JPH0448203A - Exposure device - Google Patents

Abstract

PURPOSE:To realize alignment with high accuracy and a high throughput by performing simultaneous alignment on a reticle and a wafer with two pieces of luminous flux for alignment which can interfere with each other by using the same interference fringes. CONSTITUTION:An alignment grating 11a provided on the wafer 10 at a specific position is irradiated with the two pieces 51 and 52 of luminous flux which differ slightly in frequency and interfere with each other through a window 25a provided to the reticle 20 and a reducing projection lens 40 so that interfer ence fringes 54 are formed on the alignment grating 11a. At the same time, a alignment grating 21a provided on the reticle 20 at a specific position is also irradiated so that interference fringes 53 are formed on the positioning grating 21a. The reticle 20 or wafer 10 is displaced and aligned. The positioning grating 21a on the reticle 20 nd the positioning grating 11a on the wafer 10 can be aligned at the same time by using the same interference fringes 53 and 54. Consequently, the high-throughput alignment is realized.

Description

DETAILED DESCRIPTION OF THE INVENTION Field of Industrial Application The present invention relates to an exposure apparatus for semiconductor devices and the like having a fine pattern of 1 micron or less submicron rule. Even though the density of each device is becoming higher and higher, and the dimensions of the fine patterns of each element are less than 1 micron, the conventional method of aligning the photomask and LSI wafer during LSI manufacturing is done using alignment marks provided on the wafer. This was done by overlapping the marks on the photomask and the marks on the wafer by rotating and parallelly moving the stage on which the wafer was mounted. When forming elements, the alignment accuracy is poor and it is not practical.

Therefore, in order to improve the accuracy of alignment between the reticle and the wafer, for example, lf:, USP 4,771,
As disclosed in No. 180, a method using two-beam interferometry has been proposed. In this method, as shown in FIG. 16, an alignment grating 121 provided on the reticle 120 is used to align the wafer 110 with a reticle 120 on which a circuit pattern is formed.

