JP3029133B2 - Measurement method and device - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention is, for example, formed on a plurality of objects such as masks and reticles (hereinafter collectively referred to as "reticle") by an exposure apparatus for manufacturing semiconductor devices, or stored as drawing data. When a plurality of fine electronic circuit patterns are sequentially aligned and printed on the same wafer or the like having a photoconductor and baked, whether the patterns at the time of each printing are accurately superimposed on the photoconductor,
That is, the present invention relates to a method and an apparatus for measuring the accuracy of alignment by an exposure apparatus.

[0002]

2. Description of the Related Art In a so-called exposure / transfer apparatus for manufacturing a semiconductor, in which a circuit pattern of a reticle is exposed and transferred onto a photosensitive member of a wafer by using ultraviolet light, X-rays, etc., the relative alignment performance between the reticle and the wafer is improved. It is one of the important factors to achieve. In particular, in recent aligners in an exposure apparatus, an aligner having, for example, a submicron or less alignment accuracy is required for high integration of semiconductor elements.

In many alignment apparatuses, a so-called alignment pattern for alignment is provided on a reticle and a wafer surface, and both alignments are performed using position information obtained from the alignment patterns. In order to actually measure and evaluate the alignment performance of an exposure system assembled as an exposure system, a fine pattern formed on a conventional reticle is superimposed on the wafer and printed, and the amount of deviation from the pattern on the wafer is measured. The measurement is performed by visual observation or image processing.

[0004]

However, among such measurement methods, the measurement by visual observation depends on the experience and skill of the reader, and the measurement accuracy is not stable.

[0005] Since it is not automatic measurement, it takes time and effort.

[0006] High measurement accuracy cannot be obtained. And so on.

[0007] Also, in the measurement by image processing, there are problems that the method is complicated and time-consuming, high measurement accuracy cannot be obtained, and the like.

The present invention has been made in view of the above-mentioned drawbacks of the prior art, and provides a method and an apparatus for measuring overlay accuracy or positioning accuracy of an apparatus, which can be automated, can shorten the measurement time, and can obtain a stable and high measurement accuracy. The purpose is to do.

It is a further object of the present invention to provide an exposure apparatus capable of measuring a printing deviation between layers on the same exposure apparatus after printing on the semiconductor exposure apparatus.

[0010]

In order to achieve the above object, the present invention provides a method of irradiating a first grating pattern and a second grating pattern on the same surface with coherent light. The relative position shift between the first and second grating patterns is detected by causing the diffracted light emitted from the light source to interfere with each other.

[0011]

DESCRIPTION OF THE PREFERRED EMBODIMENTS The principle of the first embodiment of the present invention will be described below with reference to specific drawings.

FIG. 1 shows the first time when a laser beam 5 enters as a plane wave into an equally-spaced linear diffraction grating 4 (a linear grating pattern extends in a direction perpendicular to the paper) provided on a substrate 3. The figure shows a state in which the folded light is diffracted perpendicularly to the surface of the substrate 3.

At this time, the wavefront of the diffracted light 6 generated when the light 5 is incident on the diffraction grating 4 having the pitch P has a phase of 2π (that is, only one wavelength) when the diffraction grating 4 moves in the x direction by one pitch. It is well known that it changes. That is, when the light is moved by x ₀ in the x direction, a phase change of 2mπx ₀ / P is added to the diffracted light. Here, m is the order of the diffracted light.

As shown in FIG. 2, two adjacent equidistant linear diffraction gratings (gratings) on the same plane are considered. Here, the pitches of the two gratings are equal, and there is a shift in the x direction between adjacent linear patterns of each other. One of the gratings is A, and the other is B. The frequencies of these gratings are slightly different, and the initial phases are φ _0A and φ _0B respectively.
When lights 7 and 8 are irradiated on A and B, respectively, both complex amplitudes are expressed by the following equations.

[0015] _{U A = A 0 exp {i} (ω A t + φ 0A)} ... (1) U B = B 0 exp {i (ω B t + φ 0B)} ... (2)

[0016] At this time, the first-order diffracted light 9, 10 from each of the gratings _{U A '= Aexp {i (} ω A t + φ 0A + φ A)} ... (3) U B' = Bexp {i (ω B t + φ 0B + Φ _B )｝ (4) here,

[0017]

[Outside 1] x _A, x _B is the x-direction shift from each grating A, the same reference position of the B.

[0018] Therefore (3) and causing interference light represented by (4) (overlay) the intensity change of the interference light found by typing the _{| U A '+ U B'} | 2 = A 2 + B 2 + 2ABcos {2π
_{(F B -f A) t +} (φ B -φ A) + (φ 0B -φ 0A)} ...
(5) here,

[0019]

[Outside 2]

In the equation (5), A ² + B ² is a DC component, 2AB is an amplitude and a frequency difference between two light waves. f _B −
The signal having the beat frequency component of f _A is the initial phase shift φ
_The phase is temporally phase-modulated by a phase φ _B −φ _A representing the deviation between _0B− φ _0A and the grating.

[0021] Thus, by dividing the light by a half mirror or the like before being irradiated to the grating, U _A, when photoelectric conversion by superimposing the light of U _B, the intensity change signal of the interference light, | U _A + U _B | ² = A ² ₀ + B ² ₀ + 2A ₀ B ₀ cos ｛2π
(F _B −f _A ) t + (φ _0B −φ _0A )｝ (6) When the phase difference between the signals (5) and (6) is taken as a reference signal, the initial signal phase shift of light Can be eliminated, and highly accurate phase difference detection can be performed as so-called heterodyne interference measurement.

As is well known, the heterodyne method is
Since the phase shift between the two signals is detected as time, even if there is a difference in DC component or a change in amplitude between the signals, the measurement is not affected.

As shown in FIG. 3, the phase difference ΔT between the signal under test 11 and the reference signal 12 is detected with high precision using, for example, a lock-in amplifier or the like, so that the phase difference can be measured with high precision. For example, in terms of lock-in amplifier capability, λ /
1000 to λ / 2000 (0.4 ° to 0.2 ° phase)
Is actually possible.

Since the phase difference detected as described above matches the phase difference φ _B −φ _A indicating the amount of displacement between the gratings, if the pitch of the grating is P, P (φ _B −
The deviation amount between the gratings is obtained from φ _A ) / 2π.
For example, when using a 2 μm pitch grating,
In order to detect the displacement between gratings of 0.01 μm,

[0025]

[Outside 3] Therefore, it is necessary to detect a phase difference of λ / 200 (λ: wavelength of light used for heterodyne detection).

In this embodiment, the amount of displacement between gratings obtained based on the above-described principle is used as the amount of misalignment between the pattern printed on the first time and the pattern printed on the second time. This will be described below.

First, a pattern of a first reticle is printed on a photosensitive member of a wafer, here, a resist. On the first reticle, a first real element pattern, a first alignment mark of a known shape, and a first grating pattern, which is the above-described linear diffraction grating at equal intervals, are formed. Simultaneously print on the resist with an exposure device. After baking, the resist is developed and the like, and a new resist is applied thereon. The pattern of the second reticle is printed on the new resist. On the second reticle, a second real element pattern, a second alignment mark of a known shape, and a second grating pattern composed of equally spaced linear diffraction gratings having the same pitch as the first grating pattern are formed. Have been.
The printing is performed by a well-known semiconductor exposure apparatus using a first alignment mark transferred onto a wafer and a second alignment mark of a second reticle by a known method to detect a relative position between the second reticle and the wafer. Then, the second reticle is positioned after the wafer is aligned based on the result. The second lattice pattern is adjacent to the first lattice pattern on the wafer as shown in FIG. 2 when the second real element pattern is aligned with the first real element pattern on the wafer without displacement. The adjacent linear patterns of the gratings are arranged so as to be transferred at the matching positions without being shifted in the x direction in FIG. When the resist on which the second reticle is baked is developed, both the first and second lattice patterns are formed on the wafer. The shift amount in the x direction shown in FIG. 2 between the linear patterns adjacent to each other in the two lattice patterns is detected based on the principle described above. This shift amount is equal to the overlay shift amount of the first and second real element patterns, and indicates the error amount of the alignment by the first and second alignment marks in the semiconductor exposure apparatus.

The grid pattern to be formed may not be a pattern obtained by developing a resist. For example, the pattern is transferred to the resist, and the first and second grid patterns are latent images, ie, a pattern image seen on the resist without development. Either one or both may be formed. Also, one or both of the first reticle and the second reticle may not necessarily have a real element pattern. For example, when measuring only the alignment error amount of the semiconductor exposure apparatus, a test reticle having no actual element pattern can be used for the first reticle or the second reticle. For example, in this case, the first grating pattern and the first alignment pattern are not only transferred from the first reticle, but also used in a well-known EB.
(Electron beam) It may be formed by drawing with a drawing device. Similarly, the second grating pattern and the second alignment pattern need not be from the second reticle. Further, the material may be formed by directly processing the material with an ion beam or the like without using a photoreceptor. The alignment between the second reticle and the wafer is not only performed by the alignment mark, but also includes the case where the wafer on the positioning stage is moved by moving the stage by a predetermined amount with respect to the second reticle set at a predetermined position. .

