JPH06224104A - Alignment device - Google Patents

Abstract

PURPOSE:To lessen the effect of dissymmetrical property of bottom part of a recessed part in an alignment device with which position detection is conducted using the light emitted from an alignment mark by projecting a position detection light on the alignment mark consisting of recessed parts arranged periodically in the direction of alignment. CONSTITUTION:A wafer mark WM is formed by arranging recessed parts 2a of width (d) on a metal film 2 at pitch (p) in X-direction which is the direction of measurement. The width (d) of the recessed parts 2a is set at about the wavelength of the light used for detection of position, and a linear polarized light, which is polarised in Y-direction vertical to X-direction which is the periodical direction of the recessed parts 2a, is formed.

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 disclosed in, for example, Japanese Patent Application Laid-Open Nos. 2-227603 and 2-272305. Has been. As shown in FIG. 8, in the two-beam interference method, a wafer mark WM made of a phase type diffraction grating as an alignment mark is formed in the vicinity of each shot area on the wafer 1 as a photosensitive substrate. X is the measurement direction of the position
Direction, the wafer marks WM are formed at a pitch P in the X direction. The two laser beams LB1 and LB2 are symmetrically applied to the wafer mark WM via the projection optical system. In this case, the + 1st order diffracted light LB11 from the wafer mark WM by the laser beam LB1 and the −1st order diffracted light LB21 from the wafer mark WM by the laser beam LB2 are parallel to each other and reflected vertically upward from the wafer 1, The pitch P and the incident angles of the laser beams LB1 and LB2 are set.

A signal corresponding to the position of the wafer mark WM is obtained by causing the diffracted lights LB11 and LB21 to interfere with each other in the alignment optical system via the projection optical system and photoelectrically converting the interference light. Then, the laser beam LB1
In the case of the heterodyne system in which the frequencies of WB and LB2 are made different, the signal due to the interference becomes a signal of a predetermined beat frequency, and the position of the wafer 1 in the X direction is compared by comparing the phase of this signal with the phase of the reference signal. 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 composed of an amplitude type diffraction grating or the like is formed as an alignment mark 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.

Next, a conventional example of the process of forming the wafer mark WM will be described. FIG. 9 shows the detailed structure of the wafer mark WM. As shown in FIG. 9, when the wafer mark WM is formed, first, the pitch P in the X direction is formed on the surface of the wafer 1.
Thus, the concave portion 1a is formed. After that, a metal film 2 of aluminum or the like is deposited on the surface of the wafer 1 including the concave portion 1a by sputtering, vapor deposition, or the like to form the concave portion 1a.
Corresponding to the concave portions 2a having a pitch d in the X direction and a width d.
Is formed. In FIG. 9, the width d of the recess 2a is expressed narrowly, but in the conventional wafer mark, the width d of the recess 2a is about ½ of the pitch P. As described above, the wafer marks WM are formed from the concave portions 2a arranged at the pitch P in the X direction.
Are formed. However, since it can be said that the middle of the two adjacent concave portions 2a is a convex portion, it can be said that the wafer mark WM is formed of convex portions arranged at the pitch P in the X direction, and further, the wafer mark WM is convex. It can be said that the portions and the concave portions are formed by periodically arranging in the measurement direction.

Further, conventionally, the polarization states of the laser beams LB1 and LB2 with which the wafer mark WM is irradiated are both set to the circular polarization state. FIG. 10 shows a conventional optical system (isolator) for separating the laser beam to the wafer mark WM and the laser beam from the wafer mark WM. In FIG. 10, the light is emitted from a synthetic prism (not shown). Two laser beams L
B1 and LB2 enter the polarization beam splitter 51. The polarization states of the laser beams LB1 and LB2 incident on the polarization beam splitter 51 are linearly polarized light in a direction parallel to the paper surface of FIG. 10, that is, P polarized light with respect to the joint surface of the polarizing beam splitter 51. Therefore, the laser beams LB1 and LB2 directly pass through the joint surface of the polarization beam splitter 51.

P passing through the polarization beam splitter 51
The polarized laser beams LB1 and LB2 are transmitted by the quarter wavelength plate 5
After passing through 2, the light becomes circularly polarized light, is reflected by the mirror 53, and is then irradiated onto the wafer mark WM via, for example, a projection optical system. At this time, diffracted lights diffracted vertically upward from the wafer mark WM (diffracted lights LB11 and LB21 in FIG. 8).
Is circularly polarized light in the reverse direction, and the diffracted light of the circularly polarized light in the reverse direction passes through the projection optical system and the mirror 53, and the quarter wavelength plate 52
Incident on. Therefore, the diffracted light returned in this way is 1 /
It is converted to S-polarized light by passing through the four-wave plate 52,
The S-polarized diffracted light was totally reflected by the joint surface of the polarization beam splitter 51 and was incident on the light receiving element 54. In this conventional example, the light amount loss was suppressed to the minimum by combining the polarization beam splitter 51 and the quarter-wave plate 52.

