JP2691298B2 - Positioning device and exposure apparatus including the same - Google Patents

Description

TECHNICAL FIELD The present invention relates to a diffraction grating interference type alignment apparatus, and is particularly suitable for an alignment apparatus such as a semiconductor manufacturing apparatus (stepper).

[Conventional technology]

In recent years, as an apparatus for transferring a fine pattern such as a semiconductor element onto a semiconductor wafer with high resolution, a projection type exposure apparatus,
So-called steppers have come to be used frequently. Particularly in recent years, there has been a demand for higher density of LSIs manufactured by this, and it is necessary to transfer a finer pattern onto a wafer. In order to cope with this, more accurate alignment is essential.

Therefore, many diffraction grating interference type alignment devices have been proposed.

For example, as shown in FIG. 13, there is known a positioning device using the heterodyne interferometry.

As shown, the reticle 1 has a predetermined circuit pattern.
Are held on a reticle stage 2 which is two-dimensionally movable. The pattern on the reticle is uniformly illuminated by the exposure light supplied from the illumination optical system 30, and is imaged on the wafer 4 (substrate) via the projection lens 3. A wafer mark WM of a diffraction grating for alignment (positioning) is formed on the wafer 4.

Now, when the wafer 4 is attracted to the stage that moves two-dimensionally by the step-and-repeat method and the transfer of the pattern on the reticle in one shot area on the WM is completed,
It is stepped to the next shot position.

X on reticle stage 2 and wafer stage 5
An interferometer (not shown) is provided for each stage for independently detecting the position in the direction, the Y direction, and the rotation (θ) direction, and each stage is driven in each direction by a drive motor (not shown). .

 Next, the alignment optical system will be described.

The laser beam output from the laser light source 10 is expanded into a predetermined beam diameter by the beam expander 11 and divided into two by the beam splitter 12. The light beam reflected by the beam splitter 12 is reflected by the acousto-optic element 14a (A
On the other hand, the light beam which is incident on the OM) and which is transmitted through the beam splitter 12 is incident on the acousto-optic element 14b via the reflection mirror 13. When each light beam is frequency-modulated by each acousto-optic element, it becomes a light beam having a different frequency and is relayed by the lenses 15a, 15b, 16 and the field stop FS of the diaphragm means has a predetermined incident angle (± θ _FS ) With two directions.

The field stop FS has a rectangular opening, and the irradiation of alignment light forms an interference fringe that runs through this field stop. At this time, when the wavelength of the alignment light is λ, the pitch P _{FS of the} interference fringes is And the frequency of the light beam passing through the acousto-optic element 14a is f ₁ ,
Assuming that the light beam passing through the acousto-optic element 14b is frequency-modulated with a frequency of f ₂ , this interference fringe is P _FS × | f ₁ −f ₂ |
Move at the speed of.

Each of the alignment lights that have passed through the field stop FS passes through the reflection mirror 17, the objective lens 18, the beam splitter 19, the dichroic mirror 20, and the projection lens 3 to form a diffraction grating mark WM (hereinafter referred to as “mark”) formed on the wafer (substrate). , Wafer mark) are irradiated with two coherent light beams having different frequencies at an incident angle (± θ _WM ) which is symmetrical with respect to the normal (optical axis) direction of the two coherent light beams. Then
Similar to the field stop FS described above, interference fringes running on the wafer mark WM are formed. At this time, the pitch of this interference fringe P _IWM
Is Therefore, this interference fringe moves at a speed of P _IWM × | f ₁ −f ₂ |.

Here, in the normal (optical axis) direction of this wafer mark WM, ±
The n-th order diffracted light is set to be generated, and this incident angle θ _WM is defined as follows: when the pitch of the diffraction grating mark WM is P _WM and the wavelength of the alignment light is λ, Given by

Then, for the detection ± that occurs in the direction of this normal (optical axis)
Projecting lens 3 and dichroic mirror 2 for 1st-order diffracted light
Photoelectric detection is performed by a detector 23 provided at a position conjugate with the pupil of the projection lens 3 through 0, the beam splitter 19, the objective lens 21, and the spatial filter 22.

The spatial filter 22 is provided at a position conjugate with the pupil of the projection lens and has a function of passing the diffracted light from the wafer mark WM traveling in the optical axis direction.

The phase detection system 41 is an optical beat signal obtained by the detector 23.
Drive signal difference between S _WM and acousto-optic elements 14a, 14b | DS ₁ −
The phase difference from DS ₂ | is detected, and based on this phase difference, the control system 42 controls the wafer stage 5 so that the phase difference becomes zero to align the wafer 4, that is, perform alignment. .

[Problems to be solved by the invention]

As described above, when the diffraction grating interference type alignment is performed, for example, as shown in FIG. 13 described above, the field stop FS is set at a position conjugate with the wafer 4 in the optical system for transmitting the alignment light. Placement needs to be placed. This limits the illumination area on the wafer mark of the diffraction grating formed on the wafer, and in particular, a part of the alignment light is reflected by illuminating the pattern of the transfer area or other alignment marks. It functions to prevent light from entering the detector as noise light.

However, there is a problem that the diffracted light generated by the field stop FS adversely affects the detection accuracy. Therefore, in order to facilitate the following description, the alignment optical system will be briefly described with reference to FIG.

As shown by the solid line in FIG. 14, coherent alignment lights (LB ₁ and LB ₂ ) of different frequencies irradiate the rectangular diaphragm FS in two directions with a predetermined incident angle θ _FS, and then the projection lens 3 Two beam spots (BS ₁ and BS ₂ ) are formed on the pupil plane P (Fourier plane) of the projection lens 3 through the lens 3a which constitutes one part, and further, the one of the photographing lens 3
Through the lens 3b forming the part, the illumination area of the diffraction grating mark WM formed on the wafer is moved to a predetermined incident angle (±
Irradiate in two directions with θ _WM ).

Now, it is assumed that the incident angle (± θ _WM ) is set so that ± 1st order diffracted light is generated in the direction of the normal line (optical axis) of the wafer marm WM.