This alignment grating 121 is irradiated with a light beam 122 emitted from a light source for alignment. Alignment grid 1
A light beam 122 irradiated onto the wafer 110 placed on the wafer stage 150 is diffracted by the alignment grating 121, and the diffracted light 112P passes through the alignment optical system 130 and the reduction projection lens 140 to the wafer 110 placed on the wafer stage 150.
projected on top. Reference numeral 1 is light for circuit pattern exposure, which projects the circuit pattern of the reticle onto the wafer through a lens 140. The positioning optical system 130 is shown in FIG. 17(a).
As shown in &- a pair of Fourier transform lenses 131 and 132, a spatial filter 133 . three mirrors 134
．． It has an angle of 135°136. Positioning optical system 1
301L Alignment grid 121 on reticle 120
A pair of Fourier transform lenses 131 and 132 . By the spatial filter 133, ±
Only the primary beams 123 and 124 are selected and intersected at the back focal position of the Fourier transform lenses 131 and 132 to generate interference fringes, and then the wafer 110 is projected by the reduction projection lens 140.
A grating 111 for alignment is provided both on the wafer 110 and on the wafer 110, and the light 200 +1 that is diffracted and interfered with by this grating 111 is the reduced projection lens 140.
and the photodetector 16 through the alignment optical system 130.
0 (see Figure 17(a)), and
Based on the interference light detected by the photodetector 160, the positional deviation between the grating 111 and the two-beam interference fringes 125 on the wafer 110 is detected, and based on the detection result, the wafer 110 is
The wafer 110 and the reticle 120 are aligned by being moved relative to the reticle 120, and after such alignment, the circuit pattern on the reticle 120 is exposed onto the wafer 110 with exposure light. In this way, in the method using two-beam interferometry, the grating 121 on the reticle 120 and the grating 11 on the wafer 110
1 and force <, 2 Since the reticle 120 and the wafer 110 are aligned with high precision using the beam interference fringes 125, the problem to be solved by the invention is however, with such a method. , reticle 120
An alignment optical system 130 must be provided between the lens and the reduction projection lens 140.
This requires a large number of optical members such as two mirrors 134 and 136, and the configuration is complicated. , etc. Also, there is a problem that the positional force of each optical member in the alignment optical system 130 changes over time.
is fixed to the reticle 120, so when exposing the circuit pattern 129 on the reticle 120 to the wafer 110, there is a problem in that the exposure range is narrowed by the positioning optical member 130.As described above, this method c'L reticle 120 and wafer 11
0 can be aligned with high precision. However, in an actual semiconductor process, since circuit patterns are exposed in an overlapping manner using many different types of reticles, it is necessary to form the alignment grating 121 at a different position for each reticle. The formation position of the grating 121 on the reticle 120 is as shown in FIG.
Areas 127a, 127 that are outside the exposure area 128 of the circuit pattern and are irradiated with light beams for alignment.
However, there is also the problem that the circuit pattern formed on the wafer 110 is limited. Furthermore, the alignment between the reticle 120 and the wafer 110 may be performed by moving the wafer stage 150 based on changes in the optical output of the interference light diffracted from the alignment grating 111 on the wafer 110 as the wafer stage 150 moves. This number
There is a risk that the S/N ratio of the interference light detection signal will decrease due to the vibration of the wafer stage 150, and there is also the problem of low throughput.Instead of this method, reduction projection using optical heterodyne interference may be used. Method for aligning reticle and wafer in exposure method<, 36th
Rejuvenated Proceedings of the Japanese Society of Applied Physics (1989), la-ni-
8, P., 560 (as shown in FIG. 19), the laser light emitted from the He-Ne2 frequency laser 230 is the light that is irradiated onto the alignment grating 221 on the reticle 220. wafer stage 250
Alignment grid 211 on wafer 210 placed on top
One of the lights switched by the shutter 260 is irradiated onto the alignment grating 221 on the reticle 220, and the other light is switched through the slit 222 provided on the reticle 221. After passing through the reduction projection lens 240
and is irradiated onto the alignment grid 211 on the wafer 210. The ±1st-order light ζ irradiated and diffracted by each alignment grating 211 and 221 is used as a color correction lens 271, a lens 272. Polarizing beam splitter 27 friend lens 27
East polarizing plate 275. The beams sequentially pass through the slits 276 and enter the photoelectric conversion element 277. Based on the phase difference between the respective interference lights detected by the photoelectric conversion element 277, the reticle 220 and the wafer 210 are aligned. However, with this method, there is a problem that the throughput is reduced because the light irradiated to the alignment grating 221 on the reticle 220 and the light intensity irradiated to the alignment grating 211 on the wafer 210 are switched by the shutter 260. Also, with this method, light with a polarization plane unnecessary for alignment is mixed into the alignment light,
Although it has been reported that the alignment accuracy decreases, the purpose of the present invention (Table 1) is to achieve alignment with high precision and high throughput by simultaneously aligning the reticle and wafer using the same interference fringes. In addition, by providing the alignment optical system on the reticle, the misalignment caused by the displacement of the alignment optical system, which was seen in the conventional example shown in Figure 17, no longer occurs, and the reticle and the sea urchin can be aligned with high precision. The purpose of the present invention is to realize alignment, but the present invention is a means for solving the problem.Two light beams for alignment, which have slightly different frequencies and can interfere with each other, are provided at predetermined positions on the wafer. Irradiation is performed on the alignment grating through a window provided in the reticle and a reduction projection lens so that interference fringes are formed on the alignment grating. an alignment optical system that emits light so as to form the interference fringes; a first photodetector that detects a hetero gain beat signal of the ±1st-order light diffracted and interfered by the alignment grating on the reticle; a second photodetector for detecting a heterodyne beat signal of the ±1st-order light diffracted and interfered by the alignment grating; A phase meter that detects the phase difference detects the phase difference between each hetero gain beat signal detected by this phase meter, measures the relative position of the wafer with respect to the reticle based on this phase difference, and then measures the position of the reticle or wafer. According to the present invention, there is provided an alignment optical device characterized in that alignment is performed by displacement. The overlay accuracy and throughput are higher u% compared to the conventional example in which the grids of the reticle and the wafer are aligned separately.Also, according to the present invention, unlike the conventional example, the reticle and the wafer can be aligned separately. There is no need for a positioning optical system between the reduction projection lenses.Therefore, the overlapping position does not change due to changes in the positioning optical system over time and is stable.Also, it is possible to avoid the need for a complicated positioning optical system like the conventional example. The advantages of no adjustment and the high productivity of the reduction projection exposure apparatus are also described below.Examples of the Alignment Apparatus of the Invention As shown in FIG.
is used to align the 4 semiconductor wafers with 10
(t) Even though the reticle 20 is placed on the wafer stage 30 and the surface is coated with photoresist, the center of the reticle 20 has a rectangular circuit pattern area 22 where a predetermined circuit pattern is formed, as shown in FIG. However, a reduction projection lens 40 is provided between the reticle 20 and the wafer 10 as shown in FIG. Even if it is reduced and projected, in this example, the force between the reticle 20 and the wafer 10 is
, TTR(Throughthe 1ens)-o
Alignment is performed using the n axis method.

