the average frequency of the two first beams and the average frequency of the two second beams. 2. The apparatus according to claim 1, further comprising a drive circuit for driving the second frequency shifter.