Then, this wafer mark WM detects only LB ₁ (+1) and LB ₂ (−1) of ± first-order diffracted light generated in the normal (optical axis) direction, and performs alignment.

However, in practice, when the field stop FS having a rectangular opening is irradiated with alignment light (LB ₁ , LB ₂ ) having different frequencies from two directions having a predetermined crossing angle (± θ _FS ), Diffracted light is generated by the edge of the rectangular diaphragm FS.

Here, since the two alignment lights (LB ₁ and LB ₂ ) incident on the rectangular diaphragm FS can be treated as plane waves, the projection lens 3
In the pupil plane P of, it can be analyzed as Franfofer diffraction.

Qualitatively, as shown in FIG. 15, on the pupil plane P (Fourier plane) of the projection lens 3, the alignment light shown by the solid line is drawn.
A diffraction intensity distribution DF ₁ by LB _1, the diffraction intensity analysis DF ₂ by the alignment light LB ₂ shown by a broken line is formed symmetrically pupil center O.

Here, in particular, diffracted light that adversely affects the alignment accuracy will be described.

As shown in the figure, the diffracted light of the alignment light LB ₂ having the intensity I _N is mixed into the alignment light LB ₁ having the intensity I ₀ which forms the beam spot at the position of BS ₁ , while the beam spot is formed at the position of BS _2. Alignment light LB with intensity I _0
2 is mixed with the diffracted light of the alignment light LB ₁ having the intensity I _N.

From this, in FIG. 14 described above, the noise diffracted light lb ₁ generated by the alignment light LB ₁ of the frequency f ₁ travels on the same optical path as the alignment light LB ₂ of the frequency f ₂ as shown by the broken line. Therefore, the alignment light LB ₂ is aligned with the wafer mark WM while being beaten by the noise light lb _1.
Irradiation is performed at an incident angle of θ _WM .

On the other hand, the noise diffraction beams lb ₂ generated by the alignment light LB ₂ of frequency f ₂ also, for the progress of the alignment light LB ₁ the same optical path and the frequency f _1, the alignment light LB ₁ was beat by noise light lb ₂ Wafer mark WM +
Irradiation is performed at an incident angle of θ _WM .

FIG. 16 shows how diffracted light is generated on the wafer mark WM, and as shown in the figure, the normal line (optical axis) of this mark WM.
With the -1st-order diffracted light LB ₂ alignment light LB ₂ (-1) of the -1st-order diffracted light lb ₁ noise diffracted light lb ₁ and (-1) is generated in the direction of + 1st-order diffracted light LB ₁ alignment light LB ₁ (+
1) and the noise diffracted light lb ₂ and the + 1st order diffracted light lb ₂ (+1) is generated.

Further, in FIG. 15 just described, it can be seen that the diffracted light of each alignment light is also mixed in the center O of the pupil position of the projection lens 3. Therefore, in FIG. 14, the noise diffracted light (lb ₁ ′, lb ₂ ′) of each alignment light by the field stop FS having the rectangular opening advances on the optical axis in a beat state, and the wafer mark It will illuminate the WM vertically.
Then, as shown in FIG. 16, 0th-order diffracted light (lb ₁ (0) ′, lb ₂ (0) ′) is generated in the normal direction of the mark WM.

As described above, since the detector provided at the position conjugate with the pupil plane P of the projection lens detects all the light generated in the normal direction (optical axis direction) of the wafer mark WM, noise in the detected light is generated. When the diffracted light is included, the detected optical beat signal includes a large detection error.

Therefore, the present invention solves all the above problems,
An object of the present invention is to provide a high-performance position detection that eliminates the adverse effect of the diffracted light generated by the aperture stop on the alignment accuracy and can always perform highly accurate and stable position detection.

[Means for solving the problem]

An alignment optical system is provided, and the coherent light from the alignment optical system is applied to a diffraction grating formed on an object to be position-detected in two predetermined directions to cause interference of the diffracted light. In the alignment device for photoelectrically detecting the intensity of the interference fringes, diaphragm means is provided at a position conjugate with the object to be inspected in the alignment optical system, and the edge forming the aperture of the diaphragm means is It is designed to be inclined with respect to the direction corresponding to the groove direction of the diffraction grating.

Further, the length of the edge forming the aperture of the diaphragm means is L, and the pitch of the interference fringes formed on the diaphragm is
P _FS , where m is an integer, at least one pair of opposing edges may satisfy L = mP _FS .

In particular, it is preferable that the aperture of the aperture is formed in a parallelogram shape. Specifically, the inclination of the edge of the aperture is θ, and the pitch of the interference fringes formed on the aperture means.
P _FS , P is the pitch of the diffraction grating marks formed on the wafer
_WM , W is the length of the edge of the aperture of the diaphragm means corresponding to the array direction of the diffraction grating, and h is the length of the edge of the aperture of the diaphragm means in the direction corresponding to the groove direction of the diffraction grating.
When the allowable alignment error on the wafer is x _WM , It is desirable to configure so as to satisfy the following.

Further, a light blocking means for blocking the diffracted light generated by the diaphragm may be provided closer to the object to be inspected than the diaphragm means.

More preferably, It is better to configure so that

(Operation) Up to this point, the adverse effect on the alignment accuracy due to the noise diffracted light generated by the rectangular diaphragm FS has been briefly described, but below, this physical phenomenon will be further analyzed.

As shown in FIG. 16, + 1st-order diffracted light LB ₁ generated by the alignment light LB ₁ in the normal direction of the wafer mark WM (+
The amplitude of 1) is A ₁ , the alignment light LB ₂ causes the −1st-order diffracted light LB ₂ generated in the normal direction of the wafer mark WM (−1) to have the amplitude of A ₂ , and the alignment light LB ₂ travels on the same optical path. The noise diffracted light lb ₁ causes the amplitude of the −1st order diffracted light lb ₁ (-1) generated in the normal direction of the wafer mark WM to be b ₁ , and the noise diffracted light lb ₂ traveling on the same optical path as the alignment light LB ₁ + 1st order diffracted light generated in the normal direction of the wafer mark WM
When the amplitude of lb ₂ (+1) is b ₂ , the optical beat signal S _WM photoelectrically detected by the detector is given by the following equation.