The wafer stage 30ζ, lt is capable of moving in the X-Y-Z directions, tilt correction, and rotation in the θ direction. Each window group 25 (Y) is located outside the outer periphery of the three straight lines of the rectangular circuit pattern area 22. has,
The windows 25a in each window group 25 are arranged along each outer periphery of the circuit pattern area 22. When the reticle 20 is exposed to the LSI chip area of the wafer 10, each window group 25 is formed within the LSI chip area.
Even if the reticle 20 is provided so that a projected image of 21a, each alignment grating group 21, reticle 20, and wafer 10.
It is used to detect positional deviations in the X, MY, and θ directions. Even if the lower surface of the peripheral edge of the reticle 20 is covered by a light-shielding band 23, this light-shielding band 23 may cover only the alignment gratings 21a of each alignment grating group 21 without covering the windows 25a of each window group 25. Even if three alignment grating groups 11 (see Fig. 1) are provided above, each alignment grating group 11ζ 3
Even if each of the alignment gratings 11a in each alignment grating group 11 has two alignment gratings 11a, a predetermined window 25a of each window group 25 provided on the reticle 20 passes through the reduction projection lens 40 to the wafer 1.
When projected onto the window 25a, the window 25a is located within the projected portion of the window 25a. As shown in FIG. 1, each alignment grating group 21 and each window group 25 forming a pair of reticle 20 are irradiated with two light beams 51 and 52 from an alignment optical system (not shown).
and 52 are slightly different, and both light beams 51 and 52 may interfere with each other. Each of the light beams 51 and 52 intersects on the alignment grating 11a to form a thousand madder stripes 53.
The ±1st-order light diffracted by the alignment grating 21a of the reticle 20 and interfering with each other is irradiated onto the photodetector 61 provided above the reticle 20. The optical heterodyne beat signal intensity of the diffracted and interfered ±1st-order light is emitted, and each light beam 51 irradiated to the 8 windows 25a of the window group 25
and 52 (y) After passing through the 8 window 25a, the light beams 51 and 5 pass through the reduction projection lens 40.
2 is reduced and projected onto the wafer 10 by the reduction projection lens 40 to form interference fringes 54 on the alignment grating 11a on the wafer 10. Each light beam is diffracted by the alignment grating 11a of the wafer 10 and interferes with the ±1st order. When light passes through the reduction projection lens 40 and the window 25a of the reticle 20 and is irradiated onto the photodetector 62 provided above the reticle 20, the light is diffracted by the alignment grating 11a on the wafer 10. The output of each photodetector 61 and 62 is fed to the phase meter 63.
3c. Control the drive circuit 64 of the wafer stage 30 so that the phase difference between the optical heterodyne beat signals detected by each photodetector 61 and 62 becomes 0.
The principle of the alignment device of the present invention is shown in FIG. 3(a).
As explained based on and (b), two luminous fluxes 51 and 52
are respectively Ll+ and 02, and the respective frequencies are fl and f
Let it be a. If the incident angles of the two light beams 51 and 52 onto the wafer IO are respectively θ, then the two light beams 51 and 52
The pitch of the interference fringes formed by P is given by the following equation.