Next, a first embodiment of the present invention will be described with reference to FIG. 4 together with a specific configuration of a grating diffraction light detection system. Those with the subscript 1 attached to the numbers in the figure are S waves,
Those with subscript 2 indicate P-wave light. The electric waves of the P wave and the S wave vibrate in a direction parallel to and perpendicular to the paper.

In FIG. 4, on a wafer 14 on which light 13 including light f _A1 and f _B2 of two frequencies (respectively f _A and f _B ) is to be measured,
The grating U.S. L, each frequency light is diffracted by the grating U, and the diffracted lights f _A1U , f _B2U, and the diffracted lights f _A1L , f _B2L diffracted by the grating L, respectively, are _relayed by the relay lenses 15, 16. After that, the optical paths of the light f _A1U and the light f _B2L are made to coincide with _each other by the Savart plate 17 (light flux 22), the optical path of the light f _A1L and the light f _B2U are moved away from this common optical path, and the polarization plane is inclined by 45 ° with respect to the paper. After passing through the polarizing plate 18, only the light beam 22 is condensed by the lens 19 and detected by the sensor 20. The sensor 20 beat frequency | may be selected in consideration of responsiveness | f _A -f _B. Therefore, for example, an APD (avalanche photo diode) having excellent responsiveness at a beat frequency of 1 MHz is preferable.

The seventeenth Savart plate will be described in more detail. As described above, the Savart plate is used to superimpose the light f _A1L and f _B2L diffracted by the wafer. As shown in FIG. 5, two parallel plane plates of the same type having a thickness d and a normal to the boundary surface forming an angle θ with the optical axis (θ = 45 ° for quartz, θ = 44.6 ° for calcite) A Savart plate is attached to a substrate so that a main cross section (a plane including an optical axis and a propagation direction) with respect to perpendicularly incident light is orthogonal. In the Savart plate, the traveling directions of the ordinary ray (hereinafter referred to as S-wave) and the extraordinary ray (hereinafter referred to as P-wave) are different. The S-wave at the first board becomes the P-wave at the second board, and the P-wave coming out of the first board becomes the P-wave. In two plates, it becomes an S wave. Therefore, each light is given one direction displacement by one of the first plate and the second plate. Assuming that the displacement amount is L ₀ ,

[0032]

[Outside 4] It is represented by the following equation. Here, a = 1 / _ne ( _ne is the refractive index of extraordinary light), b = 1 / n ₀ , c = a ² sin ² θ + b ²
cos θ. The distance between the S wave and the P wave on the emission surface is √2
・ L ₀ , and the two lights emitted from the Savart plate are √2 ・ L ₀
Are parallel to each other while maintaining the distance of. It is assumed that the Savart plate is arranged in the coordinate system shown in FIG. 5, and that light parallel to the x-axis is incident on the Savart plate surface (0, y ₀ , z ₀ ).
The positions of the S wave and the P wave on the emission surface are (2d,
y ₀ , z ₀ + L ₀ ) and (2d, y ₀ + L ₀ , z ₀ ). FIG. 6 shows a front view of the emission surface. When the light A is vertically incident on the point a (from the back side of the paper), the positional relationship between the S wave and the P wave is displaced by L _{0 in the} y-axis direction and the z-axis direction, respectively. Here, assuming that the light B is incident on a point b which is separated from the point a by √2 · L ₀ and is inclined downward by 45 ° from the y axis, the P wave of the light A and the S wave of the light B overlap on the exit surface. . The diffracted light from the wafer is superposed by utilizing such characteristics of the Savart plate. That is, FIG. 5 shows the state of the optical path when the light incident from the left exits to the right, but the light entering from the right also reverses, and in the configuration of the optical system of this embodiment, the light enters from the right. , The light coincides and overlaps on the exit side.

Returning to FIG. 4, how light travels near the Savart plate is shown.

Two light beams f _A1 and f _B2 are applied to the entire surface of the two gratings (U on the upper side and L on the lower side) on the wafer.
At this time, the diffracted light from the grating has four components, each of which is represented as f _A1U , f _B2U , f _A1L , f _B2L . (The subscript U means that it has the phase information of the upper grating U. Similarly, the subscript L means that it has the phase information of the lower grating L. At this point, f _A1U and f _B2U , f _A1L and f _B2L
Are on the same optical axis. These four components light enters the Savart plate through a relay lens, f _A1U and f _B2L are overlapped by the characteristics of the Savart plate as described above.

FIG. 7 shows the entire system including the illumination system of the measuring apparatus according to the first embodiment of the present invention. The same members as those in FIG. 4 are denoted by the same reference numerals.

In FIG. 7, 40 is a light source such as a laser, and 41 is a frequency shifter. If the light source 40 is a two-frequency generating light source such as an axial Zeeman laser, the frequency shifter 41 becomes unnecessary. In this case, a phase plate such as a λ / 4 plate is placed at the position of the frequency shifter 41 instead. Reference numeral 42 denotes an optical system for converting a beam diameter, and reference numeral 43 denotes a half mirror. The light flux from the light source 40 is a frequency shifter 41,
The wafer 14 is illuminated via the optical system 42 and the half mirror 43 as shown in FIG. 44 is a polarizing plate, 45 is a condensing optical system, 46 is an optical sensor, 47 is a lock-in amplifier,
55 and 56 are optical fibers (for example, single mode fibers).

In the figure, two gratings U and L
Is detected by the sensor 20 as described with reference to FIG. 4, and a signal represented by Expression (5) is obtained.

The light beam split by the half mirror 43 is P
Wave, the incident two-frequency light incident on the polarizing plate 44 having a polarizing plate inclined at 45 ° with respect to the vibration plane of the S wave interferes with each other,
A sensor 46 passes through a condensing optical system 45 and an optical fiber 55.
Incident on. The sensor 46 detects this interference light,
The signal represented by the equation (6) is obtained. Sensor 20, 4
6 are input to the lock-in amplifier 47, and as described above, a phase shift amount signal corresponding to the positional shift amount between the gratings is obtained. From this signal, a calculator (not shown) calculates the relative displacement of the grating,
That is, a pattern overlay error amount or a position alignment error amount is obtained.

If there is no phase shift between the grating patterns U and L, the wavefronts of f _B2L and f _A1U have no mutual phase shift due to diffraction of the grating, and the phase shift between the reference signal R and the signal S Does not occur. When there is a phase shift between the grating patterns U and L, the phase shift amount of the output of the lock-in amplifier 47 is obtained as a phase angle (actually, ΔT) as shown in FIG.
In FIG. 7, the Savart plate 17 is arranged so that the diffracted light shifts in the direction in which the gratings L and U are arranged (the straight line direction of the straight grating). FIG. 8 shows a specific example of the frequency shifter 41. In FIG.
60 is a polarization beam splitter, 61 and 62 are acousto-optic modulators, 63 and 64 are mirrors, and 65 is a polarization beam splitter. Here, for example, the acousto-optic modulator 61 is set to 80
If the acousto-optic modulator 62 is 81 MHz acousto-optic modulation, a 1 MHz frequency difference is given between the two lights.

As described above, according to the present embodiment, the evaluation of the amount of superimposition printed on the wafer by the application of the heterodyne interferometer technology can be performed with high precision and automatic measurement, and in the future, the miniaturization will be further increased. It is possible to provide a burn-in evaluation device that can be continuously integrated with higher integration of semiconductors.

Next, the principle after the second embodiment of the present invention will be described.

FIG. 9 shows a laser having an absolute value of the same incident angle and a slightly different frequency on an equally-spaced linear diffraction grating 104 (a linear grating pattern extends in a direction perpendicular to the paper) provided on a wafer 103. This indicates that the lights 105-1 and 105-2 are incident as plane waves. Here, the diffracted light on the left side in the traveling direction of the 0th-order diffracted light is represented by + m-order diffracted light,
The right side is the -m order diffracted light (m is an integer). + 1st-order diffracted light with respect to 105-1 grating pattern 104, 105-2
1 shows that -1st-order diffraction with respect to 104 is diffracted perpendicular to the surface of the wafer 103. At this time, the lights 105-1 and 105-2 are incident on the diffraction grating 104 having the pitch P, and the diffracted lights 106-1 and 106-2 generated therefrom.
It is well known that the phase of the wave front changes by 2π (that is, by one wavelength) when the grating 104 moves in one pitch x direction. Therefore, when the diffraction grating 104 moves by x _{0 in the} x direction, a phase change of ± 2mπx ₀ / P is added to the diffracted light. Here, m is the order of the diffracted light.
This is the same as that described in the first embodiment. The following embodiment detects a pattern printing misalignment detecting method and apparatus different from the first embodiment using this principle.