[0008]

In the prior art as described above, as shown in FIG. 9, the bottom of each recess 2a of the wafer mark is not generally parallel to the measurement direction (X direction), but the measurement direction is not parallel. Is formed asymmetrically with respect to. Therefore, the light reflected from the bottom of the recess 2a of the wafer mark has a predetermined directionality, and when the reflected light is detected to detect the position of the wafer mark, it is measured as the actual position of the wafer mark. There is an inconvenience that an error occurs between the position and the position.

In order to eliminate such an error in the position detection result, the applicant of the present application has disclosed that the wafer mark is a metal film in Japanese Patent Laid-Open No. 4-284370, and the width of the recesses 2a in the measuring direction is defined by the laser beam wavelength. It has been proposed to make it about 3 times or less, especially smaller than about the wavelength. By the way, in the conventional example, the laser beams LB1, L radiated on the wafer mark
B2 is in a circularly polarized state, and the electric vectors (electric field vectors) of the laser beams LB1 and LB2 with respect to the measurement direction of the wafer mark include a polarization component parallel to the measurement direction and a polarization component perpendicular to the measurement direction. Has been. In this case, the polarization component in the measurement direction of the wafer mark, that is, in the direction perpendicular to the periodic direction (X direction) of the concave portions 2a is absorbed and attenuated before reaching the bottom of the concave portions 2a, but is parallel to the periodic direction of the concave portions 2a. The polarized component has no attenuation and reaches the bottom of the recess 2a and is reflected. Therefore, the laser beam LB
When 1 and LB2 are irradiated on the wafer mark in a circularly polarized state, even if the width of the recess 2a in the measurement direction is set to about the wavelength of the laser beam, a measurement error due to the asymmetry of the recess 2a may occur. was there.

In view of the above point, the present invention irradiates a position detecting light to an alignment mark composed of concave portions periodically arranged in the alignment direction, and detects the position by using the light generated from the alignment mark. It is an object of the present invention to reduce the influence of the asymmetry of the bottom of the recess in the alignment apparatus for performing the above.

[0011]

An alignment apparatus according to the present invention comprises, for example, as shown in FIGS. 1 and 2, concave portions (2a) which are periodically arranged in the alignment direction on an object (1) to be inspected. An alignment mark (WM), an irradiation optical system (12) that irradiates the alignment mark (WM) with light for position detection, and receives light generated from the alignment mark (WM) by the irradiation of the light for position detection. It has a light receiving optical system (12) and has an alignment mark (W
In the alignment apparatus that detects the position of the object (1) using the light generated from M), the polarization state of the light for position detection on the alignment mark (WM) is determined by the period of the alignment mark (WM). It is a linearly polarized light whose direction is determined by the direction.

In this case, when the alignment mark (WM) is formed of metal, the direction of the electric vector (electric field vector) on the alignment mark (WM) of the light for position detection is changed to the alignment mark (WM). ) It is desirable to make it perpendicular to the cycle direction. However, when the alignment mark (WM) is made of metal, the alignment mark (WM) of the light for position detection
The direction of the electric vector above is aligned with the alignment mark (W
In some cases, the direction M) may be used. Further, it is desirable that the width of the concave portion (2a) of the alignment mark (WM) in the periodic direction is set to about 3 times the wavelength of the light for position detection or less. Further, it is desirable to provide polarization state changing means for setting the direction of the linearly polarized light of the light for position detection on the alignment mark (WM) to the direction determined by the periodic direction of the alignment mark (WM).

[0013]

According to the present invention, the position detecting light applied to the alignment mark (WM) is linearly polarized light. The alignment mark (W
By setting the direction to be determined according to the periodic direction of M), the influence of the asymmetry of the bottom of the recess (2a) is reduced, and the position detection error is reduced.

When the alignment mark (WM) is made of metal, the direction of the electric vector of the position detecting light on the alignment mark (WM) is
When the alignment mark (WM) is perpendicular to the periodic direction, the following action is obtained. That is, when the direction of the electric vector of the light for position detection is parallel to the inner surface of the recess (2a) (perpendicular to the period direction), the electrons on the inner surface of the recess (2a) also vibrate in response to the vibration of the electric vector. Therefore, the light for position detection is exponentially attenuated when propagating in the recess (2a). Therefore, even if the bottom of the recess (2a) is asymmetric, the influence of the asymmetry becomes very small, and the error in position detection becomes very small.

On the other hand, if the direction of the electric vector of the light for position detection is perpendicular to the inner surface of the concave portion (2a) (parallel to the period direction), the concave portion (2a) will be generated even if the electric vector vibrates. The electrons on the inner surface of No. 2 do not vibrate, and the light for position detection reaches the bottom of the recess (2a) without being attenuated even when propagating through the recess (2a). Therefore, it is affected by the asymmetry of the bottom of the recess (2a). However, when the width of the concave portion of the alignment mark is about three times the wavelength or less, particularly narrower than the wavelength, the bottom portion of the concave portion (2a) of the alignment mark comes to be formed symmetrically (it can be regarded as being formed). Sometimes). At this time, since the light in the polarization state parallel to the periodic direction is not attenuated, the intensity of the diffracted light is larger than that of the polarized light perpendicular to the periodic direction, and the detection accuracy is improved. In addition, depending on the alignment mark, it may be possible to make the concave portion (2
It may be advantageous in terms of S / N ratio as compared with the case where the direction along the groove in (a) and the polarization direction are matched.