S _WM = A ₁ cos (ω ₁ t + φ ₁ ) + A ₂ cos (ω ₂ t + φ ₂ ) + b ₁ c
os (ω ₁ t + δ ₁ ) + b ₂ cos (ω ₂ t + δ ₂ ) ... (1) Since the signal detected by the detector is a time average of intensity (square of amplitude), Δω = ω ₁ −ω The detection signal strength when the value is ₂ is as follows.

Now, assuming that the pitch of the wafer mark WM is P _WM and the wafer is displaced by x in the X direction, the phase difference corresponding to this is Becomes

Therefore, in order to simplify the explanation, A ₁ = A ₂ ≡A,
Assuming that b ₁ = b ₂ ≡b and the initial phases other than the phase difference are all zero due to the movement of the wafer, the equation (2) is given by the following equation. As can be seen from the equation (3), the optical beat signal observed by the detector is only the second and third terms on the right side of the equation (3), and the second term according to the displacement of the wafer in the X direction. , The third term becomes a noise component that has nothing to do with the displacement of the wafer in the X direction.

Therefore, the optical beat signal S _W observed by the detector is
It can be understood that the phase difference between the drive signal DS of AOM and the drive signal DS of AOM is detected as the difference between the combined waveform of the second term and the third term on the right side of Expression (3).

Here, the phase change α associated with the displacement x of the wafer in the X direction α
Can be expressed by the following equation.

Due to the adverse effect of the error component term of the third term of the equation (3) described above, the phase difference actually detected is as shown in FIG.
It becomes α _P and the phase error becomes maximum when α = π / 2.

In this case, if A ² + b ² >> 2Ab,
The phase error amount α _e (= α−α _P ) is Can be approximated.

Further, if A >> b, then this equation (5) becomes Can be approximated.

Thus, alignment, drive signal difference AOM | DS ₁ -DS ₂ | and the relative phase difference between the optical beat signal S _WM obtained by the detection is performed in correspondence to the displacement of the wafer Therefore, it can be understood that the phase error directly becomes a position error and greatly affects the detection accuracy.

Here, when the phase error alpha _e, the pitch of the wafer mark WM of the diffraction grating formed on a wafer and P _WM, the pitch of the interference fringes formed on the wafer mark because P _WM / 2, the wafer The position error x _e is Becomes

As an example, if the wafer mark WM pitch is 8 μm and the wafer position error is 0.02 μm or less, from (4) in the above equation, Becomes

In this case, the noise diffracted light (lb ₁ , lb ₂ ) contained in each alignment light (LB ₁ , LB ₂ ) by the rectangular diaphragm FS
As for the ratio of b / A = α _e / 2 in the equation (6), it is necessary to satisfy the following condition.

Therefore, the amplitude of the noise diffracted light that actually mixes with each alignment light on the pupil plane P (Fourier plane) of the projection lens is obtained. On this pupil plane P, as described above, it can be analyzed as Franfofer diffraction, so that the field stop FS
If the sine value of the exit angle of the diffracted light generated from is ξ, and ξ is the coordinate on the pupil plane (Fourier plane), this pupil plane P
The amplitude distribution of each diffracted light at is given by the following equation.

However, the amplitude of the diffracted light is b, the amplitude of the plane wave incident on the field stop FS is A ₀ , the wavelength of the alignment light (LB ₁ , LB ₂ ) is λ, the field stop width w, and the beam from the center O of the pupil plane P. The distance to the spot (BS ₁ , BS ₂ ) is ± ξ ₀ (= sin ± θ _FS = ± λ
/ 2P _FS ), alignment light (L
The incident angles of B ₁ and LB ₂ ) are ± θ _FS , and the pitch of the interference fringes formed on the field stop is P _FS and sin0 / 0 = 1.

FIG. 18 is a graph of the equation (9), and FIG. 18A shows the diffraction amplitude distribution by the alignment light LB _{2 on} the pupil plane P of the projection lens.
(B) is the alignment light LB on the pupil plane P of the projection lens
_The diffraction amplitude distribution by ₂ is shown. It is understood that these two figures are supported by the diffraction intensity distribution shown in FIG. 15 described above on the pupil plane P of the projection lens.

For example, in the pupil plane P of the projection lens, FIG.
As shown in, the beam spot position B of the alignment light LB ₁
The amplitude b of the noise diffracted light lb ₂ at S (ξ = + ξ ₀ ) or the noise diffracted light lb at the beam spot position BS ₂ (ξ = −ξ ₀ ) of the alignment light LB ₂ as shown in FIG. 18 (b). _When the magnitude of the amplitude b of ₁ is calculated, the equation (9) is as follows.

As an example, the pitch of the interference fringes formed on the field stop
_Assuming that P _FS is 4 μm and the width W of the rectangular diaphragm is 50 μm, equation (10) becomes Is larger than the value of b / A ₀ ≦ 0.016 in the equation (7) described above.

That is, in this case, in the configuration in which the rectangular field stop FS is applied, 0.02 due to the influence of diffraction from the stop FS.
It can be seen that it becomes difficult to secure accuracy of μm or less.

Further, the amplitude b of one noise light that travels in the optical axis direction by the rectangular diaphragm FS and passes through the center O (ξ = 0) of the pupil plane P of the projection lens is given by the following equation (9): Therefore, if the width w of the field stop and the pitch of the interference fringes formed on the field stop are set to the same values as P _FS , Therefore, the value becomes larger than 0.016 in the above-mentioned expression (7), and as a result, it becomes difficult to perform highly accurate alignment.

As described above, the present invention focuses on and reduces the physical phenomenon that the noise diffracted light generated by the field stop FS provided at the conjugate position with the wafer or reticle adversely affects the alignment accuracy in the alignment optical system. Therefore, the shape of the edge of the field stop FS with respect to the groove direction of the diffraction grating mark formed on the wafer is sufficiently considered.

[Embodiment 1] In FIG. 1 showing a schematic configuration diagram of this embodiment, the same members as those in FIG. 13 described above are designated by the same reference numerals.