λ (where λ is the wavelength of the interference light) is placed on the semiconductor wafer 10 (mu) with a pitch force equal to the pitch of this interference fringe ＼ Alignment grating 11a with a pitch that is an integral multiple
is formed, the ±1st-order light U+(-1) and U2(1) diffracted by the alignment grating of the wafer 10
C They are respectively expressed by the following formulas: lJ+ (-1)-A(f+)exp[1(2yr
f+ t-δ)) IJ2(+1)-A(fa)exp(
i(2yr fat+δ)) where δ is the phase difference of the diffracted light when the alignment grating on the wafer moves by ΔXll, and is expressed by the following equation.

△
+ (-1>+U2(1) Q-A-2(f+)+
13t+” (:fa)+2Aw (fa)Bm (
fa) cos(2rr (f+-fa)t-2δ)
Jm'(f+)+B-"(fa)+2Aw(b)By(fa)cos2yr
((f+ -b)t-2(△x 1./P)] From this equation, it is clear that the intensity of the heterodyne beat signal detected by the photodetectors 61 and 62 is Similarly, the intensity IRζ of the interference light incident on the alignment grating 21a on the reticle 20 and diffracted is expressed by the following equation.

1++=A*'(f+)+BR2(fa)+2Aa(
fa)B++(b)cos27r((f+-fa)
t-2(△x */P)) Therefore, the phase difference between the interference light from the wafer 10 and the interference light from the reticle is measured using a phase meter, and alignment is performed so that the phase difference becomes O. Although it is possible to align the reticle 20 and the wafer 10 using the fringes as a medium, in the present invention, even if two-beam interference fringes are generated in the area where the two light beams intersect, they are defocused and used as alignment light because the depth of focus is deep. It is easy to use a positioning wavelength different from the exposure wavelength of the reduction projection lens. FIG. 4 shows a phase change measured by the phase meter 63 when the wafer 10 is moved relative to the reticle 20. The phase change CL measured by the phase meter 63 The phase change CL that changes periodically with the pitch of the interference fringes The wafer stage 304 on which the wafer 10 is placed X As shown in FIG.
A module having a coarse movement stage 31 in the direction and a coarse movement stage 32 in the Y direction placed on the coarse movement stage 31 in the X direction.
Even if an X-Y-θ direction stage 33.2, an α-β direction drive stage 3, and a θ-direction coarse movement stage 35 are placed in sequence on the direction coarse movement stage 32, and The wafer 10 is placed on the θ-direction coarse movement stage 35, and a laser interferometer 36 is provided on the 4Y-direction coarse movement stage 32 to detect the position of the Y-direction coarse movement stage 32. As described above, the wafer stage 30 is driven by the output of the drive circuit 64, and the drive circuit 64 is controlled based on the detection result of the phase meter 63. The operation of the wafer stage 30 by the drive circuit 63 is shown in FIG. The explanation will be based on a flowchart.The wafer 10 is placed on the coarse movement stage 35 in the θ direction (wafer load), and the entire wafer stage 30 is moved to the position where the circuit pattern of the reticle 2 is exposed by the exposure light source. (coarse alignment).This coarse alignment ζdai For example, the wafer l
This is carried out by moving the wafer stage 304 to a position where the rough alignment mark provided at the wafer 10 is observed by a global microscope.
The X-direction coarse movement stage 31, the Y-direction coarse movement stage 32. and the θ-direction coarse movement stage 35 are driven (X, Y, and θ-direction global alignment). This global alignment is performed with an accuracy of about 0.5 μm, similar to global alignment performed in normal wafer processes. At this time, Y
The Y-direction coarse movement stage 32 is moved based on the detection result of the laser interferometric length measuring device 36 provided on the direction coarse movement stage 32. The X-direction coarse movement stage L is similarly moved based on a laser interferometer (not shown).
The alignment gratings 21 of each alignment grating group 21 on the reticle 20 and the alignment gratings 11 in each grating group 11 of the wafer 10 are irradiated with two light beams 51 and 52 to form the alignment gratings of the reticle 20 as described above. 4. The phase difference between the interference light produced by the interference light beam 21a and the interference light produced by the alignment grating 11a of the wafer 10 is detected by the phase meter 63. ,
Although the fine movement stage in the X-Y-θ direction is driven,
When the reticle 20 and the wafer 10 are aligned in this way, the exposure light is directly irradiated onto the reticle 20 and the circuit pattern on the reticle 20 is transferred onto the wafer 10. 4 The reticle 20 and wafer 10 in this example
