JP2773779B2 - Displacement measuring method and displacement measuring device using diffraction grating - Google Patents

Description

Description: TECHNICAL FIELD The present invention relates to a displacement measurement technique, and more particularly to a displacement measurement technique for measuring displacement with high accuracy.

2. Description of the Related Art In recent years, displacement measurement with a high accuracy of the order of several nm has been required in semiconductor device manufacturing technology and the like. The present invention enables such highly accurate displacement measurement.

[Prior Art] Conventionally, a technique based on laser interference is known as a displacement measuring technique for an XY stage or the like. One laser beam is split into two beams, one of which is reflected by a mirror provided on the object to be measured, and is combined with the other beam that has traveled a constant optical path length.
Cause interference. When the two laser beams are expressed as ASin (wt) and ASin (wt + 2δ), the composite beam is ASin (wt) + ASin (wt + 2δ) = 2ASin ｛(2wt + 2δ) / 2｝ · Cos (2δ / 2) = 2ASin (wt + δ) Cos (δ). That is, when the reflecting mirror is displaced by △ x, the optical path length reciprocating there changes by 2 △ x, and the light of wavelength λ is 2δ = 2π
(2 △ x) / λ phase is changed. The term Cosδ oscillates with this phase change, and oscillates the amplitude of the electric field intensity of the light, and thus the energy of the light, to draw a so-called interference fringe.
By detecting this interference fringe, the displacement of the reflecting mirror can be known. If the amplitudes of the two light beams are equal, the light intensity of the interfering composite wave oscillates between the maximum value and zero.
In this case, it is relatively easy to determine the phase.

However, if the amplitudes of the two light beams are different, the light intensity of the composite wave does not become zero at least. Further, if the amplitude changes due to disturbance or the like, it becomes difficult to determine the maximum and minimum positions of the light intensity of the interference fringes, and it is difficult to determine the phase.

In a semiconductor device manufacturing process, it is necessary to measure a relative position between a mask and a wafer with high accuracy. Here, an oxide film, a wiring layer, a resist layer, and the like are stacked on the surface of the wafer, and the reflectance thereof greatly changes. Therefore, there is a limit in using the laser light interference method for wafer alignment.

Optical heterodyne interference has been proposed as a method for solving this problem. Optical heterodyne interference has wavelengths (λ ₁ , λ
₂ ) or slightly different frequencies (f ₁ , f ₂ ) (λ ₁ ≒
λ ₂ , f ₁ ≒ f ₂ ) Two laser beams are made incident on a diffraction grating on the object to be measured, and a beat-difference component is detected similarly to heterodyne detection. A typical example of the conventional optical heterodyne interference technique will be described below. These techniques are disclosed in the following documents, for example.

1) Autumn Meeting of the Japan Society of Precision Optics 1989 B66 "Optical Heterodyne Interference High-Precision Registration Method Une et al. 2) 1989 Applied Physics Vol. 58, No. 10, P1511 (P85)-P1512 (P86)" Position in X-ray Sylography 3) Suzuki et al. 3) 1986 Autumn Meeting of the Japan Society of Applied Physics 28a-ZF-10 (P317) "Study of optical heterodyne interference type position detection method using diffraction grating" Suzuki et al. 4) JP-A-63-172904 "Diffraction "Position detection method and position detection device using lattice" Suzuki et al. 5) JP-A-62-274216 "Micro-displacement measurement method and device" Suzuki et al. 6) JP-A-62-261003 "Position alignment method and alignment" Suzuki et al. 7) JP-A-62-58628 "Positioning method and positioning apparatus" Suzuki et al. 8) J. Vac. Sci. Technol. B6 (1), Jan / Feb 1988 "An alignment system for synchrotron radiation"
x-ray Lithography ”Ito et al. 9) Applied Physics Vol. 56, No. 11, 1987, P1490 (P80)-P1494 (P84)“ High-precision alignment technology by optical heterodyne measurement using three diffraction gratings ”Ito et al. 10) SPIE Vol.773 Electron-Beam, X-ray and Ion-Be
am Lithographies V1 (1987) "A new mask-to-wafer alignment technique for
synchrotron radiation x-ray lithography ”Ito et al. 11) 1986 SSD86-113“ High-precision alignment method for X-ray lithography ”Ito et al. 12) Japanese Journal of Applied physics Vol. 25, No.
8, August, 1986, pp. L684-L686 `` Optical-Heterodyne Detection of Mask-to Wafe
r Displacement of Fine Alignment ”Ito et al. 13) Japanese Journal of Applied physics Vol. 25, No.
6, June, 1986, pp. L487-L489 `` A New Interferometric Displacement-Detection
−Method for Mask−to−Wafer Alignment Using Symme
Ito et al. 14) Japanese Patent Application Laid-Open No. 62-172203, "Relative Displacement Measurement Method" Kanayama et al. 15) 1988 Autumn Meeting of Precision Optics E26 "Application of Heterodyne Interference to Holographic Alignment" Yamamoto et al. 16) No. 64-89323, "Positioning device" Yamashita et al. 17) Japanese Patent Application Laid-Open No. 64-82625, "Positioning device" Sato et al. 18) Japanese Patent Application Laid-Open No. 64-82624, "Positioning device" Yamaguchi et al. No. 64-82623 "Positioning device" Aoki et al. 20) JP-A-64-82622 "Positioning device" Sato et al. 21) Spring Meeting of the Japan Society of Applied Physics 28a-H-5 "Diffraction light heterodyne in reduction projection exposure equipment" Interference Alignment Detection Method ”Kataoka et al. 22) Fall Meeting of the Japan Society of Applied Physics 1987, 18a-F-8 (P427)“ Nanometer position detection method for steppers by photography (IV) ”Nomura et al. Background First, Fig. 11 (A), Referring to (B), the base of the diffraction grating The essential properties will be described. The diffraction grating is formed by, for example, arranging grooves parallel to the Y direction in the X direction at regular intervals (pitch P), and has a function of dispersing incident light at an angle corresponding to the wavelength.

Given the case of normal incidence, when the incident light rays l ₁ , l ₂ , and l ₃ are reflected at the reflection angle θ in the adjacent grooves in FIG. 11 (A), the optical paths of the reflected lights l ₁₁ , l ₁₂ , and l ₁₃ The length differs by PSinθ. When the optical path difference PSinθ is an integral multiple of the wavelength, the reflected light reinforces each other, and when the optical path difference PSinθ is (integer + 1/2) times the wavelength, the reflected light decomposes. In this way, the diffraction grating performs wavelength dispersion of light. The number of times the optical path difference between light beams reflected by adjacent grooves is equal to the wavelength is called the order of diffraction.

As shown in FIG. 11 (B), when light of a predetermined wavelength (frequency) is vertically incident on the diffraction grating, diffracted light of each order is generated symmetrically. These are referred to as ± 1st order, ± 2nd order, ± 3rd order diffracted light. When two coherent lights having slightly different frequencies f ₁ and f ₂ are incident, each diffracted light is generated in substantially the same direction.

Incidentally, the diffraction grating has a periodic structure in the X direction, and its period is P. When the diffraction grating itself is displaced by Δx in the X direction, the diffracted light also changes according to the displacement. Since the period of the diffraction grating is P, an equivalent state occurs for all the diffracted lights due to the displacement of P. However, in the case of the m-th order diffracted light, a periodic displacement having a displacement of P / m as a unit (2π) is shown. That is, when the diffraction grating performs a displacement of △ x, the m-th order diffracted light has 2π · △
A phase change of x / (P / m) occurs. If this phase change can be detected, the displacement can be measured. However, the frequency f of the light is very high, and it is not easy to detect the phase as it is.

By the way, when two electromagnetic waves having slightly different frequencies are superimposed, the difference component can be extracted as a beat wave. That is, the frequencies f ₁ and f ₂ = f ₁ + {f (f ₁ >>>>
A signal having a frequency Δf can be obtained from the two electromagnetic waves f). For example, if two sine wave having the same _{amplitude, ASin (w 1 t) +} ASin (w 2 t + 2δ) = 2ASin {(w 1 t + w 2 t + 2δ) / 2} -Cos {(w 1 t-w 2 t- 2δ) / 2｝, and a signal having a low angular frequency of (w ₁ −w ₂ ) can be obtained. In the case of light, a signal having such a low frequency component can be created by using two coherent lights having very similar wavelengths (frequency). As a light source capable of generating coherent light having such slightly different wavelengths (frequency), for example, a transverse Zeeman laser is known. The transverse Zeeman laser generates two coherent lights having frequencies f ₁ and f ₂ (f ₁ ≒ f ₂ ) with their polarization planes orthogonal to each other. If optical heterodyne detection is performed using such a light source, a low-frequency signal can be obtained.

If this optical heterodyne detection can be combined with the displacement of the diffracted light due to the displacement of the diffraction grating to include a phase change representing the displacement of the diffraction grating in the low-frequency signal, the displacement of the diffraction grating can be easily known from the detection of the diffracted light. be able to.

Provide a diffraction grating (grating pitch P) on the object to be measured,
When the frequency f _1, f ₂ enters the two coherent beams of different diffraction grating, the two lights f _1, f ₂ respectively ± n order (n = 1,
2, 3,... N) diffracted light is obtained.

Here the next m for light of a frequency f ₁ diffracted light and the frequency f ₂
The n-th order diffracted light with respect to the light is caused to interfere. The following relationship is established between the obtained phase φ of the beat signal and the displacement Δx of the object.