There is a problem that the intensity of the diffracted light is reduced when the width of the concave portion (2a) is about 3 times the wavelength or less, particularly the wavelength or less. When the width of the recess is less than the wavelength, the bottom is formed symmetrically (may be considered to be formed).
Therefore, it is not affected by the polarization direction of the incident light. In this case, it is easier to obtain the amount of light when the light is made incident in parallel to the alignment mark (WM) than when it is made perpendicular to the periodic direction.
The / N ratio is improved. In addition, the width of the recess (2a) is set to 3
If it is about double or less, the accuracy is further improved. In addition, the polarization state variable for setting the direction of the linearly polarized light of the position detecting light on the alignment mark (WM) to the direction determined by the periodic direction of the alignment mark (WM) and the material of the alignment mark (WM). When the means is provided, it is possible to reduce the amount of light for position detection reaching the bottom of the concave portion (2a) of the alignment mark (WM) even when the period direction and material of the alignment mark (WM) change. it can.

[0017]

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. FIG. 2 shows a main part of a projection exposure apparatus provided with the alignment apparatus of this example. In FIG. 2, the reticle R is held on the reticle holder 3 and the circuit pattern formed on the reticle R is omitted from the illumination. Illuminate with exposure light IL from the optical system. The circuit pattern is, for example, 5 by the projection optical system 4.
It is reduced by a factor of 1 and transferred onto the wafer 1 coated with a photosensitive material such as resist. Wafer marks WM as alignment marks are formed in the vicinity of each shot area of the wafer 1 at a predetermined pitch in the measurement direction.

FIG. 1 shows a wafer mark WM formed in the vicinity of each shot area of the wafer 1. In FIG. 1, a metal film 2 is formed on a substrate (for example, a silicon substrate) of the wafer.
Are deposited, and recesses 2a having a width d are formed on the metal film 2 at a pitch P in the X direction. Therefore, the cycle direction of the wafer mark WM is the X direction, and this X direction is also the measurement direction (positioning direction). Further, in this example, the width d of the concave portion 2a in the X direction is 3 of the wavelength λ of the light for position detection.
It is set to about twice or less, particularly smaller than the wavelength λ. Thereby, the light for position detection is easily attenuated in the recess 2a.

Returning to FIG. 2, the wafer 1 is placed on the wafer holder 5
It is mounted on the wafer stage 6 via the. The wafer stage 6 is composed of an XY stage for positioning the wafer 1 in a two-dimensional plane perpendicular to the optical axis of the projection optical system 4, a Z stage for positioning the wafer 1 in the optical axis direction, and the like. A movable mirror 7 is installed on the wafer stage 6, 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 6. Further, by operating the wafer stage 6 via the driving device 9, for example, step and
Positioning of each shot area of the wafer 1 is performed by the repeat method.

The pattern already formed in each shot area on the wafer 1 and the pattern of the reticle R usually need to 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 6, and an alignment microscope 11 provided with a light source of illumination light having the same wavelength as the exposure light IL is arranged on the upper side of the reticle R. To do. Further, an alignment optical system 12 is arranged on the side of the projection optical system 4, and the position of the wafer mark WM on the wafer 1 is detected by the alignment optical system 12 (details are disclosed in JP-A-2-272305, etc.). Exist).

Two laser beams LB1 and LB2 are emitted from the alignment optical system 12, and these laser beams are irradiated onto the wafer 1 through the folding mirror 13 and the projection optical system 4. In this case, the wafer mark WM
The incident angles of the two laser beams and the pitch P of the wafer marks WM are set so that the diffracted light LB0 is generated substantially vertically above the wafer 1. 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 12 via the projection optical system 4 and the folding mirror 13.

In this case, the relative position between these two laser beams LB1 and LB2 irradiated on the wafer 1 and the reticle R is determined by using the alignment microscope 11 with the reference mark 10 provided on the wafer stage 6. It can be known by observing through the reticle R and detecting the same reference mark 10 by the alignment optical system 12.

FIG. 3 shows the alignment optical system 12 shown in FIG.
In FIG. 3, the laser beam LB emitted from the laser light source 14 is the two-division prism 1
The laser beam is divided into two laser beams LB1 and LB2. One laser beam LB1 is incident on the combining prism 20 via the first acousto-optic modulator (AOM) 16 and the rectangular prism 17, and the other laser beam L
B2 enters the combining prism 20 through the right-angle prism 18 and the second acousto-optic modulator (AOM) 19. The acousto-optic modulators 16 and 19 have different frequencies f
Acousto-optic modulators 16 and 19 driven by ₁ and f _2.
Of the light diffracted by, the + _1st order diffracted light becomes higher in frequency by the driving frequencies f ₁ and f ₂ . Two + 1st order diffracted lights having different frequencies (also called laser beams LB1 and LB2) enter the combining prism 20.