We analyzed the error factors in the alignment between the reticle 20 and the wafer 10 and investigated the alignment accuracy (error) between the reticle 20 and the wafer 10 due to each error factor. , by the two-beam interference sensor Measurement reproduction error of the alignment grating Manufacturing error of the alignment grating Position control error of the X-direction fine movement stage and Y-direction fine movement stage 33 Position control error of the θ-direction fine movement stage 35 Manufacturing error of the reticle itself I can think of it. In this embodiment, the circuit pattern of the reticle 20 is exposed on the wafer lO by the TTR-・n・・I・ method (mis). No, ma?Q
X-direction 11) Positional deviation error caused by movement error of the θ-direction coarse movement stage and positional deviation when fixing the reticle (reticle rotation). (-The overall total error was 50 ntn/3σ. Figures 8(a) and 8(b) show histograms of overlay accuracy.) Conventional method for alignment grid with 5iCh steps The overlay accuracy is 0.15μm
However, with the alignment method of the present invention, an overlay accuracy of 55 nm/3σ is achieved even for A1 alignment gratings with surface roughness. In the method of the present invention, the variation in the interference fringe generation position due to environmental changes is within 10 nm after 18 hours. Even though alignment is performed with high accuracy, the method of the present invention can achieve alignment with a high accuracy of 55 nm/3σ, and can achieve stable alignment even in the face of environmental changes, resulting in a high throughput. This makes TTR-0゜a possible, making it extremely effective for integrating highly integrated semiconductor devices such as 64-megabit DRAM.
xis exposure method, but TTL (Through t
When exposing a circuit pattern on a reticle to a wafer using the exposure method, a reticle 20' shown in FIG. 10 is used.A circuit pattern is formed in the center of this reticle 20'. Even if a rectangular circuit pattern area 26 is formed, one window 27 is provided at a position outside the outer periphery of this circuit pattern area 26, and eight windows 27 are provided.
Even if one alignment grid 28 is provided to the outside of the ware 110, as described above, each of the ware 110 is 3 to the outside of the LSI chip area where the circuit pattern is projected.
Alignment grating group 1 having two alignment gratings 11a
1 is provided, the lower surface of the outer peripheral edge of the reticle 20' is covered by the light-shielding band 29. In the TTL-0ffaxl exposure method in which a reticle 20' is used, a wafer stage 90 shown in FIG. 11 is used. X-Y composed of piezo elements
Even if the -θ direction fine movement stages 93 are mounted in order, the X-Y-θ direction fine movement stages 93 and
α-β direction fine movement stage 94. Even if the θ-direction coarse movement stage 95 and the θ-direction coarse movement stage 95 are placed in order, the wafer 1o is placed on the θ-direction coarse movement stage 95. On the X-Y-θ direction fine movement stage 93, there is a laser interferometer 9 for detecting the position of the X-Y-θ direction fine movement stage 93.
6 is provided. The alignment of the reticle 20° and the wafer 10 is performed as follows. First, the entire height of the wafer stage 90 is positioned approximately at a position where the circuit pattern of the reticle 20' can be projected onto the wafer 10.
The positioning grating 11a is moved by a predetermined distance based on the laser interferometer 46, and the positioning grating 11a is used for detecting a deviation in the X direction on the wafer 1o (the window 26 of the reticle 20'' for detecting the X direction is reduced on the wafer 10). It is aligned with the position projected by the projection lens. What is the positional deviation error in this case?
! , even if it is about 0.5 μm. In this state, the alignment light is projected, and as mentioned above, the phase difference of the heterodyne beat signal is measured using the two-beam interference fringes as a medium.
The amount of positional deviation between the reticle 20'' and the wafer 10 in the
Each positioning grating 11a force (reticle 20
' is moved to the position corresponding to the window 26 for detecting a deviation in the Y direction and the window 26 for detecting a deviation in the θ direction. In this way, when the positional deviation amounts of the reticle 20" and the wafer 10 are detected in the X direction, Y direction, and θ direction, the entire force of the wafer stage 90 is determined. The wafer stage 90 is moved based on 96 and placed at the initial exposure position. At this position, the position of the wafer stage 90 relative to the reticle 20' is roughly adjusted. Fine movement stage 93 in X-Y-θ direction to correct positional deviation amount