φ = 2π △ x / {P / (m-n)} ... (1) The frequency f ₁ of the light m n order diffracted light and the frequency f ₂ light
Having described the case of interference-order diffracted light, obtained similar phase term also causes interference m order diffracted light of the light of light n-order diffracted light and the frequency f ₂ of the frequency f _1. For example, first-order diffracted light
When it interferes with the first-order diffracted light, mn = 1 + 1 = 2, and φ = 2π · △ x / (P / 2) = 4π · △ x / P (2)

In the equation (2), if the phase difference φ is known by the optical heterodyne measurement, the displacement amount Δx of the object is obtained as follows.

Δx = P · φ / 4π (3) When measuring the relative position of two objects, a diffraction grating is provided on each object, and the expression (1) or (2) is applied to each diffraction grating (that is, each object). Is obtained, and the relative position is obtained by comparing the phase φ or the displacement Δx.

The measuring methods applying the above principle are classified into the following four types.

Conventional Example 1.1 Principle FIG. 12 shows a detection optical system. Two light beams having different Zeeman laser frequencies (f ₁ , f ₂ ) are transmitted from the Zeeman laser 51 through the half mirror 59 and the lens 52 to the diffraction grating 53 on the wafer.
Enters the (pitch P), the -1st-order diffracted light of the + 1st order diffracted light and the frequency of the frequency has been reflected diffracted f ₁ is f ₂ lens 52,
The light is caused to interfere by a mirror 54, a spatial filter 55, a polarizer 56 and the like, and is received by a photomultiplier (photomultiplier tube) 58 through a slit 57.

In the figures, reference numerals 60 and 61 indicate lenses. Between the Zeeman laser 51 is a light source to the polarizer 56 is in the independent relationship mutually orthogonal to the light of the light and the frequency f ₂ of the frequency f _1, polarizer
56 is 45 for the electric vector of these two light beams.
Is inserted at an angle of ゜ and the two light beams overlap and interfere.

Phase difference φw of generated beat signal (wafer beat signal)
Separately, the phase difference φo of the reference beat signal obtained by directly interfering the interference light having the frequency f ₁ and the frequency f ₂ is optically heterodyne-measured, and the difference between the two phases is compared to obtain the displacement Δx of the wafer diffraction photon. Is measured.

Here, the beat intensity Iw from the wafer caused by the interference of the ± 1st-order diffracted light can be expressed by equation (4), and the intensity change Io of the reference beat signal can be expressed by equation (5).

Iw = Aexp [i {2π ( f 1 -f 2) t + 4π · x / P}] ... (4) Io = Bexpi [{2π (f 1 -f 2) t}] ... (5) However, A, B Is a constant representing the amplitude, and x is the position of the diffraction grating. In equations (4) and (5), the phases φw, φ
o is represented by the following equations (6) and (7).

φw = 4π · x / P (6) φo = 0 (7) In equations (6) and (7), the phase φw of the wafer beat signal changes by 4π · x / P with respect to the reference beat signal. It indicates that From the equations (6) and (7), the position x of the wafer, that is, the displacement △ x, is obtained by the equation (8).

φw−φo = 4π · x / P ∴x = φw · P / 4π (8) A diffraction grating is provided on a mask or a reticle, and coherent light having frequencies f ₁ and f ₂ is incident on the diffraction grating. to give the -1 order diffracted light 1 order diffracted light and the frequency f ₂ of the frequency f _1, may be used as the reference signal a beat signal obtained by the interference of these. In this case, assuming that the phase term of the reference signal is φm, φw−φm = 4π / x / P Δx = (φw−φm) · P4π Here, Δx indicates the relative position of the wafer.

1.2 Problems (1). In FIG. 12, when an optical path difference (△ 1) occurs in the ± first order light, the phase of the beat signal represented by the equation (6) becomes 4π.
・ Changes by Δl / p, resulting in a detection error.

That is, as shown in FIG. 13, to interfere in a state of -1 shifted the phase of the diffracted light of the frequency f ₂ with respect to the + 1st order diffracted light of the phase of the frequency f ₁ by the optical path difference △ l, the resulting beat light The phase is shifted by 4π △ l / p.

This can be expressed on equations (4) and (6) as (9)
Equation (10) is obtained.

I′w = Aexp [i ｛2π (f ₁ −f ₂ ) t−4π · x / P + 4π △ l / p｝] = Aexp [i ｛2π (f ₁ −f ₂ ) t + (4π / P) (x + Δl)｝] (9) φ′w = 4π / P (x + Δl) (10) (2). The optical path branch is affected by disturbance. Since the optical path of the ± 1st-order diffracted light is branched, it is affected by disturbance due to fluctuations in the refractive index in the air during that time. For details of this conventional example, see, for example, Reference 21).

Conventional Example 2.2.1 Principle Fig. 14 shows the configuration of the detection system. The diffraction gratings 68 and 69 having the same pitch P are provided on the mask 66 and the wafer 67, and the frequencies (f ₁ , f ₂ ) are different from the ± 1st-order diffraction directions of both diffraction gratings.
Light simultaneously incident on the same beam spot position 71,
A ± 1st-order diffracted light having different frequencies reflected and diffracted on the same path interferes to form a beat signal. A membrane window 72 is provided in the beat spot 71, and the wafer diffraction grating 69 thereunder is exposed. The ± 1st-order diffracted lights having different frequencies reflected and diffracted by the wafer diffraction grating 69 also interfere with each other,
Form another beat signal. The phase of the beat signal from each of the mask and the wafer is separately measured, and the phases are compared to determine the relative displacement between the mask and the wafer.

The conventional example 2 solves the problem shown in the conventional example 1. That is, there is no position error due to the optical path difference Δl. This will be described below. Put the φE a phase difference of light of the light and the frequency f ₂ of the frequency f ₁ when incident from ± N following symmetrical angles.

Beat light intensity IM synthesized by mask diffraction grating 68
Is represented by equation (12).

IM = Aexp [i {2π ( f 1 -f 2) + φM + φE}] ... (12) However, A is the amplitude, .phi.M represents the phase change due to displacement xM mask diffraction grating represented by equation (13).

φM = 4π · xm / P (13) Similarly, the intensity IW of the beat light synthesized by the wafer diffraction grating 69 is expressed by equation (14).

Iw = Aexp [i {2π ( f 1 -f 2) + φw + φE}] ... (14) φw = 4π · xw / P ... (15) However, xw represents a change in the wafer diffraction grating. (13)
From equation (14), the phase difference Δφ between both beat signals obtained from the mask diffraction grating and the wafer diffraction grating is expressed by equation (16).

Δφ = φM−φW = 4π · xM / P−4π · xW / P = (4π / P) · (xM−xW) (16) As is apparent from equation (16), the phase difference Δφ is measured. Doing this is equivalent to obtaining the relative change between the mask and the wafer. Further, in the phase difference △ phi does not include a phase difference φE generated between incident at f ₁ and f _2. This is because the light is incident from a direction symmetrical with respect to the diffraction grating,
Since the same phase difference φE occurs in the beat signal from the mask and the wafer expressed by the equation (14), the phase difference 両 者 E between the two is obtained.
This is because when φ is taken, it is erased as can be seen from equation (16). Also, since the beat signals of the mask and the wafer expressed by the equations (12) and (14) have no relation to the gap at all, the gap between the _two is free in the region where the lights of f ₁ and f ₂ overlap. Can be set.

2.2 Problems (1). The device configuration is complicated.

FIG. 15 shows a basic configuration for detecting a relative displacement in one direction between a mask and a wafer by this method.

A wafer 67 is placed below the SOR aligner mask 66, and diffraction gratings 68 and 69 are formed on each of them. The mask 66 is provided with a membrane window 72, and the wafer 67
The upper diffraction grating 69 is exposed. Displacement measuring apparatus includes a Zeeman laser 51, a polarization beam splitter 75 for separating the two light beams orthogonal frequency f ₁ and frequency f ₂ are emitted from the Zeeman laser 51 by the polarization.

By the polarization beam splitter 75, the light of the frequency f ₁ is straight, the light of the frequency f ₂ is reflected. Reflected frequency f ₂
Is reflected by mirrors 76 and 77 and is incident on mask diffraction grating 68 and wafer diffraction grating 69. Also, light straight frequencies f ₁ is reflected by the mirror 78, similarly incident on the mask gratings 68 and wafer grating 69. Each light light and the frequency f ₂ of these frequencies f ₁ ±
The light enters from the direction of the first-order diffracted light, and is reflected and diffracted in the zero-order diffraction direction. The reflected and diffracted light is reflected by the mirror 79, forms a beat signal by the polarizing plate 93, and is detected by the photodiodes 80 and 81. Photodiode
The signals from 80 and 81 are supplied to a phase shifter 82, and a control signal is sent from the phase shifter 82 to an alignment controller 94. The alignment controller 94 controls a relative position between the mask 66 and the wafer 67. Incidentally, the light of the light and the frequency f ₂ of the frequency f ₁ is incident symmetrically from ± 1-order direction, the mask
The detection signal does not change due to the gap between 66 and wafer 67.

Incidentally, a signal of a frequency f ₂ that is reflected by the mirror 76 is divided by the beam splitter 86, is reflected by the mirror 87 and 88, 2 is incident from the direction of the diffracted light, and + 1st-order diffracted light of the laser light of frequency f ₁ interference is caused by the reflection by the mirror 89, by detecting a photodiode 90 and 91, the light of the light and the frequency f ₂ of the frequency f ₁ is incident in asymmetrical form, the detection signal of the mask 66 and the wafer 67 It depends on the gap between them. Using this signal, the gap between the mask 66 and the wafer 67 can be detected.

If at least three optical systems are configured in the alignment system, six-axis alignment servo control of relative displacement and gap can be performed.

However, if three systems as shown in FIG. 15 are prepared, the structure of the apparatus becomes considerably complicated.

(2). A change in the angle of the object in a direction other than the measurement direction affects the detection accuracy in the measurement direction.