As shown in FIG. 10, two adjacent equidistant linear gratings (gratings) on the same plane are considered. Here, the pitches of the two gratings are both P '
And there is a shift (Δx = x _B ′ −x _A ′) between the gratings in the direction of the arrow (x direction). If the one of the grating A ', the other end B' and _{(x A ', x B'} x -direction deviation from the same reference position of each grating A is ', B').

Here, the frequencies slightly differ (ω _f1 , ω
_f2 ), and the initial phases are φ _0f1 and φ _0f2 , respectively. The complex amplitudes U _f1 and U _f2 of the two lights 109 and 110 are respectively expressed as U _f1 = A _{0 expｅｘi} (ω _f1 t + φ _0f1 )｝ (8) U _f2 = B ₀ exp ｛i (ω _f2 t + φ _0f2 ) ｝ ... (9) The two lights 109 and 110 are irradiated on the entire surface of the two gratings. For example, light 109 from the left
Light 110 is irradiated from the right at the same incident angle of absolute value. At this time, the + 1st-order diffracted light of the light 109 with respect to the gratings A ′ and B ′ is set to 111 and 112, and the light 110
-1st order diffracted light with respect to gratings A and B
When the complex amplitudes of the lights 3 and 114 are displayed,
The complex amplitudes of the lights 111, 113, _{112, and 114 are calculated as} U ' _AF1 (+1), U' _AF2 (-1), and U ' _Bf1 (+
1), U ' _Bf2 (-1)

[0045]

[Outside 5] Becomes

Here, φ _A ′ = 2π × _A ′ / P, φ _B ′ =
2πx _B ′ / P, and the gratings A ′,
The shift amount of B 'in the x direction is expressed as a phase amount. Then, diffracted lights 111 and 1 from grating A '
13, and diffracted light 112 from grating B '
And 114 cause interference light intensity changes U
_A, U _B is as follows.

U _A = | U ′ _Af1 (+1) + U ′ _Af2 (−1) | ² = A ² _f1
+ A ² _f2 + 2A _f1 A _f2 cos ｛2π (f ₂ −f ₁ ) t +
_{(Φ 0f2 -φ 0f1) -2φ A} '} ... (14) U B = | U' Bf1 (+1) + U 'Bf2 (-1) | 2 = B 2 f1
+ B ² _f2 + 2B _f1 B _f2 cos ｛2π (f ₂ −f ₁ ) t +
(Φ _0f2 −φ _0f1 ) -2φ _B ′｝ (15)

Here, f ₁ = ω _f1 / 2π, f ₂ = ω _f2 / 2
π, A ² _f1 + A ² _f2 , B ² _f1 + B ² _f2 are DC components, 2A _f1 A
_f2, 2B _f1 B _f2 are amplitudes. Equations (14) and (15) indicate that the signal having the beat frequency component of f ₂ −f ₁ is the initial phase shift φ _0f2 −φ _0f1 and the respective shift amounts φ _A ′, φ _{A of the} gratings A ′, B ′. _The phase is temporally modulated by _B '. Therefore, if one of the signals represented by the equations (14) and (15) is used as a reference signal and the other is used as a signal to be measured and two time shifts are detected, the initial phase of light can be eliminated, so-called heterodyne interference High-precision phase detection for measurement

[0049]

[Outside 6] Becomes possible. Heterodyne interferometry, as described above,
Since the phase shift between the two signals is detected as time, even if there is a difference in the DC component between the signals or a change in the amplitude, there is no effect on the measurement. Here, as shown in FIG.
5. Assuming that the time lag of the signal under measurement 116 is ΔT,
By detecting T with high accuracy using, for example, a lock-in amplifier or the like, high-precision measurement of the phase difference can be performed.

Since the phase difference detected as described above coincides with the phase difference Δφ indicating the amount of shift between the gratings, the amount of shift between the gratings is obtained from P ′ △ φ / 2π.

Therefore, based on the above-described principle, the amount of deviation between the first printed grid pattern and the second printed grid pattern is determined in the same manner as in the first embodiment. Similarly to the example, it is possible to detect the alignment accuracy of the semiconductor exposure apparatus and the shift amount between the actual element patterns formed by the first and second printings.

FIG. 12 shows an entire system of the measuring apparatus according to the second embodiment of the present invention. Light emitted from a light source 117 such as a laser is converted into two lights f _A1 and f _B2 whose polarization planes are orthogonal to each other and have slightly different frequencies by a frequency shifter 118 (here, the subscript 1 in the code is (P-polarized light, subscript 2 represents S-polarized light). Two lights f _A1 on a common optical path,
The beam diameter of f _B2 is reduced by the beam expander 119 and the traveling direction is changed by the mirror 120.
Light f _A1 , f _B2 incident on the polarizing beam splitter 121
Are divided into two directions according to the difference in the polarization plane. Where light f
_A1 is light 122 and light _fB2 is light 123. The lights separated in the two directions are reflected by mirrors 124 and 125, respectively, and the two lights are incident on the wafer 103 at incident angles having the same absolute value.
The entire surface of the upper gratings 107 and 108 is irradiated. Here, the gratings 107 and 108 in FIG.
FIG. 13 shows a view from the axial direction.

FIG. 14 is an enlarged view of an optical system at a portion where incident light is diffracted and interferes. The same angle lights 122 and 123 are incident on the normal line of the surface of the wafer 103 from the left and right directions,
The gratings 107 and 108 are irradiated. Where light 1
22 and 123 have slightly different frequencies, and the planes of polarization are orthogonal to each other. In FIG. 14, light 1 incident from the left
+ 1st order diffracted light 126-1 for 22 grating
And the -1st-order diffracted light 126-2 of the light 123 incident from the right with respect to the grating 107 passes through the Glan-Thompson prism 128 and is superimposed, and the wavefronts have the same polarization plane and interfere with each other. Similarly, the + 1st-order diffracted lights 127-1 and 123-1 of the light 122 with respect to the grating 108-
The first-order diffracted light 127-2 interferes. Each interference light S
₁ and S ₂ include a phase term indicating the amount of deviation from the initial phase of the grating 107 or 108, and when the interference signal is expressed by an equation, the interference signal is expressed by the equations (14) and (15), respectively.

That is, the expression (14) indicates that the light 126-1 and 126
-2, (15) is an expression for the signal light 127-1 and 127-2 interferes, 2φ _A ', 2φ _B' is a phase term representing a deviation amount. The interference lights S ₁ and S ₂ corresponding to the respective gratings 107 and 108 that have passed through the Glan-Thompson prism 128 are shifted according to the arrangement of the gratings, and are spatially separated by the edge mirror 129 into two directions. Are optical sensors 130, 1 such as photoelectric converters (eg, avalanche photodiodes);
The signal is converted into an electric signal at 31 and the lock-in amplifier 132
It is led to.

FIG. 15 is a view of the optical system shown in FIG. 14 as viewed from the x-axis direction. The incident light and the diffracted light overlap on a straight line.

Here, the pitch between the gratings 107 and 108 is 2 μm, and the wavelength λ of the light emitted from the light source 117.
= 0.6328 μm, ± 1st order diffracted light 126−
1, 126-2, 127-1, and 127-2 on the wafer 10
In order to diffract vertically upwards with respect to 3, light 122,
Letting the incident angle of the 123 with respect to the gratings 107 and 108 be θ ± ₁ , θ ± ₁ = sin (mλ / P) (16) (m is the order of the diffracted light) From the relational expression, θ ± ₁ = sin ⁻¹ ( 0.6328 / 2) =
18.4 °.

If the phase difference is detected at λ / 1000 in this measurement, the displacement amount of the grating pattern is equivalent to 0.0002 μm.

The frequency shifter 118 shown in FIG. 8 may be used, for example.

In the second embodiment, the gratings 107, 10
An example of detecting the amount of deviation using ± 1st-order diffracted light from No. 8 has been described. By moving the positions of the mirrors 124 and 125 in FIG. 12, higher-order diffracted light (± mth order; m = 2, 3, 4,
..) May be adjusted to diffract vertically upward with respect to the wafer 103, and the diffracted light may be used for measurement.

If high-order diffracted light is used for measurement, a phase amount representing a shift amount of the grating in the x direction can be obtained with high sensitivity. For example, if ± m-order diffracted light is used for measurement, measurement can be performed with m times higher sensitivity than that of ± 1st-order diffracted light.

Here, as in the second embodiment, the pitch P of the gratings 107 and 108 is 2 μm, and the light source 1
Assuming that the wavelength of No. 7 is λ = 0.6328 μm, ± 1st-order diffracted light can be diffracted vertically above the wafer 103 by (1
From equation (6), the incident angle θ ± _{2 is given} by θ ± ₂ = sin (2 × 0.6388 / 2) = 39.3 ° (17) Similarly, in the case of ± _third -order diffracted light, θ ± ₃ = sin (2 × 0 .6328 / 2) = 71.7 ° (18).