13 is different from FIG. 13 in structure in that a field stop FS having a parallelogram-shaped opening as shown in FIG. 2 is provided in place of the field stop FS having a rectangular opening. Therefore, description of other configurations and operations will be omitted. The operation is described in detail, for example, in JP-A-62-56818 by the same applicant as the present case.

The edges (e ₁ , e ₂ ) of the field stop FS extend along the Y direction corresponding to the arrangement (pitch) direction of the wafer marks WM, as shown in FIG. 2, and the edges (e ₃ , e ₄₎ ) Extends along a direction inclined by θ with respect to the Z direction corresponding to the groove direction (perpendicular to the arrangement direction) of the wafer marks WM.

Then, as shown in FIG. ₁ , the coherent light beams (LB ₁ , LB ₂ ) having different frequencies are incident on the field stop FS having the parallelogram-shaped opening so that they are symmetrical with respect to the optical axis. Irradiate in two directions with an angle (± θ _FS ).

Then, as shown in FIG. 3, Z of the aperture of the field stop FS is
Depending on the edge (e ₃ , e ₄ ) inclination θ with respect to the direction, the generation direction of the diffracted light of each coherent light on the pupil plane P of the projection lens 3 is defined by ξη coordinates of the pupil plane P (Fourier plane) of the projection lens. While being inclined by θ with respect to the ξ direction, the generation direction of the diffracted light of each coherent light by the edges (e ₁ , e ₂ ) of the opening along the x direction is η in the pupil plane P (Fourier plane) of the projection lens. Direction. The sine value of the exit angle corresponding to the XY direction of the diffracted light generated from the field stop FS in FIG. 1 is set as the ξη coordinate on the pupil plane.

Therefore, on the pupil plane P (Fourier plane) of the projection lens 3, each beam spot position (BS ₁ , BS ₂ ) is centered,
The diffraction intensity distribution and η along the direction inclined by θ with respect to the ξ direction
A diffraction intensity distribution along the direction is formed.

At this time, specifically, the diffraction intensity of the alignment light LB ₁ forming the beam spot BS ₁ becomes a diffraction intensity distribution DF ₁ that attenuates from the origin O toward the periphery as shown in FIG. However, it has a two-dimensional spread in which the maximum value of each diffraction intensity exists at the position indicated by a black dot.

As shown in the figure, the noise diffracted light which is distributed two-dimensionally is at the pupil center O and the beam spot BS ₂ position (−ξ ₀ , 0).
It can be seen that it has a slight adverse effect on.

Therefore, as described above, in order to reduce the noise light reaching the light receiving surface of the detector provided at the position conjugate with the pupil plane P (Fourier plane) of the projection lens 3 (or the position conjugate with respect to the wafer). , When the alignment light (LB ₁ , LB ₂ ) is transmitted to the wafer mark WM, at the ξ η coordinate of the pupil plane P of the projection lens, this pupil center position O or the position where the beam spot of each alignment light is formed (BS ₁ , B
The inclination θ of the edges (e ₃ , e ₄ ) of the field stop must be determined so as to reduce the influence of the diffracted light of the rectangular stop FS on S ₂ ).

Therefore, when the amplitude distribution in the ξη coordinate of the pupil plane P (Fourier plane) of the projection lens 3 shown in FIG. 4 is analyzed, the following equation is obtained.

However, the length of the horizontal edge of the field stop FS in the direction corresponding to the arrangement direction of the wafer marks WM is W, and the length of the vertical edge in the direction perpendicular to this horizontal edge (the direction corresponding to the groove direction of the wafer marks WM). Is h, the inclination of the lateral edge is θ, sin0
/ 0 = 1.

Therefore, first, as shown in FIG.
Diffracted light DF ₂ by LB ₂ are coordinate (+ ξ _{0, 0)} the beam spot position BS noise light amplitude mixed in alignment light LB ₁ that passes through _one or diffracted light DF ₁ by the alignment light LB _1, of the coordinates Beam spot position BS of (-ξ ₀ , 0)
The amplitude of the noise light mixed in the alignment light LB ₂ passing through ₂ is obtained from the above equation (11) as follows.

However, the light reaching the center O of the pupil and the beam spot BS
The sine of the angle θ _FS formed by the light reaching ₁ (or BS ₂ )
sin θ _FS is ξ ₀ , the wavelength of the alignment light (LB ₁ , LB ₂ ) is λ, the pitch of the interference fringes formed on the field stop is P _FS , ξ ₀
= Λ / 2P _FS .

Here, as shown in FIG. 5, when the lens 17 and the projection lens 3 satisfy the sine condition, the composite magnification M of the lens arranged between the field stop FS and the wafer is Becomes

The pitch of the wafer mark WM can be calculated by the equation (13).
P _WM , the incident angle of each alignment light that illuminates the diffraction grating mark WM on the wafer mark is θ _w , sin θ _w = ξ ₀ ′, ± n from the wafer mark WM by the alignment light (LB ₁ , L ₂ )
When the next diffracted light is detected, the relationship of ξ ₀ = Mξ ₀ ′ = nMλ / P _WM P _WM = 2nMP _FS is established.

Therefore, (12) above Can be expressed as

As described above, as can be seen from the expressions (12) and (14), by finding the optimum range of the edge inclination θ of the aperture shape of the field stop FS, it is possible to eliminate the factors that adversely affect the alignment accuracy. it can.

Therefore, first, considering the situation where the error included by the diffracted light is maximum, if the value of the term of sin is set to 1 so that the value of the numerator in the above formula (14) becomes maximum, (12) From the formula, From equation (14), Becomes

As an example, as described in the section of action, in order to secure the alignment accuracy of 0.02 μm or less on the wafer, b / A ₀ ≦ 0.016 is set from the formula (8), and the distance between the field stop FS and the wafer is set. The composite magnification M of the lenses arranged in
(Θ _W = θ _FS ), the diffracted light to be detected is ± first-order diffracted light, and the pitch P _WM (= 2P _FS ) of the diffraction grating mark WM on the wafer is 8 μ.
m, the horizontal edge length W of the field stop FS is 50 μm, and the vertical edge length h in the direction perpendicular to the horizontal edge is 30 μm, the vertical edge inclination θ is given by the above equation (15) or ( 1
From equation 6), θ ≧ 3.9 °.