FIG. 16 shows a state in which the diffraction grating is inclined with respect to the incident system. In this method, light is incident from a direction symmetrical with respect to a normal line of a symmetrical surface. When the surface is tilted by an angle θ, the direction that was the direction of ± 1st order light with respect to the vertically incident light is now shifted, so that the light reflected and diffracted by the diffraction grating is in the same direction. Does not proceed and separates in space. For this reason, in some cases, a state where no interference beat light is generated occurs.

(3). The asymmetry of the diffraction grating causes a detection error.

FIG. 17 illustrates a cross-sectional shape of an asymmetrical diffraction grating. It has been clarified that if the diffraction grating has an asymmetric shape, a phase lag occurs between the + n-order diffracted light and the -n-order diffracted light (for example, see Reference 22). Therefore, if the diffraction grating has an asymmetrical shape, a similar phase delay is included in the above-described prior art 2 technique for heterodyne measurement of the phase difference between the interference beat light of the + 1st-order diffracted light and the −1st-order diffracted light. An error occurs.

Conventional example 3.3.1 Principle Fig. 18 shows the basic configuration diagram. Basically, as in Conventional Example 2, light is incident from the direction of ± 1st-order diffracted light with respect to the vertically incident light, and output light diffracted and interfered from the 0th-order light is obtained.

In the figure, orthogonal frequency f ₁ from Zeeman laser 51
And the light of frequency f ₂ are emitted, and _two lights of ± first-order light which enter the reference diffraction grating 101, diffract and emit the light are used. One of these two lights converts the phase through the λ / 2 plate 102, and the Fourier lens 103, the spatial filter 10
4. Through the Fourier lens 105, the light beam enters the diffraction grating 106 for detecting displacement as two light beams. The output lights reflected and diffracted from the misalignment detection diffraction grating 106 interfere with each other through a polarizing plate 107 arranged at an intermediate angle, and are detected by a photodetector 108. In the lower part of the figure, light components in two directions perpendicular to the traveling direction of light are displayed as vectors.

In this method, since the ± first-order light is simultaneously treated using the Fourier lens, optical members such as mirrors can be omitted, and the configuration is simplified.

In the second conventional example, the relative displacement was measured based on the phase difference between the two objects. In the third conventional example, apparently, the displacement was detected with respect to the position of the diffraction grating (reference diffraction grating 101) provided in the incident optical path. The relative displacement of the object provided with the diffraction grating is determined. It is also possible to arrange a plurality of objects and determine their relative displacement. In principle, this is the same as in Conventional Example 2, and no phase error occurs due to the optical path difference. Further, as compared with Conventional Example 2, the number of optical path branches in the incident optical system is small, and the stability of the device is excellent.

3.2 Problems (1). FIG. 19 shows a basic configuration for measuring a relative displacement in one axis direction. A beam expander 110 is arranged on the optical axis of the light emitted from the Zeeman laser 51. The light whose beam diameter has been expanded by the beam expander 110 is
The light enters the reference diffraction grating 101 via 112. One of two diffracted lights diffracted by the reference diffraction grating 101 passes through the λ / 2 plate 102 and the other enters the Fourier lens 103 as it is, and is focused through the spatial filter 104 and the other Fourier lens 105. The light enters the mask 114 and the wafer 115. Diffraction gratings are provided on the mask 114 and the wafer 115, respectively. The output light reflected and diffracted by these diffraction gratings is guided to the photodetection system via the imaging lens 116 and the polarizing plate 107. Knife edge 118 to separate the two lights
Are arranged on the optical path. One light reflected by the knife edge is incident on the photomultiplier 108b via the mirror 119, and the other light is directly incident on the photomultiplier 108a. The output signals of these photomultipliers 108a and 108b are supplied to a phase meter 122 via band-pass filters 121a and 121b, and the phase signals are processed by a control device 124 such as a personal computer. Drive the wafer
Control 115 positions.

Although the configuration is simpler than that of the conventional example 3, the entire optical system still includes a plurality of mirrors and lenses, and it is desired to measure the displacement with a simpler configuration.

(2). The inclination of the object in a direction other than the measurement direction affects the detection accuracy in the measurement direction. As in the conventional example 2, if the mask or the wafer is tilted, the measurement accuracy is affected.

(3). The asymmetry of the diffraction grating causes a detection error.

Since the incident light is supplied from the symmetrical direction of the diffraction grating and the output light is obtained from the center similarly to the conventional example 2, the symmetrical shape of the diffraction grating causes a difference between the + 1st-order diffraction light and the −1st-order diffraction light. A phase difference occurs. This causes a measurement error. The details of the conventional example 3 are described in, for example, the above-mentioned literature 15), 1
Please refer to 8).

Conventional example 4.4.1 Principle Fig. 20 shows the basic configuration. On the mask 131, two diffraction gratings 133a and 133b having the same pitch are arranged apart from each other. A diffraction grating 134 is provided on a wafer 132 disposed below the mask 131, and is disposed at a position intermediate between the two diffraction gratings 133a and 133b of the mask 131. Diffraction gratings 133a and 133b on mask 131
Is for diffracting the incident light and making it incident on the diffraction grating 134 on the wafer 132. In the figure, u represents incident light,
u (-1) indicates that the incident light is the -1st-order diffracted light diffracted by the diffraction grating 133. Similarly, u (1) indicates that the incident light is the incident light diffracted in the + 1st order by the diffraction grating 133. U (-1, 2) indicates that the incident light is
Indicates that the light is diffracted in the −1 order direction and then diffracted in the +2 order direction by the diffraction grating 134. Similarly, u (1,-
1) indicates that the incident light is light that has been diffracted in the primary direction by the diffraction grating 133 on the mask 131 and then diffracted in the −1 order by the diffraction grating 134 on the wafer 132. When the pitch of the diffraction grating 134 is set to 1.5 times the pitch of the diffraction grating 133, as shown in FIG.
1, 2) and u incident on the diffraction grating 134 from the left side and diffracted
(1, -1) proceed in the same direction. In contrast, u
(-1, 1) and u (1, -2) travel in the same direction. A beat signal can be obtained by causing these diffracted lights traveling in the same direction to interfere with each other via the polarizing plates 136 and 137 and detecting them with the photodetector 138.

Further, the polarizing plates 141 and 142 and the photodetectors 143 and 144 receive the light diffracted in the higher order by the diffraction gratings 133 and 134, and the distance between the mask 131 and the wafer 132, and therefore, the mask 131
And a system for detecting the inclination of the wafer 132.

By the light intensity Is ₁ and Is ₂ received by the photodetector 138 and 139, the principle of detecting the relative displacement between the mask 131 and wafer 132 will be described in more detail below.

The two beat signals Is ₁ and Is ₂ are generated by the following process. Is ₁ has two diffracted lights u (1, −1) and u (−1,
Caused by the interference of the composite wave of the two different frequencies f ₁ and f ₂ in the synthesis wave 2). Here, u (1, -1) indicates that + 1st-order diffracted light by the mask diffraction grating is-by the wafer diffraction grating.
1 shows light diffracted in the first order.

Frequency f _1, the composite wave V ₁ by f _2, V ₂ is (17), represented by equation (18).

V ₁ = [u (1, −1) + u (−1,2)] = C ₁ exp {i (w ₁ t− △ ₁ )} (17) Similarly, V ₂ = C ₂ exp ｛i (w ₂ t− △ ₂ )｝ (18) where w ₁ = 2πf ₁ , w ₂ = 2πf _{2 １1} = tan ^-1 ｛Sin2δ / (α + Cos2δ)｝ ₂ = tan- ¹ Sin2δ / (β + Cos2δ)｝ ^{_{c 1 = (1 + α 2}} + 2αCos2δ) 1/2 c 2 = (1 + β 2 + 2βCos2δ) 1/2 α, β represents the diffraction efficiency for f _1, f _2.

(17), the beat signal Is ₁ from (18) is expressed by the equation (19).

Is ₁ = (V ₁ + V ₂ ) ² = ACos {(w ₁ −w ₂ ) t− (△ ₁ − △ ₂ )} (19) Similarly, the beat signal Is ₂ can be expressed by equation (20).

Is ₂ = ACos ｛(w ₁ −w ₂ ) t + (△ ₁ − △ ₂ )｝ (20) From (19) and (20), the phase difference Δφ between Is ₁ and Is ₂ is expressed by equation (21). It is.

Δφ = − (△ ₁ − △ ₂ ) − (△ ₁ − △ ₂ ) = − 2 (△ ₁ − △ ₂ ) 2tan ^-1 [(α-β) Sin2δ / ｛(1 + αβ + (α + β) Cos2δ｝] (21) The phase difference Δφ expressed by the equation (21) becomes zero when the relative displacement δ = 0.

4.2 Problems (1). First to generate diffracted light with incident optical axis and phase information
The change in the orthogonality of the object (mask 131 in FIG. 20) is a detection error.

In the configuration shown in FIG. 20, light perpendicularly incident on the mask 131 is diffracted by the diffraction gratings 133a and 133b, and ± first-order light is incident on the diffraction grating 134, so that the mask 131 is inclined with respect to the incident light. And the directions of ± 1st order light are also inclined. 21st
The figure shows a state where the mask 131 is inclined with respect to the direction of the incident light.

When the incident optical axis is inclined with respect to the mask plane by θ, the phase advances at the same phase, and is transmitted and diffracted at the positions of points A and B on the mask diffraction grating to generate ± 1st order light at points A and C. The state changes to transmission diffraction.

That is, when the incident point moves from the point B to the point C, the optical path length of the light incident and diffracted at the point C is shifted by about Sθ with respect to the phase of the light incident and diffracted at the point B.