When the measurement is performed using ± m-order diffracted light, the phase terms φ ′ _Am and φ ′ _Bm corresponding to φ _A ′ and φ _B in equations (10) to (13) are φ ′ _Am = 2mπx _a become '/ P ... (19) φ ' Bm = 2mπx B '/ P ... (20). Therefore, if the amount of displacement between the gratings 107 and 108 is represented by the amount of phase,

[0063]

[Outside 7] Becomes

FIG. 16 is a diagram showing a third embodiment according to the present invention. Hereinafter, the same members as those described above are denoted by the same reference numerals.

In FIG. 16, in the second embodiment, light 122
From the left and light 123 from the right at the same incident angle (θ ₁ and θ ₂ in FIG. 15 are θ ₁ = θ ₂ ),
The entire surface of the sample was irradiated with 108, and diffracted lights having the same absolute value of the order (for example, + m order and −m order, m = 1, 2, 3,...) Interfered with each other and used for measurement. In this embodiment, the light 122 and the light 1
The angle of incidence on the gratings 107 and 108 can be adjusted by changing the positions of the mirrors 124 and 125, and diffracted lights having different absolute values of the order (for example, + m order and -n order, m
= 1, 2, 3,..., N = 1, 2, 3,... / M / ≒ / n /)
The movement amount can be detected by interfering with each other. Other configurations and methods are the same as in the second embodiment, and a description thereof will be omitted. FIG. 16 shows an example in which light beams 122 and 123 are incident on the gratings 107 and 108 at incident angles of θ ₁ and θ ₂ (/ θ ₁ / ≒ / θ ₂ /), respectively. Here, + m-order diffracted light of the light 122 with respect to the gratings 107 and 108 is 133-1 and 134-1, and -n-order diffracted light of the light 123 with respect to the gratings 107 and 108 is 133-2 and 134-2. At this time, equation (1)
0), (11) or φ in equations (12), (13)
_A, where the phase term corresponding to φ _B φ _Amn, When phi _Bnm

[0066]

[Outside 8] Can be obtained as

Next, a fourth embodiment of the present invention will be described. In the second embodiment, the angle formed by the optical paths of the incident lights 122 and 123 and the direction in which the grating lines of the gratings 107 and 108 are drawn was 90 ° (see FIGS. 14 and 15). The present embodiment is an example in which the incident light and the grating are incident at an angle other than 90 °. Other configurations and methods are the same as in the second embodiment. The incident lights 122 and 123 and the diffracted lights 126-1 and 126 are used.
17, 18, and 19 show views of the relationship between −2, 127-1, and 127-2 when viewed from each direction. Here, looking at the optical path on the projection plane to the cross-section xz plane orthogonal to the grating line direction including the normal line of the surface of the wafer 103 in FIG. 18, it includes the normal line and is parallel to the grating line. The incident lights 122 and 123 are incident from both sides of the surface (which overlaps the Z axis in this figure).

FIGS. 20 and 21 are views showing a grating section according to a fifth embodiment of the present invention. Between the gratings U and L or 107 and 108 used for the measurement in the first to fourth embodiments, if there is no overlay error between the actual element patterns that have been printed and superimposed, the misalignment detection direction (here In the embodiment, the pattern position shift (offset) in the x direction was set so as not to occur.
Known offset X between gratings 107 and 108
Is given. If the offset amount is subtracted from the measured value obtained as a result of the detection, the first to fourth
The deviation amount Δx can be similarly measured by any of the configurations and methods of the embodiment.

Here, FIG. 20 shows the case where an offset is set in the direction of detecting the displacement and an interval is provided in the direction perpendicular thereto, and FIG. 21 is an example where the offset is set only in the direction of detecting the displacement and the interval is provided in the direction perpendicular thereto. It was not opened.

Next, FIGS. 22 and 23 are views showing a grating section of a sixth embodiment according to the present invention. In this embodiment, at least one of the two gratings of the above embodiment is formed in a divided form. In FIG.
Reference numerals -1 and 107-2 depict what was originally a continuous grating, with the portion between them being drawn, and the grating 108 was inserted between the portions. One grating group is constituted by 107-1 and 107-2, and a displacement between the grating group and the grating 108 is detected. As a measuring method, the entire grating group is irradiated with light in the same manner as one of the gratings of the first to fourth embodiments, and light from the entire grating group is applied to one of the above-described embodiments.
Light from one or each grating is received in the same manner as light from one of the gratings. The obtained signal is handled in the same manner as the signal based on the diffracted light of one grating in the above embodiment. Other measurement methods and device configurations are the same as those in any of the above-described embodiments.

In FIG. 23, both of the two gratings of the first to fourth embodiments are formed separately.

In FIG. 23, 107-1 and 107-
Reference numerals 2, 108-1, and 108-2 denote a part of the same grating, respectively.
And the first graying group consisting of
This is to detect the displacement of the second grating group constituted by 08-2. Each of the grating groups is irradiated with light in the same manner as the corresponding one of the above-described embodiments, and similarly receives the diffracted light from the corresponding one of the above-described embodiments. The light from each grating group is received by one or a plurality of light receivers, and the signal from each grating group is processed in the same manner as the signal by the diffracted light of the corresponding one grating in any of the above-described embodiments. The displacement amount can be measured as in any of the embodiments.

FIG. 24 is a view showing a measuring apparatus according to a seventh embodiment of the present invention. This embodiment is different from the second embodiment in the optical system. The light emitted from the light source 117 passes through the frequency shifter 118, and becomes two lights whose frequencies are slightly different and whose polarization planes are orthogonal to each other. The light is converted by the beam expander 119 into a beam of a designated size. After that, the polarization beam splitter 140 splits the two light beams into two directions according to the difference in the polarization plane. Here, the angular frequency ω _{1 in} P-polarized light
Light 146, and S-polarized light having an angular frequency ω ₂ of light 147.
And The light 146 transmitted through the boundary surface of the polarizing beam splitter 140 has its direction changed by a mirror 141 and is irradiated by the objective lens 142 onto the gratings 107 and 108 on the wafer 103 while maintaining a predetermined incident angle. At this time, the form of incidence of each light on each grating is the same as in the second embodiment. However, at this time, the light source 11
The objective lens 142 is designed so that the beam waist portion of the light emitted from 7 comes to the gratings 107 and 108 on the wafer 103. Similarly, the light 147 reflected on the boundary surface of the polarizing beam splitter 140 is
2, gratings 107 and 10 on the wafer 103
8 is irradiated. The light 146 and the light 147 are irradiated to the gratings 107 and 108, and the −1st-order diffracted light from the grating 107 of the light 146 is referred to as a light 146-1 and the −1st-order diffracted light from the grating 108 is referred to as a light 146-2. The + 1st-order diffracted light from the grating 107 is referred to as light 147-1, and the + 1st-order diffracted light from the grating 108 is referred to as light 147-2. The traveling directions of these four lights are changed by a mirror 143, and the lights 146-1 and 147-1, and the lights 146-2 and 147-2 are caused to interfere by the polarizer 144. Then, light 146-1 and light 1
The light generated by the interference of the light beam 47-1 is converted into light 148 and light 146-2.
The light generated by the interference between the light 147-2 and the light 147-2 is referred to as light 149. These two lights pass through the relay lens 145, and the light 148 is transmitted to the photoelectric converter 130 and the light 149 is transmitted to the photoelectric converter 131 by the edge mirror 129. And split so as to be incident. At this time, the wavefront of light on the light receiving surface of each photoelectric converter is
The beam waist portion on the wafer 103 is imaged by the relay lens 145. And the photoelectric converter 13
0 and 131 are input to the lock-in amplifier 132, and the light 148 and the light 149 are input in the same manner as in the second embodiment.
Is obtained.

FIGS. 25 to 27 are views of a grating section for explaining an eighth embodiment of the present invention. FIG. 25 shows a grid pattern 148 on the first resist to be printed on the resist (first layer) on the wafer for the first time.
6 shows a grid pattern 149 on the second resist for developing the first layer on the wafer and then baking it on the resist (second layer) applied thereon for the second time.
7 shows a state on the wafer when both lattice patterns 148 and 149 are printed on the wafer. The grid pattern 148 is baked on the wafer as a first layer, and
And Pattern a consists pattern a ₁ and a pattern a ₂ Prefecture. Next, the lattice pattern 149 is baked on the wafer as a second layer, and this pattern is referred to as b. By combining these two grid patterns a and b, a grid pattern group as shown in FIG. 27 is formed on the wafer. Here, the upper half of the grid pattern group is the area 154, and the lower half is the area 15
5 is assumed. Further, the inside of the area 154 is divided into two, and the left side is the area 1
50 (pattern a part ₂ ), the right side is divided into two regions 152 (pattern b part), and similarly, the inside of the region 155 is also divided into two parts, the left side is a region 151 (pattern a _two part), and the right side is a region 153 (pattern a _one part). And Then, this lattice pattern group includes light having an angular frequency ω ₁ and an initial phase φ ₀₁ from a predetermined direction in a plane including a normal to the wafer surface and forming an angle of 90 ° with the direction in which the lattice lines are drawn. Light having an angular frequency ω ₂ and an initial phase φ ₀₂ is applied to the region 155. The complex amplitude of the first-order diffracted light obtained from each region is expressed by the following equation.

