JPH05315223A - Alignment apparatus - Google Patents

Abstract

PURPOSE:To achieve that recesses parts in an alignment mark are made substantially symmetric with respect to a beam of light for position detection use and to reduce the detection error of the title apparatus in an alignment operation by a method wherein the width in an alignment direction of the recessed parts in the alignment mark is set to be narrower than the length of three times the wavelength of the beam of light for position detection use. CONSTITUTION:The width (d) of recesses parts 2b in a wafer mark is set to be narrower than the length 3lambda (or 3lambda/n) of three times the wavelength lambda of laser beams LB1, LB2 as beams of alignment detection light. At the inside of the recesses parts 2b, the laser beams LB1, LB12 are attenuated exponentially and hardly reach the bottom part. Consequently, even when a difference in level DELTAh1 is caused at the bottom of each recessed part 2h and the wafer mark becomes asymmetric in the X-direction, the wafer mark is equivalent to a symmetric mark with reference to the laser beams LB1, LB2. Consequently, a measured result by means of the laser beams LB1, LB2 becomes accurately the position of the mark and the detection error of the title apparatus is not caused.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an alignment apparatus suitable for application to an alignment system of a projection exposure apparatus used when manufacturing a semiconductor element, a liquid crystal display element, a thin film magnetic head or the like.

[0002]

2. Description of the Related Art When a semiconductor element, a liquid crystal display element, a thin film magnetic head, or the like is manufactured by using a photolithography technique, an image of a pattern of a photomask or reticle (hereinafter referred to as "reticle") is projected onto an optical projection system. There is used a projection exposure apparatus that transfers the image onto a photosensitive substrate via the. Generally, a semiconductor element or the like is formed of a multi-layered pattern, and therefore, when manufacturing it, it is necessary to transfer various reticle patterns onto a photosensitive substrate in a predetermined order. In this way, when transferring the circuit pattern of another reticle to the photosensitive substrate on which the pattern has already been transferred to each shot area, the alignment between each shot area of the photosensitive substrate and the reticle pattern to be exposed this time is performed. You need to be accurate.

As a method for accurately performing such alignment, a two-beam interference method using two laser beams is known. As shown in FIG. 6, in the two-beam interference method, a wafer mark WM as an alignment mark made of a phase type diffraction grating is formed in the vicinity of each shot area on the wafer 1 as a photosensitive substrate. When the position measurement direction is the X direction, the wafer marks WM are formed at the pitch P in the X direction. Two laser beams LB1 and LB2 are projected onto the wafer mark WM via a projection optical system.
Are irradiated symmetrically. In this case, the laser beam LB1
From the wafer mark WM by + 1st order diffracted light LB11 from the wafer mark WM and − from the wafer mark WM by the laser beam LB2.
The pitch P and the incident angles of the laser beams LB1 and LB2 are set so that the first-order diffracted light LB21 is parallel to each other and is reflected vertically upward from the wafer 1.

When the diffracted lights LB11 and LB21 are caused to interfere with each other by the alignment optical system via the projection optical system, a signal due to predetermined interference is obtained. Then, in the case of the heterodyne system in which the frequencies of the laser beams LB1 and LB2 are made different, the signal due to the interference becomes a signal having a predetermined beat frequency, and the phase of this signal is compared with the phase of the reference signal to obtain X. The position of the wafer 1 in the direction can be accurately detected. Similarly, a wafer mark is also formed in the Y direction orthogonal to the X direction, and the position in the Y direction is detected by this wafer mark. Further, a reticle mark as an alignment mark made of an amplitude type diffraction grating or the like is also formed on the reticle side, and the position of the reticle is detected by this reticle mark. Then, the alignment is performed by directly or indirectly setting the wafer mark and the reticle mark in a predetermined positional relationship.

FIG. 7 shows a conventional process of manufacturing a wafer mark WM as an alignment mark. First, as shown in FIG. 7A, the concave portions 1a which are periodical in the X direction at a pitch P are etched on the surface of the wafer 1. Is formed by. afterwards,
As shown in FIG. 7B, the metal film 2 is deposited on the surface including the recess 1a by the sputtering method. Since the portion of the metal film 2 corresponding to the concave portion 1a becomes the concave portion 2a, a phase type diffraction grating having a high reflectance and a pitch P is formed. Further, the pitch P is, for example, 6 μm, and assuming that the width of the concave portion 2a of the metal film 2 in the pitch direction (measurement direction) is D, it has been conventionally set to approximately D / P = 1/2.