The displacement detection error ε due to the optical path length Sθ is expressed by equation (22).

ε = θ · S · d ₁ / 2λ (22) (2). A change in the posture of the object in a direction other than the measurement direction affects the detection accuracy in the measurement direction. That is, in the configuration shown in FIG. 20, a change in the degree of parallelism between the mask 131 and the wafer 132 affects the detection accuracy. Next, consider a case where the wafer 132 is relatively inclined with respect to the mask 131. When the light diffracted by the diffraction grating 133 on the mask 131 is incident on the diffraction grating 134 on the inclined wafer 132, the diffracted light u is deviated from the direction in which the light should be reflected and diffracted.
(1, -1) and u (-1, 2) advance. Therefore diffracted light u (1, -1) and u (-1, 2) is traveling in different directions, the interference beat light Is ₁ is sometimes not occur.

(3). Between objects, there is a correlation between a relative position other than the measurement direction and the pitch of the diffraction grating used between the objects, and the constraints are severe. In FIG. 20, this constraint condition is established between the gap between the mask 131 and the wafer 132 and the pitch of the diffraction grating. That is, the beat interference light I
In order to obtain an intensity that can be detected by s ₁ and Is ₂ , it is necessary that the luminous flux of the ± first-order diffracted light diffracted by the mask 131 be superimposed over a certain area of the diffraction grating 134 on the wafer 132. For this purpose, as the gap becomes narrower, the angle of the ± first-order diffracted light diffracted from the mask must be increased, and therefore, the pitch of the diffraction grating must be reduced.

Further, the pitch of the diffraction grating between the mask 131 and the wafer 132 must be set to a non-integer multiple in order to avoid the influence of multiple reflection diffraction light.

Also, the light diffracted by the wafer 132 interferes on the same path,
The pitch must be set so as to form a beat signal.

When applied to X-ray lithography, conditions from the X-ray source and the process are added to the gap and the pitch of the diffraction grating.

(4). The asymmetry of the diffraction grating causes a detection error. As in the above-described conventional example, light is incident from the left and right directions of the diffraction grating and diffracted light is emitted in a predetermined intermediate direction.
The asymmetry of the groove type of the diffraction grating itself causes a phase error and a detection error.

For details of the conventional example 4, refer to the above-mentioned documents 8) to 14).

[Problems to be Solved by the Invention] As described above, in the optical heterodyne detection in the related art, a diffraction grating having a constant pitch is provided on an object whose displacement is to be measured, and the diffraction grating has slightly different wavelengths.
Displacement was measured by injecting two lights and detecting a beat from an interference signal of the diffracted light.

Each of the conventional techniques has a problem in terms of configuration and accuracy.

An object of the present invention is to provide a displacement detection method based on a new principle employing an optical heterodyne method.

Another object of the present invention is to provide a displacement measuring device for implementing the new displacement measuring method.

[Means for solving the problem]

The displacement measuring method according to the present invention includes a step of preparing one object to be measured in which a plurality of diffraction gratings having different pitches are arranged in close proximity to each other, and first and second light sources each including two wavelengths having slightly different frequencies. And a third incident beam incident on a plurality of diffraction gratings on the object to be measured from first, second and third directions, respectively, and diffracted light of the first and second incident beams. The selected two diffracted lights are made to travel in the fourth direction, and the two diffracted lights selected from the diffracted lights of the first and third incident beams are converted into a fifth symmetrical light in the fourth direction. Direction, forming a beat light of two diffracted lights traveling in the fourth direction, measuring a first phase, and beats of two diffracted lights traveling in the fifth direction. Forming a second phase and measuring a second phase;
Measuring a displacement of the one measured object from a phase difference between the first phase and the second phase.

[Operation] By arranging two diffraction gratings having different pitches in parallel on the same object and making incident beams including two wavelengths incident on these two diffraction gratings, light having more information can be obtained. Various processes can be performed. For this reason, more stable and highly accurate displacement measurement becomes possible.

Using two light beams having slightly different frequencies f ₁ and f ₂ to generate beat light and detecting the phase term at the beat light frequency is similar to a general optical heterodyne detection technique.

Embodiment First, a basic concept of the present invention using two diffraction gratings having different grating pitches will be described with reference to FIG.

The first diffraction grating 1 is P ₁ pitch of the grating, the diffraction grating 2 of the second pitch of the grating is P _2, are disposed parallel to each other. The direction perpendicular to the grating grooves of these diffraction gratings, that is, the direction of the diffraction grating is defined as the X direction. For simplicity, it is assumed that the grating pitch P1 of the _first diffraction grating is half of the grating pitch P2 of the _second diffraction grating. Consider a case where incident light u having slightly different frequencies f ₁ and f ₂ is incident on these parallel diffraction gratings 1 and ₂ vertically. Since the diffraction grating 1 has a half pitch of the diffraction grating 2, the two gratings correspond to one grating of the diffraction grating 2. Therefore, the secondary light of the first diffraction grating 1 travels in the same direction as the primary light of the second diffraction grating. That is, the second-order diffracted light u ₁ (2) by the first diffraction grating and the first-order diffracted light u by the second diffraction grating
_{2 The} two diffracted lights in (1) travel in the same direction. When the pitch of the two diffraction gratings deviates from the ratio of 1: 2, in order to make the diffracted light from the two diffraction gratings travel in the same direction, two incident light beams are taken as shown in FIG. The incident angle may be arbitrarily selected, and the other incident angle may be selected according to one incident angle. Each of these diffracted light and the component of the frequency f ₁
and a component of f _2. If u ₁ (2) and u ₂ (1) interfere with each other, interference light of different orders of diffracted light is formed,
The phase term includes information on the displacement of the diffraction gratings 1 and 2. To make this phase term easy to measure, the frequency f ₁ ,
the beat light of f ₂ may make.

More generally, using an equation, the property between the diffracted lights generated from the diffraction gratings having different pitches can be described as follows.

In the light wave of frequency f ₁ in the first-order diffracted light from the pitch P ₁ and the light wave of frequency f ₂ in the first-order diffracted light from the pitch P ₂ , the phase changes f ₁ and f _{2 with} respect to the movement amount Δx are as follows (23) Expression (24)

δ ₁ = 2π △ x / P ₁ (23) δ ₂ = 2π △ x / P ₂ (24) The phase δn of the diffracted light depends on the position x of the diffraction grating, and δn = 2πnx / d ( (n = diffraction order). According to (23) and (24), the phase difference φ between the + 1st-order diffracted lights generated from different pitches with respect to the movement amount Δx is expressed by Expression (25).

φ = δ ₁ −δ ₂ = 2π △ x / P ₁ −2π △ x / P ₂ = 2π (1 / P ₁ −1 / P ₂ ) △ x (25) Diffraction generated from a different pitch than (25) By measuring the phase difference φ between the lights, it can be seen that the movement amount Δx of the diffraction grating can be obtained. In the case of the nth-order diffracted light, the same equation holds if the pitch P is P / n.

Known techniques use two different orders of magnitude resulting from the same pitch P.
In contrast to measurement using a phase difference between two diffracted lights, this measurement method is a measurement using a phase difference between diffracted lights generated from different pitches.

In the present invention, a measuring method utilizing this property is used, which is completely different from known means.

When diffracted lights of the same diffraction order are extracted from diffraction gratings having different pitches, the diffraction angles formed by the incident light and the diffracted light become different.

 In this case, the following characteristics are obtained.

(1). For a plurality of objects, it is possible to measure displacement in a certain direction, for example, displacement on the order of nm with high accuracy.

(2). The displacement of the posture of the object other than the measurement direction does not affect the accuracy of the measurement direction at all.

(3). Changes in the relationship between the measuring instrument or the position between the measuring optical path and the object have no effect on the measuring accuracy.

(4). The measurement system is extremely simple.

(5). It does not depend on the stability of the measurement system.

(6). The branch of the measurement optical path is essentially unnecessary, and is hardly affected by disturbance.

2. Measuring Method According to Embodiment of the Present Invention 2.1 General Description A displacement measuring method based on the above principle will be described with reference to FIG.

Three incident light beams I, II and III are incident on two diffraction gratings 1 and 2 having different pitches (grating constants) P ₁ and P ₂ . Left and right ray II with respect to central ray I at normal incidence
And III are incident on the diffraction gratings 1 and 2 at an angle θ ゜.

Consider the condition that the diffracted light is emitted in the same direction, with the pitches P ₁ and P 2 of the two diffraction gratings 1 and ₂ being P ₁ <P ₂ . Ray I
And the light beam III, and the light beam II can be considered in the same manner as the light beam III.

Diffraction grating 1 (pitch P ₁ ) for vertically incident light I
It is When diffract light in the direction of the theta _1, the _{nλ = P 1 Sinθ 1 ... (} 26). When the diffraction grating 2 (pitch P ₂ ) generates the diffracted light of the light ray III in the same θ ₁ direction, if the incident angle θo is opposite to the diffraction angle with respect to the normal, nλ = P ₂ (Sin θ ₁ − Sinθo) (27) That is, θ ₁ = Sin ⁻¹ (nλ / P ₁ ) θo = Sin ⁻¹ {(nλ / P ₁ ) − (nλ / P ₂ )} (28) Considering the first-order diffracted light (n = 1), θ ₁ = Sin ⁻¹ (λ / P ₁ ) θo = Sin ⁻¹ {(λ / P ₁ ) − (λ / P ₂ )} (29)

The incident light beam I is applied to the incident light beam I by means of a half-wave plate or the like.
The polarization direction is rotated by 90 ° with respect to I and III.