The laser beam LB1 transmitted through the synthesizing prism 20 and the laser beam LB2 reflected by the synthesizing prism 20 are focused on the reference grating 22 by the objective lens 21, and the reference grating in the light transmitted through the reference grating 22 is used. Two diffracted lights that are diffracted substantially vertically downward with respect to 22 are mixed and detected by the light receiving element 23. On the other hand, the laser beam LB1 reflected by the synthesizing prism 20 and the laser beam LB2 transmitted by the synthesizing prism 20 pass through the polarization beam splitter 24 and then the objective lens 25.
Are focused by and are incident on the folding mirror 13 shown in FIG. At this time, the laser beams LB1 and LB2 incident on the polarization beam splitter 24 are linearly polarized light, and the polarization direction (direction of the electric vector) is P polarization with respect to the polarization beam splitter 24 (the incident surface with respect to the joint surface of the polarization beam splitter 24). Parallel direction) is set.

Although not shown in FIG. 3, FIG.
As shown in FIG. 4, a Faraday rotator 28 for rotating the polarization direction between the polarization beam splitter 24 of FIG.
The half wave plate 29 is arranged. In FIG. 4, the P-polarized laser beam LB that has passed through the polarization beam splitter 24.
1, 1 and LB2 enter the Faraday rotator 28. The laser beams LB1 and LB2 emitted from the Faraday rotator 28 are linearly polarized light polarized in a direction rotated by 45 ° from the polarization direction at the time of incidence. When the laser beams LB1 and LB2 further pass through the half-wave plate 29, the polarization directions of the laser beams LB1 and LB2 are further rotated by + 45 ° or −45 °.

The laser beam whose polarization state is set in this way is irradiated onto the wafer mark WM of FIG. 1 through the objective lens 25, the bending mirror 13 and the projection optical system 4. In this example, by adjusting the rotation angle of the half-wave plate 29, the polarization directions of the laser beams LB1 and LB2 irradiated on the wafer mark WM in FIG. 1 are the Y direction perpendicular to the periodic direction of the concave portions 2a. , Can be set in the X direction, which is the periodic direction.

The diffracted light LB0 diffracted substantially vertically upward from the wafer mark WM passes through the projection optical system 4, the bending mirror 13 and the objective lens 25 while maintaining the linearly polarized light, and then reaches the ½ wavelength plate 29 of FIG. Incident. 1/2 wave plate 2
The diffracted light LB0 transmitted through 9 has a polarization direction of 4
It becomes a linearly polarized beam rotated by 5 ° and enters the Faraday rotator 28. The polarization direction of the diffracted light LB0 transmitted through the Faraday rotator 28 is further rotated by 45 °, which means that the polarization direction is rotated by 90 ° before returning to the polarization beam splitter 24. Therefore, the diffracted light L
B0 is incident on the polarization beam splitter 24 as an S-polarized laser beam and is all reflected by the joint surface of the polarization beam splitter 24. Polarizing beam splitter 24
The diffracted light LB0 reflected by is incident on the light receiving element 26.

Returning to FIG. 3, the reference signal from the light receiving element 23 and the alignment signal from the light receiving element 26 are supplied to the alignment processing unit 27, respectively. Light receiving element 26
The alignment signal obtained by photoelectrically converting the diffracted light LB0 by means of 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 16 and 19, respectively. The reference signal output from the element 23 is also a sine wave having a frequency of Δf. 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 positional deviation detected by the alignment processing unit 27 is sent to a main computer that controls the operation of the entire apparatus.

Next, the operation of this example will be described with reference to FIG. FIG. 1 shows a wafer mark WM of this example. In FIG. 1, the laser beam irradiated onto the wafer mark WM is conventionally irradiated with circularly polarized light as shown by a circular arrow P3. On the other hand, the polarization state of the laser beam in this example is linearly polarized light parallel to the periodic direction (X direction) of the concave portions 2a of the wafer mark WM as shown by an arrow P1, or the periodic direction of the concave portions 2a as shown by an arrow P2. It is a linearly polarized light perpendicular to.

When the wafer mark WM is composed of the concave portions 2a which are periodically formed in the X direction on the metal film 2 as in the present example, the linearly polarized laser beam irradiated to the wafer mark WM. The polarization direction (direction of electric vector) is set to the Y direction perpendicular to the X direction as indicated by arrow P2. As a result, when the laser beam passes through the recess 2a, the electrons on the side surface of the recess 2a also vibrate in the Y direction as the electric vector vibrates in the Y direction.
The laser beam decays exponentially and sharply. Therefore, the laser beam is almost attenuated before reaching the bottom of the recess 2a, and the reflected light from the bottom is almost nonexistent. Even when the bottom of the recess 2a is asymmetric in the X direction, the wafer mark WM The error in the position detection result is extremely small.