[0006]

However, as shown in FIG.
As shown in (b), in the conventional wafer mark WM, a step of Δh is generated at the bottom of the recess 2a of the metal film 2 by using the sputtering device, and the recess 2a is formed.
Was asymmetrical. When such asymmetrical wafer mark WM is irradiated with two laser beams to detect the position, there is a disadvantage that a detection error occurs. this is,
When the alignment is performed by the two-beam interference method, only the component of the wafer mark WM as the alignment mark in a specific period is extracted and the position of the center of gravity thereof is obtained.
It is considered that a detection error occurs for the asymmetrical wafer mark WM.

In view of the above point, the present invention has an object to provide an alignment apparatus in which a detection error is unlikely to occur even if the alignment mark is asymmetric.

[0008]

An alignment apparatus according to the present invention is, for example, as shown in FIGS. 1 and 4, an alignment mark composed of a convex portion and a concave portion formed in an alignment direction on a test object (1). (WM), an irradiation optical system (5, 13, 14) for irradiating the alignment mark (WM) with light (LB1, LB2) for position detection, and light generated from the alignment mark (WM) by the irradiation of the light. An alignment mark (WM) having a light receiving optical system (5, 14, 13) for receiving the light, and detecting the position of the object (1) using light generated from the alignment mark (WM). The width d of the concave portion (2b) in the alignment direction is set to be narrower than the length 3 · λ which is three times the wavelength λ of the light for position detection. When the surface of the alignment mark (WM) is coated with a photosensitive material having a refractive index n, the width d is set to be narrower than 3 · λ / n.

In this case, the alignment mark (W
M) is, for example, as shown in FIG. 2, a pattern (30a, 30b, 30) including convex portions (29b) and concave portions (29a) formed at the first pitch P1 in the alignment direction.
c) is a second pitch P larger than the first pitch P1
It may be formed repeatedly in 2 in the alignment direction. The width d1 of the concave portion (29a) in the alignment direction is also set to be narrower than the length 3 · λ that is three times the wavelength λ of the light for position detection. Further, it is desirable to form the alignment mark (WM) on the metal film (2) deposited on the test object (1).

[0010]

According to the present invention, since the width d of the concave portion (2b) of the alignment mark (WM) is narrower than the length 3 · λ which is three times the wavelength λ of the light for position detection, the position detection is performed. Light is exponentially attenuated when propagating in the recess (2b), and is not affected by the asymmetric bottom surface even if the bottom surface of the recess (2b) is asymmetric. Therefore, the alignment mark (W
M) is almost equivalent to a completely symmetrical mark, and the detection error becomes very small.

Further, when the two-beam interference method, for example, is used for detecting the alignment mark (WM), the present invention is different from the conventional alignment mark, that is, the mark in which the convex portion and the concave portion have substantially the same width. In this case, the intensity of the first-order diffracted light from the alignment mark may be reduced. However, by increasing the detection sensitivity of the diffracted light, it is possible to deal with the decrease in the intensity of the diffracted light.

The alignment mark (WM)
Is a pattern (3 including a convex portion (29b) and a concave portion (29a) formed at the first pitch P1 in the alignment direction.
0a, 30b, 30c) are repeatedly formed in the alignment direction with a second pitch P2 larger than the first pitch P1.
Then, when the light for position detection propagates in it, it is attenuated to the extent that it is not affected by the asymmetry of the bottom surface. Furthermore, the concave portion (29) with respect to the entire area of the alignment mark (WM) is
Since the area ratio of a) increases compared with the case of FIG. 1, 1
It is possible to increase the intensity of the detection light such as the second-order diffracted light.
Therefore, the SN ratio of the detection signal can be improved.