Therefore, in order to reduce the influence of noise diffracted light mixed in the alignment light (LB ₁ , LB ₂ ), it is sufficient to configure the edges (e ₃ , e ₄ ) of the opening satisfying θ ≧ 3.9 °. .

Generally, (right side) in equation (15) or (16)
The alignment accuracy is guaranteed by providing the field stop FS having the edge inclination θ such that the value of is greater than or equal to the value on the left side.

Now, assuming that A in the expression (6) is A ₀ , x _WM is an allowable alignment error (μm) on the wafer, and x _e in the expression (7) is x _WM , the following expression is obtained.

Therefore, the following equation can be derived from the equations (15) to (17).

When a field stop having an aperture that satisfies the above formulas (18) and (19) is provided in the alignment optical system, the field stop FS
Since the influence of the diffracted light (lb ₁ ′, lb ₂ ′) generated in the optical axis direction is not taken into consideration in the optical path on the wafer side of the field stop FS in the alignment optical system, for example, the broken line in FIG. It is desirable to dispose a light shielding member 50 that allows only the 0th order light to pass therethrough, as shown in FIG. This makes it the first time
Alignment accuracy on the wafer of x _WM (μm) or less is guaranteed.

Next, in order to reduce the influence of diffracted light (lb ₁ ′, lb ₂ ′) generated in the optical axis direction by the field stop FS, the above (11)
If ξ = 0 and η = 0 in the equation, the effect of noise diffracted light mixed in the optical path on the optical axis can be obtained, and (15)
Similar to the formulas and the formula (16), the following formula is obtained when the numerator in the formula (11) is maximum, that is, when the value of the term of sin is 1.

However, A ₀ is the amplitude of one alignment light that irradiates the field stop FS in two directions, and b is the amplitude of noise diffracted light generated in the optical axis direction by the field stop FS.

In order to remove this noise light, similar to the above,
When A in the equation (6) is A ₀ and x _e in the equation (7) is x _WM , the above equations (20) and (21) are
It can be expressed as follows.

Therefore, generally, if the edges (e ₃ , e ₄ ) of the aperture of the field stop FS are configured so as to satisfy the expression (22) or the expression (23), the field stop FS generates in the optical axis direction. It is possible not only to reduce the influence of the diffracted light (lb ₁ ′, lb ₂ ′) that is generated, but also to reduce the noise diffracted light mixed in the alignment light at the same time.

That is, by configuring the field stop FS so as to satisfy the condition (22) or (23), it is possible to eliminate all the influence of noise diffracted light that adversely affects the optical beat signal to be detected. be able to.

As an example, similarly to the above-mentioned value, M = 1 (θ _WM = θ
_FS ), P _WM (= 2P _FS ) = 8 μm, w = 50 μm, h = 30 μ
If m, x _WM ≦ 0.02 μm, θ ≧ 16.7 ° is uniquely obtained from the formula (22) or the formula (23), and it may be configured to satisfy this condition.

When the edge of the field stop FS is configured to satisfy the expression (22) or the expression (23), the light blocking member shown in FIG.
50 can be eliminated.

Second Embodiment Next, a second embodiment of this embodiment will be described.

As shown in FIG. 13, when the alignment is performed by the conventional configuration in which the field stop FS having the rectangular opening is provided, as described in the item of action, 2Ab cos ω of the third term of the equation (3) is used.
t has a direct adverse effect on the detection signal, and the value of this Ab is calculated from the equation (9) previously obtained, Is required.

An actual graph of the equation (24) is shown in FIG. As shown in the figure, this function is an optical beat signal that causes a maximum error in the light flux that has passed through the ± ξ positions on the pupil plane P (Fourier plane) of the projection lens 3 regardless of the movement of the wafer. I understand.

Now, assuming that the composite magnification of the lens arranged between the field stop FS and the wafer is M (= sin θ _FS / sin θ _WM ), the wavelength of the alignment light is λ, and ± 1st order diffracted light is detected from the wafer mark WM. The pitch P _{FS of the} interference fringes formed on the field stop FS is as follows.

However, sin θ _WM = λ / P _WM .

Now, as shown in FIG. 7, in the field stop FS having a rectangular opening, when the lengths of the lateral edges corresponding to the arrangement (pitch) direction of the wafer marks WM are W and m are integers, W = When configured so as to satisfy mP _FS (= mP _WM / 2M), the equation (24) described above becomes Becomes

Here, from equation (26) above, ± ξ ₀ (= Mλ / P _WM )
When the value of Ab at the position is calculated, Therefore, as shown in FIG. 8 which is a graph of the above equation (26), the value of Ab is inverted between positive and negative in the vicinity of ± ξ ₀ on the Fourier plane, so that the noise light obtained by the detector is It can be seen that the influence can be suppressed. In the above-described embodiment, the case where the ± first-order diffracted light from the wafer mark WM is detected has been described, but ± n from the wafer mark WM is detected.
It goes without saying that it is also effective when detecting the second-order diffracted light.

Further, in the present embodiment, only the influence of noise light mixed in the alignment light (LB ₁ , LB ₂ ) for irradiating the diffraction grating mark WM formed on the wafer in two directions with a predetermined incident angle was considered. Since this is the case, as shown by the broken line in FIG. ₁ , only the alignment lights (LB ₁ , LB ₂ ) of different frequencies are allowed to pass through the optical path on the wafer side of the field stop FS in the alignment optical system. It is desirable to arrange the light shielding member 50.

In the first and second embodiments described above, the optical beat signal including the position information from the wafer is detected, and the alignment is performed so that the phase difference between the acousto-optic device (AOM) and the signal to be driven becomes zero. However, the reference light obtained by dividing the alignment light (LB ₁ , LB ₂ ) frequency-modulated by the acousto-optic device by a beam splitter etc. is photoelectrically detected, and this reference signal and the position from the wafer are detected. The alignment may be performed so that the phase difference from the optical beat signal containing information becomes zero.