When the wafer mark WM is formed by forming the concave portions 2a in the Y direction at a predetermined pitch, the polarization direction of the laser beam with which the wafer mark WM is irradiated is in the X direction perpendicular to the Y direction. Set. This allows
The error in the position detection result in the Y direction is also small. However, when the wafer mark WM is one in which concave portions are periodically formed in the X direction on a non-conductive object (insulator or the like), the polarization direction of the laser beam is not necessarily the Y direction perpendicular to the X direction. Need not be set to.

Next, in the present embodiment, the wafer mark WM
A method of changing the polarization direction of the laser beam incident on the laser beam will be described. In this case, as described above, as shown in FIG. 4, the Faraday rotator 28 can obtain a linearly polarized beam rotated by 45 °. Then, in order to make the linearly polarized beam rotated by 45 ° polarized in a direction perpendicular to, or parallel to, the periodic direction of the concave portions 2a of the wafer mark WM and to be incident, a ½ wavelength plate is used as a first method. 29 may be rotated. 1/2
By adjusting the rotation angle of the wave plate 29, it is possible to irradiate the wafer mark WM with a laser beam linearly polarized in a desired direction.

A second method for setting the direction of linearly polarized light of the laser beam with which the wafer mark WM is irradiated is to use a Pockels element instead of the half-wave plate 29 of FIG. That is. The Pockels element is an element that can change the angle of linearly polarized light of the emitted laser beam depending on the magnitude of the applied voltage. Therefore, by applying a voltage such as ± 45 ° rotation to the Pockels element with respect to the incident light of linearly polarized light rotated by 45 °, the incident light is perpendicular or parallel to the periodic direction of the concave portions 2a of the wafer mark WM. A linearly polarized beam polarized in the direction can be obtained.

Next, the alignment optical system 12 of the present embodiment.
A case where a laser diode is used as the laser light source 14 of FIG. 3 will be described with reference to FIG. FIG. 5 shows a laser diode 30, in which the laser beam output from the laser diode 30 has a greater spread in the direction of the electric field of the laser diode 30 than in the direction along the joint surface. Therefore,
The cross-sectional shape of the laser beam emitted from the laser diode 30 is elliptical, and the laser beam is linearly polarized in the direction in which the width of the elliptical cross section is short.

Generally, the cross-sectional shape of the laser beam with which the wafer mark WM of FIG. 1 is irradiated is a parallelogram,
This sectional shape is formed by a field slit placed at a position conjugate with the wafer. Therefore, when the laser diode 30 is used as the light source, the laser diode 3
By adjusting the longitudinal direction of the elliptical cross section of the laser beam emitted from 0 to match the longitudinal direction of the field slit, the loss of the light quantity is reduced. Further, the laser beam irradiated onto the wafer mark WM in FIG. 1 is set such that the width of the cross section is narrow in the Y direction which is generally perpendicular to the periodic direction of the recesses 2a, so the laser diode 30 in FIG. 5 is used. In this case, the concave portion 2 of the wafer mark WM
There is an advantage that it is easy to obtain a linearly polarized beam polarized in the Y direction perpendicular to the periodic direction of a.

Next, a second embodiment of the present invention will be described with reference to FIGS. 6 and 7. FIG. 6 shows a wafer mark WM formed on a wafer in the present embodiment. As shown in FIG. 6, a metal film 2 is deposited on the wafer and arranged on the metal film 2 at a pitch P1 in the X direction. The recesses 2a are formed, and the recesses 2b are arranged at a pitch P2 (normally equal to the pitch P1) in the Y direction. The width of the recess 2a in the X direction is d1, the width of the recess 2b in the Y direction is d2, and the widths d1 and d2 are each 3 times the wavelength of the laser beam for position detection.
It is set narrower than about twice. From the two-dimensional diffraction grating composed of these recesses 2a and 2b, the wafer mark WM
Is configured. In this example, with respect to the wafer mark WM, two laser beams for position measurement in the X direction and Y laser beam are used.
By mixing and irradiating with two laser beams for position measurement in the direction, the X direction and the Y direction of the wafer mark WM can be obtained.
Directional position detection is performed simultaneously.

FIG. 7 shows a main part of the alignment optical system of this example. In FIG. 7, two parallel laser beams emitted from an optical system (not shown) (for example, the combining prism 20 of FIG. 3) are emitted. LB1 and LB2 are incident on the polarization beam splitter 31 whose joint surface is perpendicular to the paper surface of FIG.
The laser beams LB1 and LB2 are linearly polarized in a direction inclined by 45 ° from the direction perpendicular to the paper surface of FIG. 7, and the S-polarized component (polarization component perpendicular to the paper surface of FIG. 7) of these laser beams is a polarized beam. A laser beam LB1 for position measurement in the X direction, which is reflected by the joint surface of the splitter 31
It is taken out as X, LB2X. On the other hand, the P-polarized component (polarized component parallel to the paper surface of FIG. 7) of these laser beams passes through the joint surface of the polarization beam splitter 31, and the laser beams LB1Y, L for position measurement in the Y direction are obtained.
It is taken out as B2Y.