is driven and the wafer 10 is aligned with the reticle 20 degrees. When the wafer 10 is aligned with the reticle 20 degrees, the circuit pattern of the reticle 20'' is exposed onto the wafer 10 and the holes are opened. After alignment, the stage moves to the exposure position.The same thing can be done even if the reticle stage moves.Here are the results of analyzing the error factors in alignment between reticle 20' and wafer 101 in this case. It is shown in Figure 12. In this case ζ
Friend: Unlike the TTR-0, 1-1 exposure method mentioned above,
The overall force length of the wafer stage 90 is to detect the amount of positional deviation between the reticle 20'' and the wafer 10 in the X direction, rKY direction, and θ direction.
0 movement causes an error (alignment sequence error) force <, 30nm/3σ, but other error factors Get,
TTR-, +1 auto x 1 ・Same force as in the case of exposure method (the amount of movement error of the coarse movement table in the X direction and the coarse movement table in the Y direction increases slightly).As a result, the total error as a whole is 62 nm/ 3σ, and the TTR-
11 The alignment accuracy is slightly lower than in the case of the @-+- exposure method.The formation position of the alignment grid on the wafer is not limited and can be formed at any position. Another embodiment of the present invention is shown in which the exposure area can be prevented from being narrowed by the light for positioning by an exposure light source for exposing a wafer. In this embodiment, as shown in FIG. The laser beams are divided into two beams, and each laser beam is given to the first acousto-optic element 83a and the second acousto-optic element 83b. After successively passing through the half mirror 84 and lens group 85, the interference light is irradiated onto a predetermined alignment grating and window of the reticle 20 by the mirror 86, and is diffracted by the alignment grating on the reticle 20, and alignment of the wafer lO The interference light ζ diffracted by the grating is reflected by the mirror 86,
After passing through the lens group 85, the output of each photodetector 61 and 62 is applied to each photodetector 61 and 62 at a half mirror 84, and the output of each photodetector 61 and 62 is provided to a phase meter 63. In this embodiment, it is assumed that the polarization direction of the laser light emitted from the laser light source and split by the beam splitter 82 is constant. There is no surface contamination and there is no risk of adversely affecting the phase characteristics of the heterodyne beat signal. Therefore, the reticle 20
Even if the alignment optical system 80EL of this embodiment and the wafer 10 are aligned with high precision, the alignment optical system 80EL of this embodiment may be adjusted as shown in Fig. M14. Aligning light or a pair of parallel beams that are irradiated onto the reticle 20 through the same path as these alignment lights are transmitted to the reticle 20 by the half mirror 81K and lens group 8 mirror 86.
The light reflected by the reticle 20 is irradiated onto the half mirror 84 by the mirror 8 and the lens group 85.Then, the half mirror 84 irradiates the light onto the mirror 87a, The light is reflected by the mirror 87a and then irradiated onto the reticle 20 by the half mirror 84, lens group 85 and mirror 86.The spot positional force of the light at this time (each acousto-optic element 83a and 8
3b, half mirror 84, lens group 85
．． and the lens group 8 so as to match the spot position of the light directly irradiated onto the reticle 20 by the mirror 86.
5 and mirror 86 are adjusted in this way.
When the mirror 86 and lens group 85 are adjusted, a pair of parallel lights are similarly irradiated onto the reticle 20 through the half mirror 84, the lens group 85, and the mirror 86.
The reflected light reflected by the reticle 20 is guided to the half mirror 84 through the same path, and interference fringes caused by interference between the light split by the half mirror 84 and the light incident on the half mirror 84 are detected by the CCD. The image sensor 87b captures the interference fringes on the monitor TV 87c.
lens group 8 to make the interference fringes clear.
5 and mirror 86 are adjusted. In this adjustment method, the optical power reflected by the reticle 20 (divided by the half mirror 84 and then image-formed) is explained. Even if it is made to pass through the lens 87e, the light that has passed through the imaging lens 87e (after being diffracted by the diffraction grating 87f, C
CD image sensor 87g is irradiated Imaging lens 8
A lens having the same optical ability as the imaging lens 85 used in the alignment optical system 80 is used, and a diffraction grating 87 f i! .