The alignment mark (WM)
Is formed on the metal film (2) deposited on the test object (1), the recess (2b or 29) on the metal film (2) is formed.
In a), since the light for position detection is exponentially attenuated exponentially, the influence of the asymmetry of the bottoms of the recesses is reduced.

[0014]

EXAMPLE A first example of an alignment apparatus according to the present invention will be described below.
An embodiment will be described with reference to FIGS. 1, 4 and 5.
This embodiment applies the present invention to an alignment system of a projection exposure apparatus. FIG. 4 shows a main part of the projection exposure apparatus of this example. In FIG. 4, the reticle 3 is held on a reticle holder 4, and the exposure light IL is emitted from an illumination optical system not shown in the circuit pattern formed on the reticle 3. Irradiate.
The circuit pattern is reduced to, for example, one-fifth by the projection optical system 5 and transferred onto the wafer 1 coated with a photosensitive material such as a resist. Wafer marks WM as alignment marks are formed at a predetermined pitch in the measurement direction in the vicinity of each shot area on the wafer 1. The wafer 1 is mounted on a wafer stage 7 via a wafer holder 6, and the wafer stage 7 positions the wafer 1 in a two-dimensional plane perpendicular to the optical axis of the projection optical system 5 and the wafer in the optical axis direction. It is composed of a Z stage for positioning 1 and the like.

A movable mirror 7 is installed on the wafer stage 7, and the laser beam of the laser interferometer 8 is reflected by the movable mirror 7 to constantly measure the coordinates of the XY stage in the wafer stage 7. Further, by operating the wafer stage 7 via the driving device 11, the exposure area of the wafer 1 is positioned by, for example, the step-and-repeat method.

The pattern already formed on the wafer 1 and the pattern of the reticle 3 must normally be aligned with an accuracy of about 1/5 or less of the width dimension of the pattern. In order to perform this alignment, a reference mark 10 is attached near the wafer 1 on the wafer stage 7, and an alignment microscope 12 equipped with a light source having the same wavelength as the exposure light IL is arranged on the upper side of the reticle 3. Reference numeral 13 denotes an alignment optical system, which is a wafer mark WM on the wafer 3.
The position is detected by the alignment optical system 13 (details are disclosed in Japanese Patent Application Laid-Open No. 2-272305).

Two laser beams LB1 and LB2 are emitted from the alignment optical system 13, and these laser beams are irradiated onto the wafer 1 through the folding mirror 14 and the projection optical system 5. In this case, the wafer mark WM
From the above, the incident angles of the two laser beams and the pitch of the wafer mark WM are set so that the diffracted light LB0 is generated vertically upward with respect to the wafer 1. In addition, the diffracted light LB0
Is a light in which the + 1st order diffracted light of the laser beam LB1 and the −1st order diffracted light of the laser beam LB2 are mixed, and the diffracted light LB0 enters the alignment optical system 13 via the projection optical system 5 and the folding mirror 14.

In this case, the relative position between the two laser beams LB1 and LB2 irradiated on the wafer 1 and the reticle 3 is determined by using the reference mark 10 provided on the wafer stage 7 using the alignment microscope 12. It can be known by observing through the reticle 3 and detecting the same reference mark 10 by the alignment optical system 13.

FIG. 5 shows the arrangement of the alignment optical system 13. In FIG. 5, the laser beam LB emitted from the laser light source 15 is split into two laser beams LB1 and LB2 by the bisection prism 16. .. One laser beam LB1 is incident on the combining prism 19 via the first acousto-optic modulator (AOM) 17 and the right-angle prism 18, and the other laser beam LB2 is incident on the right-angle prism 23 and the second acousto-optic modulator (AO).
M) 24 and then enters the combining prism 19. The acousto-optic modulators 17 and 24 are driven at different frequencies f ₁ and f ₂ , respectively, and of the lights diffracted by the acousto-optic modulators 17 and 24, the + _1st order diffracted light is divided by the respective drive frequencies f ₁ and f ₂ . Only the frequency gets higher. Two + 1st-order diffracted lights having different frequencies (these are also called laser beams LB1 and LB2) enter the combining prism 19.