[Third Embodiment] Next, a third embodiment of the present invention will be described with reference to FIG. 9, the same members as those in FIG. 1 are designated by the same reference numerals.

In this embodiment, by irradiating the wafer with alignment light (LB ₁ , LB ₂ ) through the reticle and the projection lens,
The relative alignment between the reticle and the wafer is performed by photoelectrically detecting the optical beat signal including the position information from the reticle and the optical beat signal including the position information from the wafer.

Since the alignment optical system has been described above, a brief description will be given. The laser light emitted from the laser 10 is split into two by the beam splitter 12, and then each AOM1
The light fluxes are modulated to have mutually different frequencies by 4a and 14b, and irradiate the field stop FS in two directions having a predetermined incident angle (± θ _FS ).

As shown in FIG. 2, the aperture of the diaphragm FS is formed in a parallelogram shape, and each alignment light (LB ₁ , LB ₂ ) through the field diaphragm FS is reflected by the reflection mirror 17 and the objective lens. 18, beam splitter 19, dichroic mirror 20
Diffraction grating mark RM formed on the reticle via
(Hereinafter, referred to as reticle mark) is a predetermined incident angle (± θ
Irradiate in two directions with _RM ).

Here, the incident angle when the alignment light (LB _1, LB ₂₎ illuminates the field stop FS and (± theta _FS), the incident angle when the alignment light (LB _1, LB ₂₎ irradiates the reticle mark RM relationship between the (± θ _RM) is, as shown in FIG. 5, when the magnification of the lens 18 which is disposed between the field stop FS and the reticle and _{m 1, ± θ FS = ±} m ₁ θ _RM .

As shown in FIG. 10 (a), the reticle has an edge inclined by θ similar to the inclination θ of the edge (e ₃ , e ₄ ) of the field stop FS described above at a position adjacent to the reticle mark RM. The parallelogram shaped transmission window portion P ₀ is provided, and the alignment light irradiates an area that simultaneously covers the reticle mark RM and the parallelogram shaped transmission window portion P ₀ .

The ± 1st-order diffracted light for detection generated in the normal direction (optical axis direction) by the alignment light (LB ₁ , LB ₂ ) irradiated with this reticle mark RM in two directions is the dichroic mirror 20,
Beam splitter 19, lens 200, spatial filter 201
Reach

This spatial filter is provided at a position conjugate with the pupil of the photographing lens 3 and travels in the optical axis direction (± first-order diffracted light from the reticle mark RM, and wafer mark WM described later).
(± 1st order diffracted light from) is transmitted, and an unnecessary light flux is filtered.

The ± 1st-order diffracted light for detection from the reticle mark RM that has passed through the spatial filter 201 is reflected by the beam splitter 203, and the optical beat signal S _RM that includes the position information of the reticle 1 at the detector 205 via the mask member 204. Is photoelectrically detected.

The mask member 204 is provided at a position conjugate with the reticle, and as shown in FIG. 11 (a), the reticle mark RM
The opening A _RM is provided corresponding to the reticle mark RM so that only the ± 1st-order diffracted light from is transmitted.

On the other hand, as shown in FIG. 10 (a), the reticle mark RM
Parallelogrammatic transmission window portion P ₀ provided at a position adjacent to
Alignment (LB ₁ , LB ₁ ) via the method of irradiating the diffraction grating mark (wafer mark) WM formed on the wafer via the projection lens 3 in two directions with a predetermined incident angle θ _WM. Become.

The wafer mark WM is provided corresponding to the transmission window portion P ₀ formed on the reticle, as shown in FIG. 10 (b).

Now, the alignment light (LB ₁ , LB ₂ ) is the reticle mark.
The relationship between the incident angle (± θ _RM ) when irradiating RM and the incident angle (± θ _WM ) when the alignment light (LB ₁ , LB ₂ ) irradiates the wafer mark WM is shown in FIG. As described above, when the imaging magnification of the lens 18 arranged between the field stop FS and the reticle is m ₂ , ± θ _RM = ± m ₂ θ _WM .

The pitch relationship between the reticle mark RM and the wafer mark WM is P _WM = m ₂ P _RM .

± 1 for detection that is generated in the normal direction (optical axis direction) of the wafer mark WM by the alignment light (LB ₁ , LB ₂ ) that irradiates the wafer mark with such a relationship of the incident angle.
The next-order diffracted light passes through the projection lens 3 and again passes through the parallelogram-shaped transmission window portion P ₀ .

At this time, the 0th-order diffracted light of the alignment light (LB ₁ , LB ₂ ) that specularly reflects the wafer mark WM at the same incident angle as the incident angle also enters the transmission window P ₀ at an incident angle of ± θ _RM , and again. Since the light is diffracted, there may be a situation in which the ± 1st order diffracted light for detection is adversely affected.

Therefore, as shown in FIG. 2, in the opening of the transmission window P ₀ in this embodiment, the inclination θ of the edges (e ₃ , e ₄ ) is derived in the first embodiment (22). ) Or (23) is preferably provided.

Now, the ± 1st-order diffracted light for detection that has passed through the parallelogram-shaped transmission window P ₀ passes through the dichroic mirror 20, the beam splitter 19, the lens 200, the spatial filter 201, the lens 202, the beam splitter 203, and the mask member 206. To the detector 207 via. The optical beat signal _{SWM including the} wafer position information is photoelectrically detected by the detector 207.

At this time, the mask member 206 is provided at a position where the mask member 204 is conjugate with the wafer, and as shown in FIG. 11 (b), only the ± first-order diffracted light from the wafer mark WM passes therethrough. An opening A _WM is provided corresponding to the wafer mark WM.

As described above, the optical beat signals from the marks (RM, WM) photoelectrically detected by the detectors (203, 207) are the field stop FS.
In addition, noise diffracted light mixed due to the transmission window P ₀ of the reticle can be detected independently while being reduced. Therefore, a highly reliable optical beat signal is obtained, and highly accurate alignment is guaranteed.

Incidentally, in the apparatus of the present embodiment, the objective lens 18 is used as proposed in JP-A-63-283129 by the same applicant as this case.
May be configured by a double focus lens, and the detection light from each mark (WM, RM) via this double focus lens may be detected.