Region 150 u ₁ = A · exp (ω ₁ t + φ ₀₁ + φ _a ) (24) Region 151 u ₂ = B · exp (ω ₂ t + φ ₀₂ + φ _a ) (25) Region 152 u ₃ = C · exp (ω ₁ t + φ ₀₁ + φ _b ) (26) region 153 u ₄ = D · exp (ω ₂ t + φ ₀₂ + φ _a ) (27) where φ _a = 2πx _a / P, φ _b = 2πx _b / P
x _a and x _b are shift amounts in the x direction from the same reference position of the grid patterns a and b, respectively. Of the four diffracted lights, when the lights expressed by the equations (24) and (25) are superposed and interfered with each other, the following equation is obtained, and this signal becomes a reference signal in the main measurement.

| U ₁ + u ₂ | ² = A ² + B ² + 2AB cos ｛(ω ₂ −ω
₁ ) t + (φ ₀₂ −φ ₀₁ )｝ (28) Then, when the lights shown by the equations (26) and (27) are superposed and interfered, the following equation is obtained. And an object signal including a phase term indicating the relative displacement between the lattice pattern of the region 153 and the lattice pattern in the region 153.

| U ₃ + u ₄ | ² = C ² + D ² + 2CD cos ｛(ω ₂ −ω
₁ ) t + (φ ₀₂ −φ ₀₁ ) + (φ _a −φ _b )｝ (29) If the signals expressed by the equations (28) and (29) are input to the lock-in amplifier, the position between the two signals is obtained. the shift amount phi _a -.phi _b of retardation, that the grating can be measured as in the previous embodiments. The configuration of the device is such that in the device of the second embodiment shown in FIG. 12, two light beams emitted from the polarization beam splitter 121 are respectively incident on the regions 154 and 155 at the same angle, and the primary diffracted light emitted from the regions 150 and 151 is emitted. The optical system may be set such that are superimposed, interfere with each other and enter the detector 130, and the first-order diffracted lights emitted from the regions 152 and 153 are superimposed and interfere with each other and enter the detector 131.

In the above measuring method, the incident light is made incident on the wafer from one side. However, as shown in the second embodiment, the measurement can be performed by making the incident light incident on the wafer from both sides. Now, it is assumed that the light having the angular frequency ω ₁ and the initial phase φ ₀₁ is incident on the region 154 from the left, and the light having the angular frequency ω ₂ and the initial phase φ ₀₂ is incident on the region 155 from the right. At this time, the diffracted light obtained from each region is represented by the following equation.

+ _1st -order diffracted light from region 150 u ₁ '= A · exp (ω ₁ t + φ ₀₁ + φ _a ) (30) -1st-order diffracted light from region 151 u ₂ ' = B · exp (ω ₂ t + φ ₀₂ + φ _a) ... (31) +1 order diffracted light u ₃ from the region 152 '= C · exp (ω 1 t + φ 01 + φ b) ... (32) folding -1st from region 153 light _{u 4' = D · exp (} ω 2 t + φ ₀₂ −φ _a ) (33) Here, when the lights expressed by the equations (30) and (31) are superposed and interfered with each other, the following equation is obtained, and this becomes a reference signal.

| U ₁ ′ + u ₂ ′ | ² = A ² + B ² + 2AB cos ｛(ω ₂
−ω ₁ ) t + (φ ₀₂ −φ ₀₁ ) −2φ _a ｝ (34) Next, when the light represented by the equations (32) and (33) is superposed and interfered, the following equation is obtained. And an object signal.

| U ₃ ′ + u ₄ ′ | ² = C ² + D ² + 2CD cos ｛(ω ₂
−ω ₁ ) t + (φ ₀₂ −φ ₀₁ ) + (φ _a −φ _b )｝ ... (3
5) Then, if the signals of equations (34) and (35) are input to the lock-in amplifier, the phase difference between the two signals, that is, the amount of displacement of the grating − (φ _a −φ _b ) can be measured.

FIG. 28 shows a ninth embodiment of the present invention.
In FIG. 28, reference numeral 311 denotes a light source having high coherence such as a laser, 312 denotes a beam expander, 313 denotes a half mirror, and 314 denotes a moving mechanism for moving the half mirror 313 in the Z-axis direction. . 3
Reference numeral 15 denotes an object provided with a printing pattern such as a wafer, and L and U are patterns printed on the wafer. 316 is a lens 317 is a half mirror, 31
8 is an optical sensor such as a CCD. 319 is a drive circuit system of the moving mechanism 314, and 320 is a drive circuit system 31 which indicates a drive amount applied to the CCD 318 and the moving mechanism 314.
9 means an arithmetic processing system (for example, a computer or the like) for calculating the detection of a shift between the patterns U and L. The light emitted from the laser 311 is converted into a plane wave light beam of a predetermined size by the lens 312, and then the light is emitted from the half mirror 313.
, And is separated into a reference wave 321 and a wavefront 322 projected on the wafer 15. The projection wavefront 322 of the plane wave is
5, the first and second order diffracted lights are generated as plane waves because L and U are linear diffraction grating patterns, and after passing through a half mirror 317 via a lens 316, a reference wave 321 is formed. Interfere with each other.
The interfering wave is photoelectrically converted by a two-dimensional area sensor such as a CCD. The lens 316 makes the patterns L and U conjugate with the sensor 318.

The state of the interference fringes obtained on the optical sensor 318 will be described with reference to FIGS. FIG.
FIG. 30 shows a state in which an uneven pattern on the wafer is provided in a linear grid pattern. In FIG. 29 and FIG. 30, the hatched area and the other area schematically show areas having different heights of irregularities due to a resist or a semiconductor process material provided on the wafer. When the pitch of the grating is P, a plane wave is obliquely incident on the wafer 315 as shown in FIG. 28, and when the first-order diffracted light is diffracted in a direction perpendicular to the wafer surface, the wavelength of the light is λ, Letting the incident angle of θ be θ, the following relationship holds: Psin θ = λ (36)

FIGS. 31 to 34 each schematically show the state of interference fringes obtained on the optical sensor 318. FIGS. 31 and 32 show the reference light 321 and the diffracted light 323, respectively.
5 shows the intensity distribution on the sensor when the angles are superimposed so that the angle becomes zero. Complete plane wave (wavefront aberration = 0) 322
And if the lens system 316 has no aberration,
31 and 32, all pixels on the sensor 318 (CCD
In such a case, a so-called one-color image having a constant intensity is obtained. However, in FIG. 31, since the original patterns U and L have no phase shift, the images of U and L have the same brightness, whereas in FIG. 32, the original patterns U and L have the same brightness.
Has a phase shift δ, a signal is obtained in which the brightness of the whole image is slightly different from that of the U image.

FIGS. 33 and 34 show the reference light 321 and the diffracted light 3.
23 shows a state of interference fringes on the sensor 318 when the half mirror 313 is tilted and slightly interfered with the half mirror 313 at the time of superposition. In FIG. 33, since there is no phase shift between the original patterns U and L, there is no phase shift of interference fringes in both the U and L images, and a brightness distribution is obtained as if they were continuous images. In FIG. 34, since the original patterns U and L have a phase shift δ, as shown in the figure, a phase shift φ occurs in the interference fringe between the U image and the L image. As described above, if there is a phase shift δ between the original patterns U and L, the intensity of the image on the sensor changes even in the case of so-called one color, and even if the pitch of the stripe is generated by giving a tilt, the interference fringe may be generated. Since a shift occurs, if the phase shift of the interference fringes obtained as the intensity or intensity distribution on these sensors is determined with high accuracy, the phase shift amount δ between the original patterns U and L can be obtained with high accuracy. As one example of a method for obtaining the phase shift of interference fringes with high accuracy, there is a method called a so-called fringe scanning method. The fringe scanning method is a method in which the optical path length of the reference light is changed stepwise during interference, and the phase (shift) of the interference fringes is obtained by a Fourier series synthesis method using each interference pattern. In FIG. 28, when the phase of the reference light is changed by a moving mechanism 314 such as a piezo, the phase change amount is assumed to be changed N times over n times 2π (one wavelength). (If the amount of phase change is △ n)

[0086]

[Outside 9] The interference pattern intensity distribution at Δn is represented by I (x ′, y ′, △
n) (where x 'and y' are coordinate systems on the surface of the sensor 18 and indicate directional axes as shown in FIGS. 29 to 34). The intensity distribution of

[0087]

[Outside 10] Then, the phase φ (x ′, y ′) at the point (x ′, y ′) is

[0088]

[Outside 11] It is well known that it is required.