The laser beam LB1 transmitted through the synthesizing prism 19 and the laser beam LB2 reflected by the synthesizing prism 19 are focused on the reference grating 21 by the objective lens 20. Two diffracted lights that are diffracted vertically downward with respect to 21 are mixed by the light receiving element 22 and detected. On the other hand, the laser beam LB1 reflected by the synthesizing prism 19 and the laser beam LB2 transmitted through the synthesizing prism 19 are transmitted through the two-division prism 25 and then focused by the objective lens 26 to enter the folding mirror 14 in FIG.

Then, the wafer mark W of the wafer 1 shown in FIG.
The diffracted light LB0 from M passes through the objective lens 26 of FIG.
It is incident on 7. The reference signal from the light receiving element 22 and the alignment signal from the light receiving element 27 are supplied to the alignment processing unit 28, respectively. The alignment signal obtained by photoelectrically converting the diffracted light LB0 by the light receiving element 27 is
It is a sine wave whose frequency is the frequency Δf of the difference between the driving frequencies f ₁ and f ₂ of the acousto-optic modulators 17 and 24. Similarly, the reference signal output from the light receiving element 22 is also a sine wave having a frequency Δf. is there. By detecting the phase difference between the reference signal and the alignment signal, the positional deviation of the wafer 1 can be known. The amount of wafer displacement detected by the alignment processing unit 28 is sent to a main computer that controls the operation of the entire apparatus.

Next, the detailed structure of the wafer mark WM of this example will be described. FIG. 1 shows an enlarged cross-sectional shape of a wafer mark WM of this embodiment. In FIG. 1, recesses are formed on the surface of the wafer 1 at a pitch P in the X direction, which is the measurement direction, and the surface including the recesses. A metal film 2 is formed on the top of the. The material of the metal film 2 is aluminum, gold, copper, chromium or the like. Corresponding to the recesses of the wafer 1, recesses 2b having a width d are formed in the metal film 2 with a pitch P, and two laser beams LB1 having a wavelength λ are formed on the wafer mark.
And LB2 are illuminated. However, both beams LB1
Since LB2 and LB2 have different frequencies, they have slightly different wavelengths in a strict sense.

In this example, the X of the recess 2b of the metal film 2 is
The following conditions are imposed on the width d in the direction. d <3 · λ (1) Further, when a photoresist having a refractive index n (about 1.7) is coated on the metal film 2 as in this embodiment,
The condition of the width d is as follows. d <3 · λ / n (2) Strictly speaking, the wavelength λ in the condition (1) or (2) is the longer wavelength of the laser beams LB1 and LB2. Therefore, the width d of the concave portion 2b of the wafer mark is determined by the laser beams LB1 and LB as alignment detection light.
The length is set to be narrower than the length 3λ (or 3λ / n) which is three times the wavelength λ of 2. For example, laser beams LB1 and LB
When a He-Ne laser beam is used as 2, the wavelength λ is 0.633 μm. The pitch P is, for example, 6 μm.

However, as a result of the experiment, the width d of the recess 26 in the X direction has a remarkable effect when the following conditions are satisfied with respect to the wavelength λ of the laser beams LB1 and LB2 and the refractive index n of the photoresist. Dripping d ≦ 2 · λ / n (3) That is, when the width d is equal to or less than twice the wavelength (λ / n) after the refractive index correction, the detection error during alignment becomes particularly small. Specifically, the refractive index n of the photoresist is 1.7,
The wavelength of the laser beams LB1 and LB2 is 0.633 μm
, The upper limit of the width d determined by the condition (3) is 0.75.
μm. Then, actually, the width d was set to about 0.7 μm, and particularly good results were obtained.

Using a wafer mark as shown in FIG. 1,
Inside the recess 2b, the laser beams LB1 and LB
2 decays exponentially and almost never reaches the bottom. Therefore, even when the step Δh1 is generated at the bottom of the recess 2b and the wafer mark becomes asymmetric in the X direction,
The wafer mark is equivalent to a symmetrical mark with respect to the laser beams LB1 and LB2. Therefore, the measurement result by the laser beams LB1 and LB2 is exactly the mark position, and no detection error occurs. in this case,
As the level difference of the concave portion on the surface of the wafer 1 is increased, the metal film 2
The depth of the concave portion 2b becomes large, and the light entering the concave portion 2b can be further attenuated. Therefore, the detection error can be further reduced.