In each of the embodiments as described above, the field stop PS or the transmission window portion P ₀ formed on the reticle is described as a parallelogram shape, but the reticle or the diffraction grating mark (WM, WM formed on the wafer If the edge direction of the field stop FS provided at a position conjugate with RM) is inclined with respect to the groove direction of these diffraction grating marks (the direction perpendicular to the array direction) The influence of noise diffracted light can be reduced, and the shape of the opening may be, for example, a rhombus or a trapezoid.

Also, the shape of the aperture of the field stop FS and the transmission window P ₀ of the reticle, the lengths of the edges of these apertures (lengths that are integer multiples of the pitch P of the interference fringes formed on them), the field of view. The same effect as in the present embodiment can be achieved by appropriately combining a light blocking member for blocking (removing) noise diffracted light generated by the aperture FS and the transmission window P ₀ of the reticle.

Further, when noise diffracted light is mixed in the detection light on the detection side, a mask member having the above-described opening shape or an integer multiple of the pitch of the interference fringes of the edges of the opening is provided on the detection side of the wafer. Alternatively, it may be provided at a position conjugate with the reticle.

Further, in this embodiment, the alignment light (LB ₁ , LB ₂ ) is applied to the reticle mark RM or the wafer mark WM in two directions having a predetermined incident angle so as to be symmetrical with respect to the optical axis. , The beam spots (BS ₁ , B 2) on the pupil plane (Fourier plane) of the projection lens in the direction of the perpendicular bisector connecting the beam spots (BS ₁ , BS ₂ ) of the alignment light.
The marks (RM, WM) may be irradiated with alignment light (LB ₁ , LB ₂ ) so that S ₂ ) shifts by the same distance.

 Next, a modification of the third embodiment will be described.

In the third embodiment described above, the reticle mark RM formed on the reticle and the wafer mark WM of the diffraction grating formed on the wafer are provided with alignment light (LB ₁ , LB ₂ ) in predetermined two directions. Alignment is achieved by detecting the optical beat signal of the ± nth-order diffracted light generated in the optical axis direction by irradiation, but the pitch P _{RM of the} diffraction grating that constitutes this reticle mark _RM is 1 / You may double and perform alignment.

At this time, if the pitch of the reticle mark RM is P _RM ′, And the reticle mark RM is established as shown in FIG.
The diffraction angle of the ± n-order diffracted light is doubled as shown in FIG. The pitch of the interference fringes formed on the reticle mark RM is the same.

Specifically, as shown in the figure, the alignment light LB ₁ changes to θ _RM.
When the reticle mark RM is irradiated at the incident angle of, first, the alignment light LB ₁ travels along the reverse optical path -1
Order diffracted light LB ₁ (-1), based on the traveling direction of the diffracted light _{LB 1 (-1), + 2θ} RM direction 0-order diffracted light LB
₁ (0), + _{1st order} diffracted light LB ₁ (+1), −4θ _RM direction −
The −2nd order diffracted light LB ₁ (−2) is generated in the 2θ _RM direction.

On the other hand, when the alignment light LB ₂ irradiates the reticle mark RM at an incident angle of θ _RM , first, the + first-order diffracted light LB that travels along the reverse optical path of this alignment light LB _2
2 (+1) and the traveling direction of this diffracted light LB ₂ (+1) as a reference, the 0th-order diffracted light LB ₂ (0), −4θ in the −2θ _RM direction.
In _RM direction -1-order diffracted light LB ₂ (-1), to + 2θ _RM direction -
Second-order diffracted light LB ₂ (+2) is generated.

However, in FIG. 12, the diffracted light generated in the clockwise direction with the 0th-order light of each alignment light as a boundary is positive, and the diffracted light generated in the counterclockwise direction is negative.

Therefore, diffracted lights that are diffracted by the reticle mark RM and travel in the same direction interfere with each other, and an optical beat signal including information on the position of the reticle is obtained by photoelectrically detecting one of the interfering lights. . Therefore, for example, as shown in FIG. 9, beat light by light LB ₂ (+1) and LB ₁ (0) that travel in a direction opposite to the alignment light (LB ₁ , LB ₂ ) is photoelectrically detected. For this purpose, first, a beam splitter or the like is arranged in the optical path between the field stop FS and the reticle 1. Then, the beam splitter splits the beat light including the position information of the reticle into another optical path, and the detector is provided at a position where the reticle 1 is optically pupil-conjugated to the reticle 1. The beat light including information can be photoelectrically detected with extremely high accuracy.

As described above, also in the configuration described above, the field stop having the opening as described in the first and second embodiments, or the field stop FS having the opening having a rhombic shape, a trapezoidal shape, or the like.
As shown in FIG. 9, by providing the position at a position conjugate with the reticle or the wafer, highly accurate alignment is guaranteed.

Although the heterodyne interference type position detecting device has been described in the present embodiment, it is needless to say that the same effect can be expected and extremely effective for the homodyne interference type position detecting device.

〔The invention's effect〕

As described above, according to the present invention, it is possible to exclude the noise diffracted light generated by the field stop or the transmission window portion formed on the reticle from being mixed in the detection signal, so that the reliability is always high. Precise alignment can be achieved.