According to the above-mentioned so-called fringe scanning, by subtracting the wavefront aberration of the optical system as a unique one,

[0090]

[Outside 12] The following measurement accuracy of the wavefront shift (that is, the shift of the interference fringe) can be obtained. At this time, the phase shift δ of the original pattern
If the L / S of the pattern (linear grid) is 1 μm,

[0091]

[Outside 13] The following pattern phase shift is measured.

The inherent aberrations associated with the optical systems 313, 316, and 317 can be measured as the phase shift of the interference fringes obtained when the pattern U and L are set to 0 and the pattern is set to 0, and thereafter unknown. What is necessary is just to correct this inherent aberration with respect to the sample of the pattern phase shift and calculate the measurement value.

Incidentally, another simpler method is shown in FIG.
With respect to the interference fringe pattern image signal as shown in (1), a method of extracting the image area from the U-pattern diffracted light and the image area of the L-pattern diffracted light as the luminance signal of the scanning line and calculating the shift amount φ is also considered. Can be

In any case, according to this embodiment, a linear diffraction grating pattern is printed on a wafer, and coherent light is projected on the diffraction grating pattern. The system is configured so as to be superimposed to form interference fringes. This interference fringe is obtained as a phase shift by so-called fringe scanning or processing calculation of the luminance signal of the image, and the phase shift of the original printing pattern is obtained. This enables automatic measurement and high-precision measurement, and realizes a highly-accurate alignment deviation test of a semiconductor exposure apparatus and a stage performance test technique.

FIG. 35 shows a tenth embodiment of the present invention.

In FIG. 35, reference numeral 413 denotes a laser;
4 is a frequency shifter, 415 is a lens for changing the beam diameter, and 416 is a PBS (polarization beam splitter). The light emitted from the laser 413 generates a wave having two slightly different frequencies due to the frequency shifter 414.

A specific example of the frequency shifter 414 is shown in FIG.
6 is shown. In FIG. 36, reference numeral 430 is a polarization beam splitter, and 431 and 433 are acousto-optic modulators 432 and 43.
5, 436 are mirrors, and 434 is a polarization beam splitter. Here, for example, if 431 is an acousto-optic modulation of 80 MHz and 433 is an acousto-optic modulation of 81 MHz, a frequency difference of 1 MHz occurs between the two lights, and the polarization planes of S-polarized light and P-polarized light are orthogonal to each other as light of two wavelengths. You can get a wave.

Therefore, in FIG. 35, the light that has passed through the lens 415 is a wave whose polarization planes having slightly different frequencies are orthogonal to each other, and after passing through the polarization beam splitter 416, the transmitted light 420 is converted into P-polarized light (refraction, reflection). An electric vector vibrates in a direction parallel to the plane) and a light of frequency f ₂ (hereinafter referred to as f ₂ (p)), and reflected light 421 is an S-polarized light of frequency f ₁ (hereinafter f ₁ ). ₁ (s)).

The light of f ₂ (p) is projected onto the grating patterns L and U printed on the wafer 417, diffracted, and passed through the lens 418 to the polarizing beam splitter 41.
Go through 9.

After passing through the polarizing beam splitter 419, f ₁ (s) and the diffracted light by the grating are superimposed. However, since the polarization planes are orthogonal to each other, the light is converted into P and P using polarizing plates 404 and 425. An interference signal can be obtained by setting the polarization at 45 ° to the direction of the S-polarized light. 426 and 427 are optical sensors composed of an APD (avalanche photo diode) having a fast response, and 428 detects a phase difference between two signals obtained from the sensors 426 and 427, which fluctuates at a beat frequency with time. For example, a lock-in amplifier.
Reference numeral 422 denotes a pattern edge mirror obtained by vapor deposition of a metal film such as chromium Cr, which is located at a position where the patterns U and L on the wafer 417 are projected by the lens 418, and diffracted light from the pattern L is reflected by the pattern mirror 422. And the diffracted light from the pattern U
2 is set to be reflected (reflected by the Cr film 423). Thus, the signal of the diffracted light of the pattern L and the signal of the diffracted light of the pattern U are separated from each other,
7, 426.

With the above configuration, the lock-in amplifier 42
8, the phase shift is detected.

In FIG. 35, in order to detect and remove in advance the inherent wavefront aberration of the optical system, for example, the aberrations generated by the lenses 415 and 418 and the polarization beam splitters 416 and 419, the phase of the printing grating pattern on the wafer is required. The state of the shift 0 may be measured, and the phase output of the lock-in amplifier 428 obtained at this time may be processed as the inherent aberration of the optical system.

The lens 413 and the frequency shifter 414
With a plane of polarization orthogonal to and with a slight frequency difference
Instead of generating a wavelength, a so-called Zeeman laser (such as a transverse Zeeman laser in which a magnetic field is applied perpendicular to the optical axis or a transverse Zeeman laser in which a magnetic field is applied in a direction parallel to the optical axis)
May be used.

Next, FIG. 37 shows an eleventh embodiment of the present invention. In FIG. 37, 441 is a light source, 442 is a frequency shifter, 443 is a beam expander, 44
4 is a half mirror, 445 and 451 are mirrors, and 446 and 450 are polarization beam splitters. As shown in FIG. 37, the beam expander 443 and the half mirror 4
The light passing through 44 is composed of an S-polarized wave having a frequency of f ₁ and a P-polarized light having a frequency of f ₂ . The light transmitted through the half mirror 444 is the polarized beam splitter 4
The light is guided to a lens 447 by 46 and a mirror 451 and is incident on a grating pattern 449 on a wafer 448. Here, f ₁ (S) means S-polarized light at the f ₁ frequency, and f ₂ (P) means P-polarized light at the f ₂ frequency. When the pitch of the grating 449 is P, Psin θ = λ (40) When the light is incident in a relationship satisfying (λ is a wavelength), the diffracted light by the pattern 449 travels perpendicularly to the wafer 448.
f ₁ diffracted light which light generated upon diffraction of _{(S) f 1 (S)} ,
The diffracted light which light is generated upon diffraction of _{f 2 (P) f 2 '} (P)
Is expressed as These lights are polarized beam splitter 45
After passing through 0, f ₁ ′ (S) is sent to the sensor 454 and f ₂ ′
(P) reaches the sensor 457. Meanwhile, half mirror 4
Lights f ₁ (S) and f ₂ (P) reflected from 44 are mirrors 445.
After passing through the polarizing beam splitter 450 through
f ₂ (P) is for sensor 454 and f ₁ (S) is for sensor 4
It reaches 57. Since the light reaching the sensor 454 is f ₁ ′ (S) and f ₂ (P) and has polarization planes orthogonal to each other as they are, the polarizing plate 452 set at an angle of 45 ° with each polarization direction is attached to the sensor. Put just before.

By the way, when a beam is projected through the lens 447 so as to irradiate the entire surface of the pattern 449 (this is a pattern like A and B in FIG. 2) on the wafer 448, the diffracted light is It is diffracted from both A and B. Here, since only the diffracted light from the pattern A enters the optical sensor 454 and only the diffracted light from the pattern B reaches the sensor 457, a lens is used so that diffracted light from other gratings does not enter the sensor. With respect to 447, 453, and 456, an aperture is placed at a conjugate position with the pattern 449 to cut off extra diffracted light. FIG. 38 is an enlarged view of the vicinity of the sensor 454. 460 is an aperture. The sensor 457
Aperture 461 is also placed just before.

The light reaching the sensor 457 in FIG. 37 is f ₂ ′ (P) and f ₁ (S).
5 is placed immediately before the sensor 457. by this,
A signal is supplied to the lock-in amplifier, and phase shift detection of the pattern 449 (actually set as A and B shown in FIG. 2) on the wafer 448 can be realized with high accuracy.

Next, FIG. 39 is a view of a twelfth embodiment of the present invention, and relates to an exposure apparatus capable of measuring a printing shift of a pattern printed on a wafer. Reference numeral 201 denotes a displacement detection / measurement unit; 202, a control unit which obtains and controls a displacement detection system and a length measurement output of a length measuring device 207 for measuring the amount of movement of the stage 205; 210, an ic circuit pattern area; 211, a so-called scribe line The shift detection pattern in the x-axis direction printed on the upper side is a pattern as shown in FIG. The displacement detection measurement unit has, for example, the same configuration as any of the above-described embodiments. 2
Reference numeral 08 denotes a laser light source for length measurement, reference numeral 206 denotes a unit of an interference system, and reference numeral 214 denotes a mirror for measuring the amount of movement of a stage mounted on the stage 205. FIG. 39 shows the stage 205
Is shown only for one axis, but there is actually another orthogonal axis (not shown). Deviation detection system measurement unit 20
Reference numeral 1 denotes a spatially different position from an exposure imaging lens (not shown), and moves the stage 205 to move the wafer under the displacement detection system measurement unit 201 to perform measurement.

Thus, there is an advantage that the exposure apparatus can print and measure the amount of deviation due to printing.