Laser beam LB1 for position measurement in the Y direction
Y and LB2Y are a mirror 39a and a quarter-wave plate 33b.
And enters the polarization beam splitter 34b in a circularly polarized state. The S-polarized light component of the laser beams LB1Y and LB2Y is reflected by the polarization beam splitter 34b, and enters the light receiving element 23b through an unillustrated objective lens and reference grating. On the other hand, the P-polarized component passes through the polarization beam splitter 34b, further passes through the polarization beam splitter 35b, and then passes through the Faraday rotator 36b and the Pockels element 37b. The laser beams LB1Y and LB2Y whose polarization direction on the wafer mark WM is adjusted to the optimum direction by the Faraday rotator 36b and the Pockels element 37b are reflected by the mirror 3
The light is reflected by 9b and is incident on the separating / combining mirror 38 including, for example, a polarization beam splitter.

Further, the laser beams LB1X and LB2X for measuring the position in the X direction are rotated by 90 ° by the dove prism 32 so that the laser beams LB1X and LB2X are irradiated on the wafer mark WM while intersecting in the direction orthogonal to the Y direction. To be done. After that, the laser beams LB1X and LB2X emit the quarter-wave plate 33a.
And enters the polarization beam splitter 34a in the state of circularly polarized light. The S-polarized component of those laser beams is
The light is reflected by the polarization beam splitter 34a, and enters the light receiving element 23a through an unillustrated objective lens and reference grating. On the other hand, the P polarization component is the polarization beam splitter 34a.
After passing through the polarization beam splitter 35a, the light passes through the Faraday rotator 36a and the Pockels element 37a. These Faraday rotators 36a
The laser beams LB1X and LB2X whose polarization directions on the wafer mark WM are adjusted to the optimum directions by the Pockels element 37a are incident on the separating / combining mirror 38.

Four laser beams LB1X, LB2X and LB1Y, LB2 reflected by the separating / combining mirror 38.
Y is irradiated onto the wafer mark WM in FIG. 6 via the objective lens 25, the folding mirror 40 and the projection optical system 4.
At this time, the optical axes of the laser beams LB1X and LB2X are orthogonal to the optical axes of the laser beams LB1Y and LB2Y, and the polarization directions are also orthogonal to each other. Wafer mark W
The laser beams LB1X and LB2X and the laser beams LB1Y and LB2Y with which M is irradiated, respectively, diffracted light LB0X whose polarization directions are orthogonal to each other from the wafer mark WM.
And LB0Y are generated, and these diffracted lights enter the separating / combining mirror 38 via the projection optical system 4, the folding mirror 40 and the objective lens 25 of FIG.

Y reflected by the separating / combining mirror 38
The diffracted light LB0Y for measuring the position in the direction is reflected by the mirror 38b,
The Pockels element 37b, the Faraday rotator 36b, and the polarization beam splitter 35b, and then the light receiving element 26b.
Incident on. The Y direction reference signal output from the light receiving element 23b and the Y direction alignment signal output from the light receiving element 26b are supplied to the alignment processing unit 27, respectively. Further, the diffracted light LB0X for position measurement in the X direction transmitted through the separating / combining mirror 38 enters the light receiving element 26a via the Pockels element 37a, the Faraday rotator 36a, and the polarization beam splitter 35a. The reference signal for the X direction output from the light receiving element 23a and the alignment signal for the X direction output from the light receiving element 26a are respectively supplied to the alignment processing unit 27. The alignment processing unit 27 detects the position of the wafer mark WM in FIG. 6 in the X direction and the Y direction from the supplied two sets of signals.

In this example, since the wafer mark WM of FIG. 6 is formed on the metal film 2, a laser beam for measuring the position in the X direction (laser beams LB1X, L of FIG. 7) is used.
B2X) is irradiated onto the wafer mark WM as linearly polarized light having a polarization direction of Y direction. Further, a laser beam for measuring the position in the Y direction (the laser beams LB1Y, L in FIG.
B2Y) is irradiated onto the wafer mark WM as linearly polarized light whose polarization direction is the X direction. As a result, the laser beam for position measurement is applied to the concave portions 2a and 2b of the wafer mark WM.
It decays rapidly inside and rarely reaches the bottom.
Therefore, even if the bottoms of the recesses 2a and 2b are asymmetric in the measurement direction, the error in the position detection result is small.