Next, the wafer mark of the second embodiment of the present invention will be described with reference to FIG. FIG. 2 shows a wafer mark of this embodiment. In FIG. 2, a metal film 2 is formed on the wafer 1 by sputtering. In this embodiment, in the X direction, which is the measurement direction of the metal film 2,
The concave portion 29a and the convex portion 29b each having the width d1 are formed with the pitch P.
1 are arranged side by side to form recessed portion groups 30a, 30b, 30c, .... Then, the recess groups 30a, 30b, 30c,
Are arranged periodically in the X direction at a pitch P2. This pitch P2 corresponds to the pitch P in the example of FIG.
These recess groups 30a, 30b, 30c, ... Are formed following the recess groups formed on the wafer 1. Further, assuming that the total width of each of the recess groups 30a, 30b, 30c, ... Is Q1, and the number of recesses 29a included in each recess group is N, the width Q1 is as follows. Q1 = (N-1) · P1 + d1 (4)

In this example, the width Q1 is set to be the group of concave portions 30a, 3
The pitch P2 of 0b, 30c, ... Is set to approximately 1/2. In order to increase the number N of recesses 29a in each recess group, the pitch P1 may be made smaller than the condition (2), and in the example of FIG. 2, the pitch P1 is set to about twice the width d1 of the recess 29a. is doing. Further, assuming that the wavelength of the alignment light with which the wafer mark WM of FIG. 2 is irradiated is λ, the following condition is imposed on the width d1 of the recess 29a in the X direction. d1 <3 · λ or d1 <3 · λ / n (5)

As a result, the alignment light incident on the recess 29a is almost attenuated midway and does not reach the bottom of the recess 29a. Therefore, even if a step Δh2 occurs at the bottom of the recess 29a and the wafer mark becomes asymmetric in the X direction, the wafer mark becomes equivalent to a symmetrical mark with respect to the alignment light. Therefore, the measurement result by the alignment light becomes the position of the mark accurately, and no detection error occurs. Further, in this example, the area ratio between the concave portions 29a and the convex portions 29b (including the convex portions between the concave portion groups) is brought close to 1 by increasing the number of the concave portions 29a while maintaining the periodicity of the pitch P2. You can Therefore, the intensity of the first-order diffracted light can be increased, and the SN ratio of the detection signal can be increased.

The wafer mark shown in FIG. 3 may be used instead of the wafer mark shown in FIG. In FIG. 3, concave portions 31a and convex portions 31b each having a width d1 are arranged at a pitch P3 in the X direction, which is the measuring direction on the metal film 2 on the wafer 1, to form concave portion groups 32a, 32b, 32c ,. .. Then, these recess groups 32a, 32b, 32
.. are periodically arranged in the X direction at a pitch P2. The entire width of each of the recessed portions is Q2, and the width Q2 is set to approximately P2 / 2. The width d1 of the recess 31a and the pitch P2 of the recess group are the same as in the example of FIG.

Comparing the example of FIG. 3 with the example of FIG.
In the example, the pitch P3 when the concave portions 31a and the convex portions 31b are arranged is larger than the pitch P1 in FIG. Therefore, the number of the concave portions 31a in FIG.
It is less than the number of 9a. Therefore, in the configuration of FIG. 3, the area of the concave portion 31a becomes small, and S of the detection signal is reduced.
It is inferior in terms of N ratio. However, in the configuration of FIG. 3, since the width of the convex portion 31b is wider than the width of the convex portion 29b of FIG. 2, there is a risk that the convex portion 31b collapses and the wafer mark collapses in the wafer mark forming process. It is decreasing.

Although the alignment mark is irradiated with two laser beams to detect the position in the above-described embodiment, the position detection can also be performed by irradiating the alignment mark with one detection beam. You can If the diffraction effect is used also in this case, two diffracted lights that are symmetrically diffracted left and right from the alignment mark may be caused to interfere in the alignment optical system to form an interference fringe.