By setting the above conditions, two interference beat lights I + and I- as shown in FIGS. 3A and 3B are obtained. I + is the + _1st-order diffracted light {f ₁ (P ₁ , 1), f ₂ (P ₁ , 1)} of the incident light beam I by the diffraction grating 1 with the pitch P 1 and the +1 of the incident light beam II by the diffraction grating 2 with the pitch P ₂ Order diffracted light ｛f ₁ (P ₂ , 1), f ₂
(P ₂ , 1) is a beat light generated by interference.

Further, I- incident by the diffraction grating 1 of the pitch P ₁ light I
-1-order diffracted light _{{f 1 (P 1, -1} ), f 2 (P 1, -1)} -1-order diffracted light of the incident light beam III by the diffraction grating 2 of the pitch _{P 2 {f 1 (P 2} , _{-1), f 2 (P 2} , -1)} is a beat light produced by interference.

Here, by measuring the phase difference φφ = [phase of I +] − [phase of I−] (30) of the two beat lights I + and I−, it becomes possible to precisely measure the relative position of the object. . Further, when two diffraction gratings are provided for each of a plurality of objects, the relative position between the objects can be measured.

Next, the principle of measurement using the interference beat light beams I + and I− obtained by setting the above conditions will be described.

2.2 Principle of measurement When an incident light beam is branched into three, a phase difference occurs between the incident light beams.

Therefore, the phase difference between the incident light I and the incident light II is φ ₁ , and the phase difference between the incident light I and the incident light III is φ ₂ .

Beat light I + is formed by light waves (P _{1, 1)} and (P _2, 1). The light wave (P ₁ , 1) is represented by the following two equations.

f ₁ (P ₁ , 1) = A ₁ exp {i (w ₁ t + δ ₁ )} (31) f ₂ (P ₁ , 1) = A ₂ exp {i (w ₂ t + δ ₁ )} (32) The light wave (P ₂ , 1) is represented by the following two equations.

f ₁ (P ₂ , 1) = B ₁ exp {i (w ₁ t + δ ₂ + φ ₁ )} (33) f ₂ (P ₂ , 1) = B ₂ exp {i (w ₂ t + δ ₂ + φ ₁ )} … (34) where A ₁ , A ₂ , B ₁ , and B ₂ are amplitudes (A ₁ ≠ A ₂ , B ₁ ≠ B ₂ , A ₁
≠ B ₁ ). Also, w ₁ = 2πf ₁ (35) w ₂ = 2πf ₂ (36) δ ₁ = 2πx / P ₁ (37) δ ₂ = 2πx / P ₂ (38) (31) to (34) As a result, two beat lights I + (0) and I + (π / 2) whose polarization directions are different by 90 ° are obtained as shown in FIG. 3 (A).

I + (0) = {f ₁ (P ₁ , 1) + f ₂ (P ₂ , 1)} ² = [A ₁ expwi (w ₁ t + δ ₁ ) ｛+ B ₂ exp 2i (w ₂ t + δ ₂ + φ ₁ )｝] ² = A ₁ ² B ₂ ² + 2A ₁ B ₂ Cos ｛(w ₁ −w ₂ ) t + δ ₁ − (δ ₂ + φ ₁ ]｝ (39) I + (π / 2) = ｛f ₂ (P ₁ , 1) + f ₁ (P ₂ , 1)｝ ² = [A ₂ exp ｛i (w ₂ t + δ ₁ )｝ + B ₁ exp ｛i (w ₁ t + δ ₂ + φ ₁ )｝] ² = A ₂ ² + B ₁ ² + 2A ₂ B ₁ Cos {(w ₂ −w ₁ ) t + δ ₁ − (δ ₂ + φ ₁ )} (40) On the other hand, the beat light I− is divided into light waves (P ₁ , −1) and (P ₂ , −2).
1) is formed.

Lightwave (P _1, -1) is represented by the following two equations.

_{f 1 (P 1, -1)} = A 1 exp {i (w 1 t-δ 1)} ... (41) f 2 (P 1, -1) = A 2 exp {i (w 2 t-δ 1 )｝ (42) The light wave (P ₂ , −1) is represented by the following two equations.

f ₁ (P ₂ , −1) = B ₁ exp {i (w ₁ t−δ ₂ + φ ₂ )} (43) f ₂ (P ₂ , −1) = B ₂ exp {i (w ₂ t−) δ ₂ + φ ₂ )｝ (44) FIG. 3 (B) shows two beat lights I- (0) and I- (π / 2) having polarization directions differing by 90 ° according to the equations (41) to (44). Obtained as shown.

_{I- (0) = {f 1} (P 1, -1) + f 2 (P 2, -1)} 2 = [A 1 exp {i (w 1 t-δ 1)} + B 2 exp {i (w _{2 t-δ 2 + φ 2} )}] 2 = A 1 2 + B 2 2 + 2A 1 B 2 Cos {(w 1 - w 2) t-δ 1 - (- δ 2 + φ 2)} = A 1 2 + B 2 ^{_{2 + 2A 1 B 2 Cos {}} (w 1 - w 2) t-δ 1 + δ 2 -φ 2} ... (45) I- (π / 2) = {f 1 (P 1, -1) + f 2 (P ₂ , -1)｝
² = [A ₂ exp {i (w ₂ t−δ ₁ )} + B ₁ exp {i (w ₁ t−δ ₂ + φ ₂ )}] ² = A ₂ ² + B ₁ ² + 2A ₂ B ₁ Cos} (w _{2 - w 1) t-δ} 1 - (- δ 2 -φ 2)} = A 2 2 + B 1 2 + 2A 2 B 1 Cos {(w 2 - w 1) t-δ 1 + δ 2 -φ 2} ... (46) I + (0) and I− (0) whose polarization directions are the same in the beat lights I + and I− are extracted by the polarizers 5 and 6, and the phase difference φ between the beats is expressed by the equations (39) and (45). Find more.

φ = δ ₁ − (δ ₂ + φ ₁ ) − (− δ ₁ + δ ₂ −φ ₂ ) = δ ₁ −δ ₂ −φ ₁ + δ ₁ −δ ₂ + φ ₂ = 2 (δ ₁ −δ ₂ ) −φ ₁ + Φ ₂ ... (47) By substituting the expressions (37) and (38) into the expression (47), φ = 2 (δ ₁ −δ ₂ ) −φ ₁ + φ ₂ = 4π ｛(1 / P ₁ ) − (1 / P ₂ )｝ x−φ ₁ + φ ₂ (48) From equation (48), it can be seen that the displacement x can be detected by measuring the phase difference φ.

Note that (−φ ₁ + φ ₂ ) is an initial value determined at the time of incidence and is a constant.

 The above is the measurement principle.

3. Phase change due to movement in the Y direction Light flux (P ₁ ,
Interference beam light I generated by interference between 1) and (P ₂ , 1)
The fact that + and I- have no phase information for movement in the Y direction will be described below.

 The phase change accompanying the movement in the Y direction is defined as δy.

δy = 2πny / 2a (49) where n is the diffraction order, y is the moving amount, and 2a is the distance between the diffraction gratings. In the beat light I +, this phase change δy
, The light beam (P _1, + ₁₎ and (P _2, + 1) together from being included as many, beat light I + (0) is expressed by the following equation.

I + (0) = {f 1 (P 1, 1) + f 2 (P 2, 1)} 2 = [A 1 exp {i (w 1 t + δ 1 + δy)} + B 2 exp {i (w 2 t + δ 2 + φ ^{_{1 + δy)}] 2 =}} A 1 2 + B 2 2 + 2A 1 B 2 Cos {(w 1 - w 2) t + δ 1 - (δ 2 + φ 1)} ... (50) (50) a phase change .delta.y is erased from the equation Has been (30)
The result is the same as the formula. That is, the measurement result is not affected by the movement in the Y direction. Remaining beat light I +
Similarly, for (π / 2), I− (0), and I− (π / 2), the phase change δy is canceled, so that I + (0) and I− (0)
The phase difference (0) and the phase difference between I + (π / 2) and I− (π / 2) do not include the phase change δy in the Y direction. Therefore,
Even when the object moves in the Y direction, the beat lights I + and I-
Is expressed by the equation (47), so that there is no change in the movement in the Y direction.

4. Diffraction image of interference beat light Y-direction diffraction images of interference beat light I + and I- are obtained. The interference beat light I + includes two light beams (P ₁ , 1) and (P ₂ ,
1) occurs as a diffraction image that interferes in the Y direction. Here, the diffraction image is obtained.

In the XZ plane, light beams (P ₁ , 1) and (P ₂ , 1) traveling in the same direction (θ ₁ = θ ₂ ) and separated by 2a in the Y direction
Is considered as transmitted light from a double slit with a width of 2 ° and an interval of 2a, the diffraction image of the transmitted light is the diffraction image of the interference beat light I +, and the Fraunhofer diffraction image along the Y axis is expressed by the following equation (51). Is represented by

I = Io [{Sin (2πξy / λf)} / (2πξ y / λf)] Cos 2 (2πay / λ 1 f) ... (51) However, Io represents a constant term.

The diffraction image of equation (51) has the shape shown in FIG. 4, and has a maximum value at the position of yn = nλ (f ₁ −f ₂ ) / 2a (51a). Here, n represents the diffraction order as an integer. This diffraction image has the same phase difference with respect to the movement in the X direction regardless of the diffraction order n, as described in the above-mentioned phase change in the Y direction. Note that to be precise, the wavelength and the frequency in the equations (51) and (51a) are (λ
₁ , λ ₂ ) and (f ₁ , f ₂ ), but the difference frequency is usually several hundred KHZ to several MHZ, and assuming that they are almost the same, λ ₁ ≒ λ ₂ ≒ λ, f ₁ ≒ ₂ ≒ f was determined.