A reference grating 76 similar to that used in the alignment optical system 70 shown in FIG. 14 may be used. Also, the respective positions of the imaging lens 87a, interference fringe grating 87f, and CCD image sensor 87g are determined by three-dimensional length measurement. It is reflected from the reticle 20 and determined by the half mirror 8.
The two beams J divided at 4 are irradiated onto the diffraction grating 87f through the imaging lens 87e, and the two beams interference fringes are formed by the diffraction grating 87f.Moiré fringes are generated between these interference fringes and the diffraction grating 87f. This moiré fringe is also captured by the CCD image sensor 87g,
Even though the mirror 86 and the imaging lens 85 are adjusted so that the moiré fringes displayed on the monitor TV 87h become one, this is not the case.In this embodiment, the alignment optical system 80 can be easily adjusted. and wafer 10 are aligned with extremely high precision.

Effects of the Invention The alignment optical device of the present invention can align a reticle and a wafer with high precision in a reduction projection exposure apparatus.The alignment grating provided on the reticle is reduced and projected onto the wafer for close-up exposure. Usually 5 for the pitch of the alignment grid provided on the mask in the device.
The pitch of the alignment grid on the reticle can be increased by increasing the pitch of the alignment grid. The alignment optical system that irradiates light does not need to be of high performance, and an inexpensive alignment optical system can be used.
Because the luminous flux is irradiated! Even if alignment is performed with high throughput,

[BRIEF DESCRIPTION OF THE DRAWINGS] FIG. 1 is a block diagram showing an example of an optical device for positioning a reticle and a wafer in an exposure apparatus of the present invention. FIG. 2 is a planar view showing an example of a reticle. Schematic diagram for explaining the principle of the alignment optical device of the present invention Part 3
Figure (b) is a graph showing the heterodyne beat signal of the interference light due to the alignment grating of the wafer and the interference light due to the alignment grating of the reticle, Figure 4 is a graph showing the phase change with respect to relative fluctuation between the reticle and the wafer, and Figure 5 The figure shows the configuration of the wafer stage. Figure 6 is a flowchart explaining the operation of the alignment optical device. Figure 7 is the TTR-07.
゜゜I@Figure 8 (a) and (b) show the relationship between error factors and errors of the alignment optical device in the exposure device.
Figure 9 is a graph showing the overlay accuracy with respect to the Oa step and A1 surface roughness grating. Figure 10 is a graph showing the amount of variation in interference fringes over time.
Figure 12 is a configuration diagram showing another example of a wafer stage. Figure 12 is a diagram showing error factors and errors of the alignment optical system in the TTR-on-owls exposure method. Figure 13 is a configuration diagram showing another example of the alignment optical system. Fig. 14 is an explanatory diagram of the adjustment method of the positioning optical system. Fig. 15 (a), (b)
) is an explanatory diagram of another adjustment method for the positioning optical system. Figure 16 is a configuration diagram showing an example of a conventional positioning optical system. Figure 17(a) is an enlarged view of the main part. Figure 17(b)
18 is an explanatory diagram of its principle. FIG. 18 is a flat plate of a reticle used in the alignment optical device. FIG. 19 is a configuration diagram showing still another example of the conventional alignment optical device. Semiconductor ueno ＼ 20...Rechikuno L
<25a...S30...Wafer stage,
11... Grating W 40... Reduction projection lens, 7
51.52... light! 61.62...Photodetector 63...Phase ratio 64...Name of driving agent Shigetaka Awano Haka1 Phase (degree) ) ■ Doufu Kagure Ward Figure 9 Figure 12 Figure 1 Figure entrance, -', -62 Figure 16 Figure 7 (a) \, -2/ 187 \ 12'lα Figure 19 27/