Although the alignment system of the present embodiment employs the heterodyne system, for example, the homodyne system or two light fluxes having different polarization directions are applied to the alignment mark, and the diffracted light from the mark is analyzed by the analyzer. The same effect can be obtained by applying the present invention to a system in which light is received by interfering with the light through the same. Further, not only the method of detecting the phase of the diffracted light from the alignment mark, but also the alignment mark or the image pickup device for detecting the position by a so-called laser step alignment method disclosed in, for example, JP-A-60-186845. The same effect can be obtained by applying the present invention to the alignment mark of the image processing type alignment apparatus using the. For example, in the case of the laser step alignment method, the alignment mark and the slit-shaped probe light are relatively scanned to detect the intensity change of the diffracted light emitted from the alignment mark in a predetermined direction by the probe light. As a result, position detection is performed.

In the alignment marks shown in FIGS. 1 to 3, only the shape in the X direction (measurement direction) has been described, but nothing has been described in the non-measurement direction (direction perpendicular to the X direction). .. However, the shape in the Y direction may be linear, or a plurality of dot marks may be arranged in the Y direction at a predetermined pitch, and any shape may be used. For example, a linear alignment mark in the Y direction is a one-dimensional grid pattern, and an alignment mark arranged in the Y direction at a predetermined pitch is a two-dimensional grid pattern.

As described above, the present invention is not limited to the above-described embodiments, and various structures can be adopted without departing from the gist of the present invention.

[0035]

According to the present invention, since the concave portion of the alignment mark is substantially symmetrical with respect to the light for position detection, there is an advantage that a detection error in alignment is reduced. In the case of the present invention, the diffracted light from the alignment mark may be weakened, but this can be dealt with by increasing the detection sensitivity of the weak diffracted light.

In addition, the alignment mark has a pattern composed of a convex portion and a concave portion formed at a first pitch in the alignment direction at a second pitch larger than the first pitch in the alignment direction. When it is formed repeatedly, the area of the concave portion can be made closer to the area of the convex portion, so that the intensity of the diffracted light can be increased,
The SN ratio of the detection signal is improved. Further, when the alignment mark is formed on the metal film deposited on the object to be inspected, the light for position detection in the concave portion is rapidly attenuated, so that the effect of the present invention becomes particularly large.

[Brief description of drawings]

FIG. 1 is an enlarged cross-sectional view showing a wafer mark of a first embodiment of an alignment apparatus according to the present invention.

FIG. 2 is an enlarged sectional view showing a wafer mark according to a second embodiment of the present invention.

FIG. 3 is an enlarged cross-sectional view showing a modified example of the wafer mark of the second embodiment.

FIG. 4 is a configuration diagram showing a main part of a projection exposure apparatus used in these examples.

5 is a configuration diagram showing an alignment optical system 13 in FIG.

FIG. 6 is a diagram for explaining an alignment method by two-beam interference.

FIG. 7 is an enlarged cross-sectional view showing a conventional wafer mark manufacturing process.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 wafer 2 metal film 2b concave part 3 reticle 5 projection optical system 6 wafer holder 7 wafer stage 8 moving mirror 10 reference mark 12 alignment microscope 13 alignment optical system 15 laser light source 17,24 acousto-optic modulator 21 reference grating 22,27 light receiving element 29a Recesses 30a, 30b, 30c Recess group

Claims (3)

[Claims] 1. An alignment mark composed of a convex portion and a concave portion formed in the alignment direction on an object to be inspected, an irradiation optical system for irradiating the alignment mark with light for position detection, and the irradiation of the light. An alignment apparatus that has a light receiving optical system that receives light generated from the alignment mark, and that detects the position of the test object using the light generated from the alignment mark, wherein the alignment of the concave portion of the alignment mark is performed. An alignment apparatus characterized in that a width in a direction is set to be narrower than a length three times a wavelength of the light for position detection. 2. The alignment mark has a pattern composed of a convex portion and a concave portion formed at a first pitch in the alignment direction at a second pitch larger than the first pitch. The alignment apparatus according to claim 1, wherein the alignment apparatus is repeatedly formed. 3. The alignment apparatus according to claim 1, wherein the alignment mark is formed on a metal film deposited on the test object.