In FIG. 39, reference numeral 203 denotes a wafer;
Is a wafer chuck, 209 is a vacuum tube,
212 is a guide for movement.

FIG. 40 shows the relative arrangement of the ic pattern area 210 printed on the wafer 203 and the printing shift detecting pattern 211 in the x-axis direction. An example is shown in which a burn-in misregistration detection pattern 211 is placed in a scribe line 214. Note that the burn-in deviation detection pattern 211 may be burned in the ic pattern area 210 instead of in the scribe line 214.

FIG. 41 shows an example of a printing misalignment detection pattern. 211-1 is the Nth layer, 211-2 is Nth layer
The printing pattern corresponding to the (+1) th layer shows a printing shift of Δx. FIG. 42 shows another example of the burn-in deviation detecting pattern.

FIG. 42 shows a pattern capable of detecting the displacement of both the x-axis and the y-axis. Detection of displacement in the y-axis direction is performed using the patterns 215-1 and 215-2, and 216-1 and 216-2.
The detection of the displacement in the x-axis direction is performed by the following pattern.

Further, the present invention is applicable to a reduction projection exposure apparatus called a so-called optical stepper as shown in FIG. FIG. 43 shows a thirteenth embodiment of the present invention. Light source 511
(For example, a so-called Zeeman laser with two-frequency oscillation) emitted from the first lens system 515 is incident on the first lens system 515. An image obtained by using the pattern of the grating 510 as a secondary light source is once collected by a first Fourier transform lens 515, and further the image of the first grating 510 is reduced and projected onto a wafer 518 through a second Fourier transform lens 517. Projection through 519.
The first focal point on the rear focal plane of the first Fourier transform lens 515 is
The diffraction light of the grating 510 is spatially distributed. In the present embodiment, a spatial filter 516 (see FIG. 45) is arranged on the Fourier transform plane and is filtered on the spectral plane, and the pattern of the grating 510 is filtered on the spectral plane, so that the interference fringes 520 are formed on the wafer 518 plane.
Generate Reference numerals 566 and 567 denote polarizing plates, which pass P-polarized light (electric scene parallel to the paper surface) and S-polarized light (electric scene perpendicular to the paper surface), respectively. The diffracted light 522 is diffracted from the second grating 521 formed on the semiconductor wafer 518 (ie, the first and second layer gratings 591 and 592 baked on the wafer as shown in FIG. 50), and the reduced projection optics is used. The system 519 and the second lens 517 are returned in the opposite direction and guided to the photodetector by a small mirror arranged at the position of the spatial filter 516. Above is wafer 5
18 is an optical system for detecting a misalignment in printing of a semiconductor process pattern, which is printed on 18. In FIG. 43, 512 is a projection light source, 513 is an illumination optical system, and 514 is a reticle.

FIG. 44 is a sectional view showing the state of diffraction when light enters the first grating 510 of FIG. The grating pattern 541 diffracts a plurality of diffracted lights such as 0 order, ± 1 order, ± 2 order .... The light blocking portion 543 surrounding the lattice pattern 541 is formed of a film such as chromium or chromium oxide, and allows the incident light 544 to pass only inside the pattern 541. In the example of FIG. 44, the amplitude grating pattern 541 is used to obtain the diffracted light. However, this grating may be a phase grating, or may be an echelette grating when the incident light is obliquely incident.

Here, the principle of the present embodiment will be described with reference to FIG. The light of wavelength λ emitted from the light source 511 is
The zero phase grating pattern 541 is illuminated. Note the phase grating pattern 541 is disposed at a position before and after the focus f ₁ of the first Fourier transform lens 515. The pitch P ₁ of the grating pattern 541 and the diffraction angle θ ₁ of the diffracted light are P ₁ sin θ _n = n · λ
(N = 0, ± 1, ± 2...) (41). The diffracted light enters the Fourier transform lens 515, and forms a Fourier spectrum image corresponding to each diffracted light on the rear focal plane as shown in FIG. Light corresponding to the Fourier spectrum of the first-order diffracted light forms an image on the Fourier transform surface in a state where it is completely separated from the Fourier spectrum of the zero-order diffracted light. A spatial filter 516 is arranged on the Fourier transform plane to block 0th order and ± 2nd or higher order diffracted light of the grating pattern 541, and pass ± 1st order diffracted light. Light source 511
In this case, light having a slightly different frequency is oscillated like a Zeeman laser. The diffracted light passing through the Fourier transform surface passes through the second Fourier transform lens 517, and is further projected on the wafer 518. Note that reference numerals 566 and 67 denote polarizing plates, and the polarization state of the light after passing through them is adjusted to P-polarized light and S-polarized light.

The ± primary light components interfere with each other on the wafer 518 to form interference fringes of a new pitch. The pitch P ₂ of this interference fringe is

[0117]

[Outside 14] Given by At this time, the second Fourier transform lens 517
Since the first Fourier transform lens 515 Fourier transform plane is set on the front focal plane of the above equation, there is a relationship of f ₁ sin θ ₁ = f ₂ sin θ ₂ (43). First and second Fourier transform lenses 51
5, 517 and the reduction projection optical system 519 shown in FIG.
Between images through a reduction projection system with a reduction ratio m of

[0118]

[Outside 15] There is a relationship. Therefore, when f ₁ = f ₂ , the pitch P _{3 of the} interference fringes formed on the wafer W has a value that is half the pitch when the grid pattern 541 is projected. A second grating G is formed on the wafer W518 by the projected image of the grating 541, and when this grating G is irradiated with light beams 611 and 612, respectively, as shown in FIG. Light is obtained. When the two light beams 611 and 612 are simultaneously irradiated on the wafer W518, interference fringes are generated.
The light diffracted by the grating G on the light 8 interferes with each other, and the interfered light is detected by a photodetector, and light intensity information indicating the positional relationship between the interference fringe and the grating G is obtained.

FIG. 48 shows the relation between the lens 517, the small mirror M, the lens 589, the relation between the wafer W and the edge mirror 584 (see FIG. 51), and the relation between the path of the light diffracted by the grating G on the wafer W. . The light diffracted by the wafer W passes through the lens 517, is condensed at the position of the small mirror M, and reaches the edge mirror 584 through the lens 589. At this time, when the focal length of the lens 589 and _{f 3,}
48, the position of the edge mirror 584 and the position of the wafer W are in a conjugate positional relationship with respect to the lenses 517 and 589. The pattern of the grid printed on the wafer W is the pattern shown in FIG. 50. On the edge mirror 584, the relationship between the position of this grid pattern and the edge is as shown in FIG. That is, FIG.
The light reflected by the edge mirror 584 is 59 as shown in FIG.
2 ′ is a light of a conjugate image of a 2nd layer pattern as shown by 2 ′, while transmitted light is 1 ′ as shown by 591 ′.
This is light of a conjugate image of the pattern of the slayer. If this is detected by a sensor 583 having a high light response such as an APD or a Si sensor, and the phase is detected by a lock-in amplifier 587, a high-precision phase detection by a so-called optical heterodyne can be performed, thereby printing on the wafer W. It is possible to check the phase shift of the grating, that is, the printing shift. 580, 5
The polarizing plate 81 is placed to cause interference between orthogonally polarized wavefronts, and the light source 511 uses a two-frequency oscillation laser such as a so-called Zeeman laser, or an AO (Aucosto-Optical) to a normal one-wavelength laser. Acousto-optic) A beat frequency oscillation light using a modulator is set. FIG. 52 shows a pattern of a temporal phase shift between the signal from the optical sensor 582 and the signal from the optical sensor 583.

When the optical heterodyne method is not adopted and the fringe scanning method described with reference to FIG. 28 is used, the light source 511 is a normal one-wavelength oscillation laser, and FIG. 53 shows a fourteenth embodiment of the present invention. As described above, the polarizing plates 580 and 581 in front of the optical sensors 591 and 592 become unnecessary.

As described above, according to the embodiments shown in FIG. 43 and thereafter, it is possible to provide an exposure apparatus capable of detecting a printing deviation on a wafer substrate with high accuracy.

In this embodiment, the diffracted light is projected on the wafer by the grating 541 to generate the diffracted light. However, the light is projected on the wafer by combining the beam splitter and the mirror without using the grating 541. You can do it.

[0123]

As described above, according to the present invention, for example, after printing on a semiconductor exposure apparatus, a deviation amount due to printing can be measured.

That is, after forming a first grid pattern on a scribe line, for example, together with the first circuit pattern, the same wafer is aligned, and a second grid pattern is formed on the scribe line, together with the second circuit pattern. Is detected by detecting the interference light of the diffracted light emitted from the two patterns, the relative displacement between the two patterns can be measured with high accuracy. .

[Brief description of the drawings]

FIG. 1 is an explanatory diagram of light diffracted by a diffraction grating.

FIG. 2 is an explanatory diagram of diffracted light by two adjacent diffraction gratings on the same plane.

FIG. 3 is an explanatory diagram of a phase difference between a signal under measurement and a reference signal.