5. Change in posture of object In FIG. 2, the effect of a change in posture other than the measurement direction (X direction) on the measurement accuracy will be described.

5.1 Change of posture in the X-Z plane In the X-Z plane, when the object is tilted by △ θ, two light beams (P ₁ , ± 1) forming the beat light I + and I-,
Since both (P ₂ , ± 1) are tilted by about Δθ, good beat light I + and I− can be continuously obtained. This is shown in FIG. A phase change Δδ caused by tilting Δθ does not occur in principle.

5. Change in posture in the 2YZ plane Beat lights I + and I when the object is tilted by △ θ in the YZ plane
The state of-is shown in FIG. 5 (B).

From FIG. 5 (B), the luminous fluxes (P ₁ , P ₂₎ forming the beat light I +, I− by tilting the object by Δθ in the YZ plane.
An incident optical path difference Δl occurs between (± 1) and (P ₂ , ± 1).

Since the incident optical path difference △ l by phase △ [delta] delayed light is incident on the diffraction grating 1 Pippi P _1, the light beam (P _1, ±
In 1), the phase is delayed by △ δ compared to the light beam (P ₂ , ± 1).

Δδ = 2π △ l / λ (52) From this, the incident light flux (P ₁ , ± 1) is expressed by the following equations.

The light beam _{(P 1, 1), f} 1 (P 1, 1) = A 1 exp {i (w 1 t + δ 1 + △ δ)} ... (53) f 2 (P 1, 1) = A 2 exp { i (w ₂ t + δ ₁ + {δ)} (54) For the light flux (P ₁ , −1), f ₁ (P ₁ , −1) = A ₁ exp {i (w ₁ t + δ ₁ + {δ)} _{... (55) f 2 (P} 1, -1) = A 1 exp {i (w 2 t + δ 1 + △ δ)} ... (56) beat light I + (0) from now (34) and (53) below It is expressed by the following equation.

I + (0) = {f ₁ (P ₁ , 1) + f ₂ (P ₂ , 1)} ² = A ₁ expwi (w ₁ t + δ ₁ + △ δ)｝ + B ₂ exp ｛i (w ₂ t + δ _{2 ^{+ φ 1)} 2 = A}} 1 2 + B 2 2 + 2A 1 B 2 Cos {(w 1 - w 2) t + δ 1 + △ δ- (δ 2 + φ 1)} also ... (57), beat light I- (0 ) Is expressed by the following equation from equations (44) and (55).

I− (0) = {f ₁ (P ₁ , −1) + f ₂ (P ₂ −1)} ² = [A ₁ expi (w ₁ t−δ ₁ + Δδ) + B2 expi (w ₂ t− ^{_{δ 2 + φ 2)} 2}} = A 1 2 + B 2 2 + 2A 1 B 2 Cos {(w 1 - w 2) t-δ 1 + △ δ - (- δ 2 + φ 2)} ... (58) (57) From equation (58), beat light I + (0) and I−
The phase difference φ of (0) is expressed by the following equation.

φ = I + (0) −I− (0) = δ ₁ + Δδ− (δ ₂ + φ ₁ ) − ｛− δ ₁ + Δδ − (− δ ₂ + φ ₂ )｝ = δ ₁ + Δδ−δ ₂ −φ ₁ + δ ₁ − △ δ− δ ₂ + φ ₂ = 2 (δ ₁ −δ ₂ ) −φ ₁ + φ ₂ (59) From equation (59), the term of the phase difference △ δ due to the gradient 、 θ is It can be seen that the difference is not included in the phase difference φ and is the same as Expression (47). That is, the phase difference φ is not affected at all by the inclination △ θ in the YZ plane.

Beat light I + (π / 2) and I 90 ° different in polarization direction
-(Π / 2) is similarly calculated for (33), (54) and (4
3) and (56) that the phase difference φ does not include the term of the phase difference Δδ due to the gradient Δθ.

5. Change in posture in the 3Z direction In FIG. 2, when the object moves by △ Z in the Z direction, with respect to the light beam I incident perpendicularly to the diffraction grating, the two light beams II and III incident from a symmetric angle θo Both phases are △ δ
Only change.

Δδ = 2π △ Z / λ (60) In FIG. 5 (C), the phase change Δδ due to the change of Z is the beam diameter of the incident light flux (I), (II), (III). The change in φD. That is, the variation of ΔZ is 3
It is limited to a range where a good beat interference wave can be obtained with the beam diameter φD of the light beam. Therefore, ΔZ is an amount determined from the beam diameter φD, the incident angle θo, and the like.

Since the phase of the diffracted light (P ₂ , ± 1) due to P ₂ changes by Δδ with respect to the diffracted light (P ₁ , ± 1) due to P ₁ due to the phase change at the time of incidence ΔP (P ₂ , ± 1) is represented by the following equation.

Light wave (P ₂ , +1): f ₁ (P ₂ , 1) = B ₁ exp {i (w ₁ t + δ ₂ + φ ₁ + △)
δ)｝ (61) f ₂ (P ₂ , 1) = B ₂ exp {i (w ₂ t + δ ₂ + φ ₁ + △)
δ)｝ (62) Light wave (P ₂ , -1): f ₁ (P ₂ , -1) = B ₁ exp {i (w ₁ t -δ ₂ + φ ₂ + △ δ)} ... (63) f ₂ (P ₂ , −1) = B ₂ exp {i (w ₂ t−δ ₂ + φ ₂ + {δ)} (64) The beat light I + (0) is given by the following equation from the equations (31) and (62). Is represented by

I + (0) = {f ₁ (P ₁ , 1) + f ₂ (P ₂ , 1)} ² = [A ₁ expwi (w ₁ t + δ ₁ ) ｛+ B ₂ exp 2i (w ₂ t + δ ₂ + φ ₁ ^{+ △ δ)}] 2 =} A 1 2 + B 2 2 + 2A 1 B 2 Cos {(w 1 - w 2) t + δ 1 - (δ 2 + φ 1 + △ δ)} ... (65) On the other hand, beat light I- (0) is represented by the following equation from equations (41) and (64).

_{I- (0) = {f 1} (P 1, -1) + f 2 (P 2, -1)} 2 = [A 1 exp {i (w 1 t-δ 1)} + B 2 exp {i (w _{2 t-δ 2 + φ 2} + △ δ)}] 2 = A 1 2 B 2 2 + 2A 1 B 2 Cos {(w 1 - w 2) t-δ 1 - (- δ 2 + φ 2 + △ δ)} (66) From the equations (65) and (66), the phase difference φ between the beat light I + (0) and the beat light I− (0) is expressed by the following equation.

φ = δ _1- (δ ₂ + φ ₁ + △ δ)-｛-δ ₁ -(-δ ₂ + φ ₂ + △ δ)｝ = δ ₁ -δ ₂ -φ _1- △ δ + δ ₁ -δ ₂ + φ ₂ + △ δ = 2 (δ ₁ −δ ₂ ) −φ ₁ + φ ₂ (67) Expression (67) does not include the term of the phase difference Δδ, is the same as Expression (47), and moves in the Z direction. It is clear that no changes have been made with this. Beat light I + with 90 ° different polarization direction
Similarly, for (π / 2) and I- (π / 2), (32) and (6
The phase difference φ is obtained from the equations (1), (42) and (63), and it is understood that the phase change due to the movement in the Z direction does not occur without including the term of the phase difference Δδ.

6. Change of incident optical axis In FIG. 2, the influence on the measurement accuracy when the incident optical axis changes with respect to the object will be described. However, the angle between the incident lights I, II, and III is invariable, and the case where the angle between the incident lights I, II, and III changes is considered. explain).

6. Change of incident optical axis in 1X-Z plane Changing the incident optical axis in XZ plane is equivalent to XZ
Equivalent to changing the attitude of the object in the plane 5.
As described in 1, the detection accuracy in the X direction is not affected.

6. Change of incident optical axis in 2YZ plane Changing the incident optical axis in the YZ plane
-Equivalent to a change in the attitude of the object in the Z plane,
As described above, it does not affect the detection accuracy in the X direction.

6. Change in the incident optical axis in the 3Z direction Changing the incident optical axis in the Z direction is equivalent to changing the object in the Z direction, and does not affect the detection accuracy in the X direction as described above.

However, the movement amount ΔZ in the Z direction is limited to the region where the three incident light beams overlap as described above.

7. Breakage of symmetry of incident optical axis FIG. 6 shows a case where symmetry is broken at the incident optical axis.

In FIG. 6, the angle between the incident light I and the incident lights II and III is θ ₁ ≠ θ ₂ (68) Δθ = θ ₁ -θ ₂ (69) (θ ₁ > θ ₂ ). This loss of symmetry causes a detection error when the posture of the object changes in the Z direction described above. If the object moves by Z in the Z direction, the incident light
The light path of II is l ₁ and the light path of incident light III is l ₂ (l ₁ <l ₂ ), and this optical path difference Δl causes a new phase difference Δδ between the incident lights II and III.

Δl = l ₂ −l ₁ (70) Δ = 2π △ l / λ (71) where l ₁ = ΔZCosθ ₁ (72) l ₂ = ΔZCosθ ₂ (73) (72) , (73) into equation (71), Δδ = 2π △ Z / λ (Cosθ ₂ −Cosθ ₁ ) (74) This phase difference Δδ is the pitch P affected by the change in the incident angle. occurs between ₂ ± 1-order diffracted light (when relative to the ± 1-order light of the pitch P _1). This is shown in FIG.