As in the alignment optical system of FIG. 7, in order to detect the positions in the X direction and the Y direction at the same time, the following method can be considered. First, in the first method, the drive frequencies f ₁ and f for the acousto-optic modulator in the X and Y directions are set.
The frequency difference Δf of ₂ is changed, and the alignment signal of the frequency difference Δf _x for the X direction and the alignment signal of the frequency difference Δf _y for the Y direction are received by one light receiving element. Then, the photoelectric conversion signal output from the light receiving element is separated into a component of frequency Δf _{x and} a component of frequency Δf _y by a frequency filter circuit, so that position measurement in the X direction and the Y direction can be performed simultaneously. You can

Further, in the second method, the pitch P1 in the X direction and the pitch P2 in the Y direction of the wafer mark WM shown in FIG. 6 are used.
(Normally equal to P1) is halved. When the pitch P1 (P2) of the wafer mark WM is halved, the diffraction angle of the laser beam with which the wafer mark WM is irradiated becomes 2
Therefore, the 0th-order diffracted light of the other laser beam LB2 and the + 1st-order diffracted light of the laser beam LB1 are emitted in an overlapping manner, for example, in the incident direction of the one laser beam LB1. Similarly, in the incident direction of the laser beam LB2,
0th-order diffracted light of laser beam LB1 and laser beam L
The -1st-order diffracted light of B2 overlaps and is emitted. Therefore,
Since no diffracted light is generated vertically above the wafer mark WM, there is no mixing of the diffracted light for position detection in the X direction and the diffracted light for position detection in the Y direction.

In the above embodiment, the width of the concave portion of the alignment mark is about 3 times the alignment wavelength or less.
If the polarized light of the alignment light (linearly polarized light) matches the direction perpendicular to the periodic direction, it is possible to reduce the influence of the asymmetry of the recess even if the width of the recess exceeds three times the wavelength. . Further, in FIG. 1, the alignment mark (WM)
The width d of the concave portion 2a is, for example, 3 times the wavelength of the alignment light.
When it is about twice or less, particularly narrower than the wavelength, it may be considered that the bottom of the recess 2a is formed symmetrically. In such a case, the polarization directions of the laser beams LB1 and LB2 with which the alignment mark WM is irradiated may be matched with the X direction which is the periodic direction. At this time, since the bottom of the recess 2a is symmetrical, the position can be accurately detected even using the diffracted light from the recess 2a. Therefore, Y perpendicular to the cycle direction
Compared with the case where the direction and the polarization direction are the same, the diffracted light is not attenuated in the concave portion 2a, so that the diffracted light intensity increases.
That is, there is an advantage that the S / N ratio is improved.

Further, the width of the concave portion of the alignment mark,
And / or the polarization direction of the alignment light may be appropriately changed depending on the degree of asymmetry of the bottom portion. That is, a first mode for matching the polarization direction with the direction perpendicular to the cycle direction and a second mode for matching the polarization direction with the cycle direction are prepared (for example, set in a control unit that integrally controls the entire device), Depending on the width of the concave portion of the alignment mark and / or the degree of asymmetry of its bottom, one of the first mode and the second mode is selected, and the polarization direction is adjusted (or confirmed) according to the selected mode. It is better to perform mark detection. For example, the first mode is selected when the width of the recess of the alignment mark on the wafer exceeds the alignment wavelength, and the second mode is selected when the width of the recess is narrower than the wavelength. At this time, even if the width of the concave portion is narrower than the wavelength, if asymmetry exists in the lower portion, it is desirable to make the polarization direction coincide with the direction perpendicular to the periodic direction.

The operator may select a mode according to the mark information (width of the recess, whether the bottom is asymmetric, etc.) and then set the mode in the apparatus via the keyboard, or the apparatus itself. May be automatically selected according to the information (width of recess, whether bottom is asymmetric, wafer name, etc.) written on the barcode of the wafer or wafer cassette. Further, for example, Japanese Patent Laid-Open No. 2-541
No. 03, Japanese Patent Laid-Open No. 4-65603 discloses an alignment sensor using an alignment sensor that employs an image processing method, and the asymmetry of the concave portion and the width of the concave portion are obtained based on the detection result. The mode may be selected accordingly. The alignment sensor disclosed in the above publication irradiates an alignment mark with illumination light (for example, white light) having a predetermined wavelength width, and an image of the mark on an index plate arranged in a plane conjugate with the wafer. Image on. Further, the image of the alignment mark and the image of the index mark on the index plate are formed on the light receiving surface of the image pickup element by the relay lens system, and the relative positional deviation amount of both marks is detected. Further, in the above embodiment, 2
Although the description has been given by taking the alignment sensor adopting the light flux interference method as an example, the present invention may be applied to the alignment sensor adopting the image processing method as described above, and the same effect as in the above-described embodiment is obtained. be able to.

Although the alignment system of the above-described embodiment employs the heterodyne method, for example, the homodyne method or two light beams having different polarization directions are applied to the alignment mark and the diffracted light from the mark is detected. The same effect can be obtained by applying the present invention to a method of receiving light by interfering via photons or the like. Further, not only the method of detecting the phase of the diffracted light from the alignment mark, for example, when performing position detection by a so-called laser step alignment method, or for an alignment apparatus of an image processing method using an image sensor, The present invention can be applied to obtain exactly the same effect. For example, in the case of the laser step alignment method, the alignment mark and the slit-shaped probe light are relatively scanned to change the intensity of the diffracted light emitted from the alignment mark in a predetermined direction by the probe light. The position is detected by the detection.