FIG. 4 is a diagram of a first embodiment of the present invention.

FIG. 5 is a diagram of a first embodiment of the present invention.

FIG. 6 is a diagram of a first embodiment of the present invention.

FIG. 7 is a diagram of a first embodiment of the present invention.

FIG. 8 is a diagram of a first embodiment of the present invention.

FIG. 9 is a diagram of a second embodiment of the present invention.

FIG. 10 is a diagram of a second embodiment of the present invention.

FIG. 11 is a diagram of a second embodiment of the present invention.

FIG. 12 is a diagram of a second embodiment of the present invention.

FIG. 13 is a diagram of a second embodiment of the present invention.

FIG. 14 is a diagram of a second embodiment of the present invention.

FIG. 15 is a diagram of a second embodiment of the present invention.

FIG. 16 is a diagram of a third embodiment of the present invention.

FIG. 17 is a diagram of a fourth embodiment of the present invention.

FIG. 18 is a diagram of a fourth embodiment of the present invention.

FIG. 19 is a diagram of a fourth embodiment of the present invention.

FIG. 20 is a diagram of a fifth embodiment of the present invention.

FIG. 21 is a diagram of a fifth embodiment of the present invention.

FIG. 22 is a diagram of a sixth embodiment of the present invention.

FIG. 23 is a diagram of a sixth embodiment of the present invention.

FIG. 24 is a diagram of a seventh embodiment of the present invention.

FIG. 25 is a diagram of an eighth embodiment of the present invention.

FIG. 26 is a diagram of an eighth embodiment of the present invention.

FIG. 27 is a diagram of an eighth embodiment of the present invention.

FIG. 28 is a view of a ninth embodiment of the present invention.

FIG. 29 is a view of a ninth embodiment of the present invention.

FIG. 30 is a diagram of a ninth embodiment of the present invention.

FIG. 31 is a view of a ninth embodiment of the present invention.

FIG. 32 is a view of a ninth embodiment of the present invention.

FIG. 33 is a view of a ninth embodiment of the present invention.

FIG. 34 is a view of a ninth embodiment of the present invention.

FIG. 35 is a diagram of a tenth embodiment of the present invention.

FIG. 36 is a diagram of a tenth embodiment of the present invention.

FIG. 37 is a diagram of an eleventh embodiment of the present invention.

FIG. 38 is a diagram of an eleventh embodiment of the present invention.

FIG. 39 is a view of a twelfth embodiment of the present invention.

FIG. 40 is a diagram of a twelfth embodiment of the present invention.

FIG. 41 is a diagram of a twelfth embodiment of the present invention.

FIG. 42 is a diagram of a twelfth embodiment of the present invention.

FIG. 43 is a diagram of a thirteenth embodiment of the present invention.

FIG. 44 is a view of a thirteenth embodiment of the present invention.

FIG. 45 is a diagram of a thirteenth embodiment of the present invention.

FIG. 46 is a diagram of a thirteenth embodiment of the present invention.

FIG. 47 is a view of a thirteenth embodiment of the present invention.

FIG. 48 is a diagram of a thirteenth embodiment of the present invention.

FIG. 49 is a diagram of a thirteenth embodiment of the present invention.

FIG. 50 is a diagram of a thirteenth embodiment of the present invention.

FIG. 51 is a diagram of a thirteenth embodiment of the present invention.

FIG. 52 is a diagram of a thirteenth embodiment of the present invention.

FIG. 53 is a diagram of a fourteenth embodiment of the present invention.

[Explanation of symbols]

 Reference Signs List 40 light source 41 frequency shifter 46 optical sensor 47 lock-in amplifier U lattice pattern L lattice pattern

──────────────────────────────────────────────────続 き Continuation of front page (72) Minoru Yoshii 3-30-2 Shimomaruko, Ota-ku, Tokyo Within Canon Inc. (56) References JP-A-2-90006 (JP, A) JP-A-2 -227604 (JP, A) (58) Fields investigated (Int. Cl. ⁷ , DB name) G01B 11/00-11/30 G01N 21/88 H01L 21/64-21/66

Claims (15)

(57) [Claims] 1. A first step of forming at least a first grid pattern on an object, a second step of forming at least a second grid pattern on the object, and a first grid pattern and a second grid on the object. Irradiating the pattern with coherent light and causing the diffracted light emitted from the first and second grating patterns to interfere with each other to detect a relative positional shift between the first and second grating patterns on the object. A measuring method characterized by having: 2. The measuring method according to claim 1, wherein said first and second grating patterns are equally spaced grating patterns having the same pitch. 3. The measuring method according to claim 1, wherein the relative positional deviation is detected by detecting a phase difference by a heterodyne method. 4. In the third step, coherent light having a first frequency is incident on the first grating pattern, and coherent light having a second frequency is incident on the second grating pattern, and diffracted lights emitted from the two interfere with each other. Detecting the first interference light, and detecting the phase difference between the first interference light and the second interference light that interfered with each other without causing the coherent light of the first and second frequencies to enter the grating pattern. 4. The measuring method according to claim 3, wherein the relative position deviation is detected by performing the detection. 5. In the third step, coherent light having a first and / or second frequency is made incident on each of the first and second grating patterns, and the coherent light is diffracted from the first grating pattern and emitted. After the first and second coherent lights interfered with each other, the first and second coherent lights interfered with each other after the first coherent light was diffracted and emitted from the second grating pattern. Detect the second interference light that caused the interference,
4. The method according to claim 3, wherein the relative position shift is detected by detecting a phase difference between the first interference light and the second interference light. 6. A first grating pattern is formed on the object simultaneously with the first or second grating pattern in the first or second step, and a first frequency is applied to the first grating pattern in the third step. The first coherent light, the first coherent light that interfered with the diffracted light emitted from each of the coherent light of the second frequency incident on the second grating pattern, and the first and second in the third grating pattern Coherent light of two frequencies is incident and the first and second diffracted lights emitted from the third grating pattern are detected as second interference lights that interfere with each other, and the first interference light and the second interference light are detected. 4. The measuring method according to claim 3, wherein the relative position deviation is detected by detecting a phase difference of the measurement result. 7. The method according to claim 1, wherein the object has a photoreceptor, and the first and second lattice patterns on the object are formed by a latent image of the photoreceptor. 8. The object has a photoreceptor, and the first step is performed by printing a pattern of a first reticle having a first lattice pattern on the photoreceptor, and the second step is a second lattice pattern. 2. The method according to claim 1, wherein the measurement is performed by printing a second reticle having the following formula on the photoconductor. 9. The measuring method according to claim 1, wherein in the second step, an object on which the first grid pattern is formed is aligned before forming the second grid pattern. 10. A means for irradiating a first grating pattern and a second grating pattern on the same surface with coherent light, and causing the diffracted light emitted from the first and / or second grating patterns to interfere with each other. A measuring apparatus comprising: means for detecting a relative displacement between the first and second grid patterns. 11. The measuring apparatus according to claim 10, wherein the relative position shift is detected by a phase difference detection by a heterodyne method. 12. A first interference light in which a coherent light beam of a first frequency is incident on the first grating pattern and a coherent light beam of a second frequency is incident on the second grating pattern, and diffracted lights emitted from the first interference light interfere with each other. By detecting the phase difference between the first interference light and the second interference light that interfered with each other without causing the coherent light of the first and second frequencies to enter the grating pattern. 11. The measuring device according to claim 10, wherein a shift is detected. 13. Coherent light beams of first and second frequencies are incident on the first and second grating patterns, respectively, and diffracted light beams of first and second frequencies emitted from the first grating pattern interfere with each other. The relative position shift detection is performed by detecting a phase difference between a first interference light and a second interference light that causes the diffracted lights of the first and second frequencies emitted from the second grating pattern to interfere with each other. 11. The measuring device according to claim 10, wherein 14. A first interference light in which coherent light of a first frequency is incident on the first grating pattern and coherent light of a second frequency is incident on the second grating pattern, and diffracted lights emitted from the first interference light interfere with each other. The first and second coherent lights of the first and second frequencies are made incident on the third lattice pattern formed on the same plane simultaneously with the first or second lattice pattern, and the third light is emitted from the third lattice pattern. Detecting the second interference light that causes the diffracted lights of the first and second frequencies to interfere with each other, and performing the relative displacement detection by detecting a phase difference between the first interference light and the second interference light. Claim 1
0 measuring device. 15. A first step of forming at least a first grid pattern on an object, a second step of forming at least a second grid pattern on the object, and at least a first grid pattern on the object. And the third step of irradiating the second grating pattern with at least a second light, the first light, the second light, the third light emitted from the same light source as the first light, the third light Of the fourth light emitted from the same light source as the second light,
A fourth step of interfering two light beams to receive light, a fifth step of interfering the remaining two light beams to receive light, and a sixth step of comparing two light reception outputs to detect a relative displacement. A measuring method characterized by the above-mentioned.