The ± 1-order light incident light I is diffracted incident on the diffraction grating of pitch P ₁ in FIG. 7 to FIG. 7 (A), the incident light II is diffracted incident on the diffraction grating of pitch P ₂ +1 Next light and incident light
III indicates the -1 order light diffracted incident on the diffraction grating of pitch P ₂ in FIG. 7 (B). However, in FIG. 7 (B), θ ₄ = θ ₃ + △ θ (75) In FIG. 7 (B), the + 1st-order diffracted light (P ₂ , 1) from the pitch P ₂ is represented by the following equation. .

f ₁ (P ₂ , 1) = B ₁ exp {i (w ₁ t + δ ₂ + φ ₁ + {δ)} (76) f ₂ (P ₂ , 1) = B ₂ exp {i (w ₂ t + δ ₂₎ + Φ ₁ + {δ)} (77) The beat light I + (0) is expressed by the following equation from the equations (31) and (77).

I + (0) = {f ₁ (P ₁ , 1) + f ₂ (P ₂ , 1)} ² = [A ₁ expwi (w ₁ t + δ ₁ ) ｛+ B ₂ exp 2i (w ₂ t + δ ₂ + φ ₁ ^{+ △ δ)}] 2 =} A 1 2 + B 2 2 + 2A 1 B 2 Cos {(w 1 - w 2) t + δ 1 - (δ 2 + φ 1 + △ δ)} ... (78) On the other hand beat light I- ( (0) is expressed by equation (45), so that the phase difference φ between I + (0) and I− (0) is (45), (7)
From equation (8), it is expressed by the following equation.

φ = δ ₁ − (δ ₂ + φ ₁ + △ δ) − (− δ ₁ + δ ₂ + φ ₂ = δ ₁ −δ ₂ −φ ₁ − △ δ + δ ₁ −δ ₂ + φ ₂ = 2 (δ ₁ −δ ₂ ) −φ ₁ + φ ₂ − △ δ (79) Comparing Equations (79) and (47), it has been found that the phase difference Δδ due to the broken symmetry becomes a measurement error due to movement in the Z direction. .

7.1 Allowable Value for Symmetry Breaking Next, an example of an allowable value in the phase change due to the symmetry breaking represented by the equation (74) is shown below.

Since the accuracy of the phase difference meter is ± 2 ° for the standard type and ± 1 ° for the precision type, the phase change {δ} expressed by the equation (74) is 1 °.

Δδ = 1 ゜ (2π / 360 radians) (80) From this, the conditional expression for the symmetry breakdown is expressed by the following expression.

2π / 360> (2π △ Z / λ) (Cosθ ₂ −Cosθ1) (81) From equation (69), θ ₂ = θ ₁ − △ θ (82) (82) is substituted into (81).

1/360> (Z / λ) ｛Cos (θ ₁ − △ θ) −Cosθ ₁ … (83) In (83), ΔZ = 5 μm, λ = 0.633 μm, P ₁
= 3 μm, θ ₁ = Sin ^-1 (0.633 / 3) = 12.18 ゜ from Equation (26) (84) Substituting (84) into (83), obtain △ θ.

Δθ ≒ 0.1 ゜ (85) From (85), the symmetry breakdown in this case is allowed up to 0.1 °.

8. Influence of Asymmetry of Diffraction Grating In FIG. 2, beat light I + includes (P ₁ , 1) and (P ₂ ,
1), and the other beat light I− is (P ₁ , −
1) and (P ₂ , -1).

That is, beat light is formed by the diffracted lights having the same sign. This property is completely different from the prior art described above, and is not affected by the asymmetry of the diffraction grating as described below.

As described above, the conventional method is affected by the asymmetry of the diffraction grating. However, the arrangement of the + _1st-order diffracted light (P ₁ , 1), (P ₂ ,
1) and the -1st-order diffracted light (P _1, -1), and changes by phase △ [delta] between the (P _2, -1).

_{(P 1, 1) ~ (} P 1, -1): △ δ (P 2, 1) ~ (P 2, -1): △ δ (P 1, 1) and (P _2, -1) The following It is expressed by an equation.

Lightwave _{(P 1, -1): f} 1 (P 1, -1) = A 1 exp {i (w 1 t-δ 1 + △ δ)} ... (86) f 2 (P 1, -1) = A ₂ exp {i (w ₂ t−δ ₁ + Δδ)} (87) Lightwave (P ₂ , −1): f ₁ (P ₂ −1) = B ₁ exp {i (w ₁ t−) δ ₂ + φ ₂ + {δ)｝ (88) f ₂ (P ₂ , −1) = B ₂ exp {i (w ₂ t−δ ₂ + φ ₂ + △ δ)} (89) I- (0) is represented by the following equation from equations (86) and (89).

I− (0) = {f ₁ (P ₁ −1) + f ₂ (P ₂ −1)} ² = [A ₁ exp {i (w ₁ t−δ ₁ + {δ)} + B ₂ exp} _{i (w 2 t-δ 2} + φ 2 + △ δ)}] 2 = A 1 2 + B 2 2 + 2A 1 B 2 Cos {(w 1 - w 2) t-δ 1 - (- δ 2 + φ 2)} (90) On the other hand, since the beat light I + (0) is expressed by the equation (39), the phase difference φ between I + (0) and I− (0) is expressed by the following equation.

φ = δ ₁ − (δ ₂ + φ ₁ ) − {− δ ₁ − (− δ ₂ + φ ₂ )} = 2 (δ ₁ −δ ₂ ) −φ ₁ + φ ₂ (91) From equation (91), the phase difference φ is the phase change due to asymmetry △ δ
It is evident that they do not contain and are not affected by asymmetry.

9. Specific Examples 9.1 Incident System In FIG. 2, a method of obtaining incident light (I), (II), and (III) is shown in FIG.

In FIG. 8, the laser beam 15 incident on the beam expander 11 becomes a beam 15a having a spot diameter expanded from φd to φD, and is divided into three light beams by the spatial filter 12. The split central light flux I is rotated by 90 ° in the polarization direction by the λ / 2 plate 13 and enters the lens 14 together with the other two light fluxes II and III. The three light beams incident on the lens 14 are diffracted by the diffraction gratings 1 and 2 at an angle determined by the numerical aperture (NA) of the lens.
(P ₁ , P ₂ ).

In FIG. 8, the incident angle θo determined by the numerical aperture (NA) of the lens 14 is set equal to the second incident angle θo.

9.2 Detection system Figure 9 shows the detection system.

In order to detect the beat lights I + and I−, the detection system includes two photoelectric detectors 3 and 4 having polarizers 5 and 6, respectively.
I just need a table. A photomultiplier tube is appropriate as the type of the photoelectric detector, but an avalanche photodiode or the like can also be used. Further, the attitude of the photoelectric detector does not affect the phase detection accuracy of the beat lights I + and I-.

9.3 Application to X-ray exposure apparatus FIG. 10 shows an example in which the above-described incident system and detection system are applied to an X-ray exposure apparatus relative position detection apparatus (hereinafter, alignment apparatus) for an X-ray mask.

In the X-ray exposure apparatus shown in FIG. 10, the same two types of parallel diffraction gratings (pitch P ₁ and pitch P ₁₎ are provided on the X-ray mask 31 and the wafer 33.
P ₂ ) 32 and 34 are provided, and the system shown in FIGS. 8 and 9 is adopted as the incident system and the detection system, and the incident light (I, II, II
Photoelectric detectors 21, 2 with polarizers 26, 27, 28, 29 from I)
Detection light (M (I +, I−): beat light from the mask diffraction grating, W (I +, I−): beat light from the wafer diffraction grating) is obtained by 2, 23, and 24.

FIG. 10 (A) shows the detection light, and the mask grating 32
And four beat lights M formed from the wafer grating 34
(I +), M (I-) and W (I +), W (I-) are detected by four photoelectric detectors 21, 22, 23, 24, respectively.
Detection angle theta ₄ is a value determined from the pitch P ₁ of the grating. P ₁
= 3 μm, the first-order diffracted light is used as the interference light, and the wavelength λ of the interference light is 0.633 μm (Hc-Ne laser).
From equation (6), θ ₄ = Sin ^-1 (0.633 / 3) = 12.18 ゜ (92).

FIG. 10 (B) shows that incident light (I, II, III) and detection light (M (I +, I−), W (I +, I−) are angle θ with respect to the normal to the mask / wafer surface. represents a side view of the set in _3. the incident / measured angle theta ₃ is any angle inclined to provide the incident system and the detection system outside the X-ray region. Note that the X-ray mask Wafer has a gap of Gμm
Away at. Here, θ ₃ = 30 ゜ (93) G = 40 μm (94)

FIG. 10 (C) shows the incident light (I, II, III). The diameter of the incident light beam is φD, and a mask and a wafer diffraction grating are provided in a region where the three light beams overlap. Incident light beam II and III are incident at an angle theta ₁ and theta _2.

Compared with FIG. 2, θo ≒ θ ₁ , θ ₂ (95) Incident angle error: Δθ = θ ₁ −θ ₂ (96) Here, assuming that the diffraction grating P ₁ = 3 μm and P ₂ = 5 μm, from equation (29), θo = Sin ⁻¹ {(0.633 / 3) − (0.633 / 5)} = 4.84} (97) The incident angle error △ θ is obtained from equation (83).

In FIG. 10, the gap G between the mask and the wafer is shown.
Is controlled within ± 3 μm using a capacitance sensor or the like, the equation (83) is expressed by the following equation.

1/360> 6 / 0.633 ｛Cos (12.8 ゜-△ θ) -Cos12.8 ゜｝ (98) ∴ △ θ = 0.076 ゜ (99) Incident system shown in FIG. , The incident angle θo is (99)
From the formula, 4.834 ゜ ≦ θo ≦ 4.986 ゜ (100).

Since this angle setting can be set relatively easily, practical detection is possible.

If the phase change due to the gap G between the mask and the wafer is φG, (79) is expressed by the following equation.