Further, in an actual process, for example, a photoresist layer having a predetermined film thickness (about 1 μm) is formed on the metal film 2 of FIG. 1, but the present invention relates to the presence or absence of such a photoresist layer. However, it is effective against the asymmetry of the bottom of the concave portion of the alignment mark. As described above, the present invention is not limited to the above-described embodiments, and various configurations can be adopted without departing from the gist of the present invention.

[0050]

According to the present invention, the polarization state of the position detecting light with which the alignment mark is irradiated is linearly polarized light. By setting the polarization direction of the light for position detection to the direction in which the light for position detection is attenuated in the recess of the alignment mark, the influence of the asymmetry of the bottom of the recess can be reduced. . Therefore, there is an advantage that an error in detecting the position of the alignment mark can be reduced.

When the alignment mark is made of metal and the direction of the electric vector of the light for position detection on the alignment mark is made perpendicular to the periodic direction of the alignment mark, the alignment mark is When the width of the concave portion in the periodic direction is narrowed to, for example, about the wavelength of the position detecting light, the position detecting light is rapidly attenuated in the concave portion. Therefore, the influence of the asymmetry of the bottom of the recess is reduced. However, when the bottom of the concave portion of the alignment mark is almost symmetrical, the direction of the electric vector of the position detection light on the alignment mark is set to the periodic direction, so that the S / N of the detection signal is changed.
The ratio can be increased.

In particular, when the width of the concave portion of the alignment mark is about 3 times the wavelength of the position detection light or less, particularly the wavelength or less, the bottom of the concave portion is formed symmetrically. At this time, if the direction of the electric vector of the light for position detection on the alignment mark is made parallel to the periodic direction of the alignment mark, there is no attenuation in the recesses, so the intensity of the diffracted light is maintained.
Further, when the polarization state changing means for setting the direction of linearly polarized light on the alignment mark of the light for position detection to the direction determined by the cycle direction of the alignment mark is provided, Even when the material changes, the influence of the asymmetry of the bottom of the concave portion of the alignment mark can be reduced.

[Brief description of drawings]

FIG. 1 is an enlarged plan view showing a wafer mark which is an object of position detection in a first embodiment of an alignment apparatus according to the present invention.

FIG. 2 is a configuration diagram showing a main part of a projection exposure apparatus equipped with the alignment apparatus of the first embodiment.

FIG. 3 is a configuration diagram showing a part of the alignment optical system in FIG. 2 with omission.

4 is a configuration diagram showing an optical system for separating a laser beam to a wafer mark and a laser beam from a wafer mark in the alignment optical system of FIG.

FIG. 5 is a perspective view showing a laser diode as another example of the laser light source.

FIG. 6 is an enlarged plan view showing a wafer mark which is an object of position detection in the second embodiment of the present invention.

FIG. 7 is a configuration diagram showing a main part of an alignment optical system of a second example.

FIG. 8 is a side view for explaining the principle of conventional wafer mark position detection.

FIG. 9 is a cross-sectional view for explaining a conventional wafer mark manufacturing process.

FIG. 10 is a configuration diagram showing an optical system for separating a laser beam to a wafer mark and a laser beam from a wafer mark in a conventional alignment optical system.

[Explanation of symbols]

 R Reticle 1 Wafer 2 Metal film 2a Recess 4 Projection optical system 10 Reference mark 11 Alignment microscope 12 Alignment optical system 14 Laser light source 16, 19 Acousto-optic modulator 22 Reference grating 23, 26 Light receiving element 24 Polarizing beam splitter 27 Alignment processing unit 28 Faraday rotator 29 1/2 wave plate 32 Dove prism 36a, 36b Faraday rotator 37a, 37b Pockels element

Claims (5)

[Claims] 1. An alignment mark composed of concave portions periodically arrayed in an alignment direction on an object to be inspected, an irradiation optical system for irradiating the alignment mark with light for position detection, and the light for position detection. And a light receiving optical system for receiving light generated from the alignment mark by irradiation of
In an alignment apparatus that detects the position of the object to be measured using light generated from the alignment mark, the polarization state of the light for position detection on the alignment mark is determined by the periodic direction of the alignment mark. An alignment device characterized by using linearly polarized light. 2. When the alignment mark is made of metal, the direction of the electric vector of the light for position detection on the alignment mark is perpendicular to the periodic direction of the alignment mark. The alignment apparatus according to claim 1. 3. When the alignment mark is made of metal, the direction of the electric vector of the light for position detection on the alignment mark is set to the periodic direction of the alignment mark. The alignment apparatus according to claim 1. 4. The alignment mark is characterized in that the width of the concave portion in the periodic direction is set to about 3 times or less the wavelength of the light for position detection. The alignment device described. 5. A polarization state changing means is provided for setting a direction of linearly polarized light of the light for position detection on the alignment mark to a direction determined by a periodic direction of the alignment mark. The alignment apparatus according to claim 1.