Claims (7)

[Claims] (1) An optical device for aligning the reticle and wafer when reducing and projecting the circuit pattern provided on the reticle onto the wafer, and is used to position the reticle and the wafer at positions where the frequencies are slightly different and may interfere with each other. Two beams of light for alignment are applied to an alignment grating provided at a predetermined position on the wafer,
Irradiation is performed through a window provided on the reticle and a reduction projection lens so that interference fringes are formed on the alignment grating, and at the same time, an alignment grating provided at a predetermined position on the reticle is irradiated onto the alignment grating. a positioning optical system that emits light so that the interference fringes are formed;
A first detecting heterodyne beat signal of ±1st-order light diffracted and interfered by the alignment grating on the reticle;
a photodetector, a second photodetector that detects a heterodyne beat signal of the ±1st order light diffracted and interfered by the alignment grating on the wafer, and detected by the first photodetector and the second photodetector. a phase meter that detects a phase difference between each of the heterodyne beat signals detected by the phase meter, and a phase meter that detects a phase difference between each of the heterodyne beat signals detected by the phase meter, and measures the relative position of the wafer with respect to the reticle based on this phase difference; An exposure apparatus that performs alignment by displacing a reticle or a wafer. (2) The reticle and the wafer each have three alignment gratings for detecting the amount of positional deviation between the reticle and the wafer in the X direction, Y direction, and θ direction.
The exposure apparatus described. (3) The exposure apparatus according to claim 1, wherein the wafer is aligned by detecting the amount of misalignment with the reticle at a position on the wafer where a circuit pattern on the reticle is exposed. (4) The wafer is moved by a predetermined amount in each of the X direction, Y direction, and θ direction with respect to the reticle from the position where the circuit pattern on the reticle is exposed on the wafer, and the amount of deviation in each direction is 3. The exposure apparatus according to claim 2, wherein the wafer is aligned by being detected and then being moved to the exposure position and then moved by the amount of deviation in each direction. (5) The alignment optical system consists of a light source that emits light of a predetermined wavelength, a beam splitter that splits the light of the predetermined wavelength emitted from this light source into two, and a frequency of each of the lights split by the beam splitter. A pair of acousto-optic elements that modulate the frequency modulated light by each acousto-optic element.
2. The exposure apparatus according to claim 1, further comprising means for making the light beam incident on a set of lens systems and irradiating the diffraction grating and window of the reticle with the two-split light beam to generate two-beam interference fringes. (6) The alignment optical system has a heterodyne holography alignment optical system composed of an alignment light source, a beam splitter, an acousto-optic element, a condenser lens, a mirror alignment grating, a photodetector, and a phase meter, and A half mirror, total reflection mirror,
An imaging device is provided, parallel light having the same wavelength as the alignment light source is incident on the reticle surface through the condensing lens, and reflected light from the reticle surface is reflected by the total reflection mirror via the half mirror. A patent characterized in that the positions of the mirror, the condensing lens, and the two light beams for positioning are adjusted while making the incident light divided by the mirror interfere with the reflected light and observing interference fringes on the image pickup device. An exposure apparatus according to claim 1. (7) The alignment optical system has a heterodyne holography alignment optical system composed of an alignment light source, a beam splitter, an acousto-optic element, a condenser lens, a mirror, an alignment grating, a photodetector, and a phase meter, a half mirror between the condenser lenses and the acousto-optic element, a condenser lens,
A diffraction grating is provided, two light beams emitted from the alignment light source and split by the beam splitter are made incident on the reticle surface, and the light reflected from the reticle is transmitted to the imaging lens, the half mirror, and the imaging lens. irradiate onto the reference grating through the reference grating to generate two-beam interference fringes, and while observing moiré fringes generated between the two-beam interference fringes and the reference grating,
Claim 1 characterized by adjusting the light beam position
Exposure apparatus described in Section 2.