φ = 2 (δ ₁ −δ ₂ ) −φ ₁ + φ ₂ − △ δ + φG (101) This φG is essentially equal to △ δ and has already been described. However, φG differs from △ δ in that φG is managed by a capacitance sensor or the like. That, [phi] G is can be regarded as a constant with phi ₁ and phi _2. When the phase difference φM of the beam light M (I +, I−) from the mask and the phase difference φW of the beat light W (I +, I−) from the wafer are expressed by the following equations, respectively.

φM = 2 (δ ₁ −δ ₂ ) −φ ₁ + φ ₂ −Δδ (102) φW = 2 (δ ₁ −δ ₂ ) −φ ₁ + φ ₂ −Δδ + φG (103) (102), (103) The relative position P between the mask and the wafer
(X) is given by P (x) = φM−φW = 2 (δ ₁ −δ ₂ ) −2 (δ ₁ −δ ₂ ) −φG = 4π (1 / P ₁ −1 / P ₂₎ ) XM−4π (1 / P ₁ −1 / P ₂ ) xW−φG = 4π (1 / P ₁ −1 / P ₂ ) (xM−xW) −φG (104) where xM = mask position change XW = wafer position change.

It can be seen from (104) that the relative position between the mask and the wafer can be measured by measuring the phases φM and φW between the mask and the wafer. Further, in the case of (104), in the case of an alignment apparatus, since φG can be accurately known by exposure, correction is easily performed.

Also, considering (100) in the above, (102),
(103) is represented by the following equation.

φM = 2 (δ ₁ −δ ₂ ) −φ ₁ + φ ₂ (105) φW = 2 (δ ₁ −δ ₂ ) −φ ₁ + φ ₂ + φG (106) 展開 δ = 0 And will be obvious to those skilled in the art.

Further, consider the case where the relative position in the Z direction between the incident optical axis and the mask / wafer changes in FIG. 10 (C). When the incident optical axis is completely θ ₁ = θ ₂ , no problem occurs as described in the previous section. On the other hand, the incident optical axis is θ ₁ = θ ₂
In the case of, a phase error 前 δ occurs as shown in (79) as described in the previous section. Here, since the relative position between the incident optical axis and the mask / wafer changes, the mask phase difference φ
The same amount of phase error △ δ is added to M and wafer phase difference φW.

φM = 2 (δ ₁ −δ ₂ ) −φ ₁ + φ ₂ −Δδ (107) φW = 2 (δ ₁ −δ ₂ ) −φ ₁ + φ ₂ −Δδ (108) (107), (108) ) Relative position P between mask and wafer
In the case of (x), △ δ is erased, and thus does not become a detection error.

P (x) = φM−φW = 4π (1 / P ₁ −1 / P ₂ ) (XM−XW) (109)

In addition, although this invention was demonstrated along Example, this invention is not limited to these. For example, although the case where ± 1st-order diffracted light is used has been described as an example, higher-order diffracted light may be used. Also, a case has been described in which the incident light is made into three light beams, one of which is vertically incident, and the other two light beams are made to enter from a symmetrical angle. However, other arrangements are also possible.
Also, relative displacement of three or more objects can be measured.
It will be apparent to those skilled in the art that various other modifications, improvements, combinations, and the like can be made.

[Effects of the Invention] As described above, according to the present invention, a stable and highly accurate displacement measurement can be performed by a displacement measurement method using a diffraction grating having a new configuration.

[Brief description of the drawings]

FIG. 1 is a diagram showing a basic embodiment of the present invention, FIG. 2 is a layout diagram schematically showing an optical system for explaining a displacement measuring method according to an embodiment of the present invention, FIG. B) is a component diagram showing the interference light obtained by the arrangement of FIG. 2 divided into polarization components, FIG. 4 is a graph showing the distribution of the intensity of the diffracted light from the two parallel diffraction gratings, FIG. (B) and (C) are diagrams for explaining a change in diffracted light due to a change in the attitude of the diffraction grating, FIG. 6 is a diagram for explaining the symmetry breakdown of the incident optical axis, and FIG. 7 (A). ) And (B) are diagrams for explaining the effect of the change of the incident angle for the three light beams, FIG. 8 is a diagram for explaining the incident system in the specific embodiment, and FIG. 9 is a diagram in the specific embodiment. Diagram for explaining the detection system, FIG. 10 (A), 11B and 11C are diagrams for explaining an optical system of an alignment apparatus according to another embodiment of the present invention. FIGS. 11A and 11B are diagrams for explaining the operation of a diffraction grating. FIG. 12 is a diagram for explaining a first example of the prior art, FIG. 13 is a diagram for explaining beat light, FIG. 14 is a diagram for explaining a second example of the prior art, FIG. FIG. 16 is a diagram for explaining a third example of the prior art, FIG. 16 is a diagram for explaining a problem in the optical system of FIG. 15, and FIG. 17 is a diagram for explaining an effect of an asymmetric diffraction grating. 18 to 20 are diagrams for explaining another example of the prior art, and FIG. 21 is a diagram for explaining a problem of the optical system shown in FIG. In the figure, 1, 2, 32, 34 ... Diffraction grating 3, 4, 21, 22, 23, 24 ... Photoelectric detector 5, 6, 26, 27, 28, 29 ... Polarizer 11 ... Beam extract Panda 12 Spatial filter 13 λ / 2 plate 14 Laser beam 31 Mask 33 Wafer P ₁ , P ₂ Pitch of diffraction grating I +, I− Interference light

Claims (4)

(57) [Claims] 1. A step of preparing an object to be measured in which a plurality of diffraction gratings having different pitches are arranged close to each other, and first, second and third light sources each including two wavelengths having slightly different frequencies. Are incident on the plurality of diffraction gratings on the object to be measured from the first, second, and third directions, respectively, and are selected from the diffracted light of the first and second incident beams. The two diffracted lights are made to travel in a fourth direction, and two of the diffracted lights of the first and third incident beams are converted into a fifth diffracted light symmetrical to the fourth direction.
A step of forming a beat light of two diffracted lights traveling in the fourth direction and measuring a first phase; and a step of measuring the first phase of the two diffracted lights traveling in the fifth direction. A displacement measuring method, comprising: forming a beat light and measuring a second phase; and measuring a displacement of the one measured object from a phase difference between the first phase and the second phase. 2. A displacement measuring device for measuring a displacement of one object to be measured in which a plurality of diffraction gratings having different pitches are arranged close to each other, wherein light of two wavelengths having slightly different frequencies are respectively measured. Incident on a plurality of diffraction gratings on the object to be measured from first, second, and third directions, respectively, including the first, second, and third beams. The two diffracted lights selected from the diffracted lights of the incident beams of the first and third incident beams travel in the fourth direction, and the two diffracted lights selected from the diffracted lights of the first and third incident beams are converted into the fourth direction. And symmetrical fifth
An incident optical system that travels in the direction of. A first light receiving system that forms a beat light of two diffracted lights traveling in the fourth direction and measures a first phase, and travels in the fifth direction. A second light receiving system that forms a beat light of two diffracted lights and measures a second phase, a first phase supplied from the first light receiving system, and a second light receiving system supplied from the second light receiving system A displacement detection system for determining a phase difference from the second phase to be measured and measuring a displacement of the one measured object from the phase difference. 3. A step of preparing first and second objects to be measured, each of which has a plurality of diffraction gratings having different pitches arranged close to each other; Disposing the first, second and third incident beams, each including two wavelengths of light having slightly different frequencies, such that the diffraction gratings are disposed in close proximity to each other; From two and third directions, the light is incident on a plurality of diffraction gratings on the first and second measured objects, and selected from the diffracted light of the first and second incident beams on each measured object. The two diffracted lights thus made travel in the fourth direction, and the two diffracted lights selected from the diffracted lights of the first and third incident beams are shifted in the fifth direction symmetrical to the fourth direction. Advancing, diffracting on the first measured object, the fourth direction Forming a beat light of two traveling diffracted lights and measuring a first phase; a beat light of the two diffracted lights traveling in the fifth direction, diffracted on the first measured object; And measuring a second phase; forming a beat light of two diffracted lights diffracted on the second object to be measured and traveling in the fourth direction; Measuring, diffracting on the second measured object, forming a beat light of two diffracted lights traveling in the fifth direction, and measuring a fourth phase; Measuring a relative displacement between the first measured object and the second measured object from the second, third, and fourth phases. 4. A displacement measuring apparatus for arranging first and second objects to be measured each having a plurality of diffraction gratings having different pitches arranged close to each other and measuring a relative displacement thereof. The first, second and third incident beams, each containing two wavelengths of light having slightly different frequencies, from the first, second and third directions, respectively, from the first and second beams; Incident on a plurality of diffraction gratings on a measurement object, and advancing two diffraction lights selected from the diffraction lights of the first and second incident beams on each measurement object in a fourth direction; An incident optical system for propagating two selected diffracted lights of the first and third incident beams in a fifth direction symmetrical to a fourth direction, and on the first measured object Diffracts to form a beat light of two diffracted lights traveling in the fourth direction. A first light receiving system for measuring a first phase, and a beat light of two diffracted lights diffracted on the first object to be measured and traveling in the fifth direction, to form a second phase. A second light receiving system for measuring, a third light for diffracting on the second object to be measured, forming beat light of two diffracted lights traveling in the fourth direction, and measuring a third phase A light receiving system, diffracted on the second measured object, forming a beat light of two diffracted lights traveling in the fifth direction, and measuring a fourth phase; From the first, second, third, and fourth phases supplied from the first, second, third, and fourth light receiving systems, the first and second phases are obtained.
And a relative displacement detection system for measuring a relative displacement of the object to be measured.