[Brief description of the drawings]

FIG. 1 is a schematic sectional configuration diagram of an alignment apparatus in a first embodiment of the present invention, FIG. 2 is a plan view of a field stop FS having a parallelogram-shaped opening, and FIG. 3 is a pupil plane of a projection lens ( Diagram showing the two-dimensional intensity distribution of alignment light (LB ₁ , LB ₂ ) on the Fourier plane. Fig. 4 is the two-dimensional intensity distribution of alignment light LB ₁ on the pupil plane (Fourier plane) of the projection lens. FIG. 5 is a diagram showing the total magnification relationship of the lens arranged between the field stop FS and the diffraction grating mark WM (wafer mark) formed on the wafer, and FIG. 6 is the value of Ab. 7 is a plan view of a field stop FS having a rectangular opening in the second embodiment of the present invention, and FIG. 8 is a view showing the edge of the field stop FS at the pitch P _{FS of the} interference fringes.
FIG. 9 is a graph showing the Ab value when it is an integer multiple of FIG. 9, FIG. 9 is a schematic cross-sectional configuration diagram of the alignment device in the third embodiment of the present invention, and FIG. 10 (a) shows the shape of the reticle mark RM. FIG. 10 (b) is a plan view showing the shape of the wafer mark WM, and FIG. 11 (a) is a plan view of the mask member 204 that transmits only the ± 1st order diffracted light from the reticle mark RM. FIG. 12B is a plan view of the mask member 206 that allows only the ± 1st order diffracted light from the wafer mark WM to pass therethrough. FIG. 12 shows diffracted light when the pitch P _RM of the reticle mark _RM is halved. Fig. 13 is a schematic cross-sectional configuration diagram of a conventional alignment device, Fig. 14 is a diagram showing a simplified alignment optical system, and Fig. 15 is the intensity of diffracted light at the pupil position of the projection lens. Fig. 16 shows the distribution, and Fig. 16 shows the diffraction grating mark WM (wafer marker) formed on the wafer. C) Figure showing how diffraction occurs, Figure 17 shows how the detected optical beat signal contains an error, and Figure 18 (a) shows alignment light LB ₂ at the pupil position of the projection lens. 18B is a diagram showing the amplitude distribution of the diffracted light of FIG. 18, and FIG. 18B is a diagram showing the amplitude distribution of the diffracted light of the alignment light LB ₁ at the pupil position of the projection lens. [Description of main parts] 1 ... Reticle 3 ... Projection lens 4 ... Wafer 10 ... Lasers 14a, 14a ... Acousto-optic element 50 ... Shading member FS ... Field stop RM ... Reticle mark WM ... Wafer Mark P ₀ ...... Transparent window

─────────────────────────────────────────────────── ─── Continuation of the front page (51) Int.Cl. ⁶ Identification number Internal reference number FI Technical indication location H01L 21/30 525L (72) Inventor Nobutaka Magome 1-6-3 Nishi-oi, Shinagawa-ku, Tokyo Stock expression Nikon Co., Ltd. (56) Reference JP 62-56818 (JP, A) JP 64-82624 (JP, A) JP 62-216231 (JP, A)

Claims (10)

(57) [Claims] 1. Diffraction generated by irradiating a diffraction grating formed on an object to be detected whose position is detected with coherent light from the alignment optical system in two predetermined directions. In a positioning device that causes light to interfere and photoelectrically detects the intensity of the interference fringes, a diaphragm means is provided at a position conjugate with the object to be measured in the alignment optical system, and an opening of the diaphragm means is formed. The alignment device is characterized in that the existing edge is inclined with respect to the direction corresponding to the groove direction of the diffraction grating. 2. Diffraction generated by irradiating a diffraction grating formed on an object to be position-detected with coherent light from the alignment optical system in two predetermined directions. In a positioning device that causes light to interfere and photoelectrically detects the intensity of the interference fringes, a diaphragm means is provided at a position conjugate with the object to be measured in the alignment optical system, and an opening of the diaphragm means is formed. The edge length is L,
When the pitch of the interference fringes formed on the diaphragm is P _FS and the integer is m, at least one pair of opposing edges is configured to satisfy L = mP _FS. apparatus. 3. The alignment device according to claim 1, wherein the aperture of the diaphragm is formed in a parallelogram shape. 4. An inclination of an edge of the opening is θ, a pitch P _{FS of} interference fringes formed on the diaphragm means, a pitch P _WM of diffraction grating marks formed on the wafer, and an array of the diffraction gratings. The edge length of the aperture of the aperture means corresponding to the direction is W, the edge length h of the aperture of the aperture means in the direction corresponding to the groove direction of the diffraction grating, and the allowable alignment error on the wafer are x. _{When using WM} , The alignment device according to any one of claims 1 to 3, characterized in that: 5. A light shielding means for shielding diffracted light generated by the diaphragm is provided in an optical path closer to the object to be inspected than the diaphragm means. The alignment device according to any one of paragraphs. 6. The inclination of the edge of the opening is θ, the pitch P _{FS of the} interference fringes formed on the diaphragm means, the pitch P _WM of the diffraction grating marks formed on the wafer, and the array of the diffraction gratings. The edge length of the aperture of the aperture means corresponding to the direction is W, the edge length h of the aperture of the aperture means in the direction corresponding to the groove direction of the diffraction grating, and the allowable alignment error on the wafer are x. _{When using WM} , The alignment device according to any one of claims 1 to 3, characterized in that: 7. The alignment device according to any one of claims 1 to 6, wherein the diffracted light is ± n-order diffracted light. 8. An illumination optical system for supplying exposure light to a reticle, a projection optical system for forming a pattern on the reticle on a wafer based on the exposure light, and alignment of the reticle and the wafer. And an alignment optical system for adjusting the coherent light from the alignment optical system to a diffraction grating formed on an object to be position-detected in predetermined two directions. In an exposure apparatus that interferes with each other and photoelectrically detects the intensity of the interference fringes, diaphragm means is provided at a position conjugate with the object to be measured in the alignment optical system, and an edge forming an aperture of the diaphragm means is An exposure apparatus which is inclined with respect to a direction corresponding to a groove direction of the diffraction grating. 9. An illumination optical system for supplying exposure light to a reticle, a projection optical system for forming a pattern on the reticle on a wafer based on the exposure light, and alignment of the reticle and the wafer. And an alignment optical system for adjusting the coherent light from the alignment optical system to a diffraction grating formed on an object to be position-detected in predetermined two directions. In an exposure apparatus that photoelectrically detects the intensity of the interference fringes by causing interference, an aperture unit is provided at a position conjugate with the object to be inspected in the alignment optical system, and an edge forming an aperture of the aperture unit is provided. Length is L,
When the pitch of the interference fringes formed on the stop is P _FS and the integer is m, at least one pair of opposing edges is configured to satisfy L = mP _FS. . 10. The exposure apparatus according to claim 8 or 9, wherein the object to be inspected is the wafer.