JPH01240803A - Detecting apparatus for pattern position - Google Patents

Abstract

PURPOSE:To enable the correction of chromatic aberration without using a chromatic aberration correction lens and thereby to improve the precision in detection of a mark, by disposing a linear-type Fresnel zone plate (FZP) on a reticle. CONSTITUTION:An alignment mark 12 on a wafer is illuminated through a projection lens 3 by a light 6 of a wavelength being different from an exposure wavelength. An image of the alignment mark 12 is projected reversely by the projection lens 3. FZP 2 is disposed on a reticle 1: this FZP 2 functions as a chromatic aberration correction lens, enabling the formation of an image I of the alignment mark 12 in an arbitrary position. The FZP 2 is of a linear type and it is so disposed as to intersect perpendicularly the alignment mark 12 on the wafer.

Description

[Detailed Description of the Invention] [Field of Industrial Application] The present invention relates to an apparatus for exposing and transferring a circuit pattern such as an LSI on a photomask to a photoresist on a semiconductor wafer, and is particularly suitable for performing highly accurate alignment. The present invention relates to a pattern position detection device.

[Conventional technology]

As an example of a conventional apparatus, a reduction projection exposure apparatus described in Japanese Patent Laid-Open No. 55-162227 will be described. A schematic diagram of the conventional device is shown in FIG.

In a reduction projection exposure apparatus, a circuit pattern 14 on a reticle 1 to be newly formed is usually formed on a circuit pattern 15 on a wafer 4 formed in a previous process using a condensing lens 1 for exposure.
8 and the reduction lens 3 perform overlapping exposure. usually,
This overlapping exposure is sequentially repeated for several reticles to form a desired circuit pattern on the wafer. At this time, the alignment of the two circuit patterns to be superimposed requires an accuracy of 1/4 to 115 of the pattern dimension to be formed. In the conventional reduction projection exposure apparatus shown in FIG. 6, such alignment is performed by detecting the position of the alignment mark 12 on the wafer and relatively moving the reticle having a newly formed pattern to match the wafer. It is done by doing. That is, the alignment mark 12 of the wafer 4 is locally illuminated by the light source 5, and the reflected light is imaged by the projection lens (reducing lens) 3 onto the reference mark 2 of the reticle 1 by back projection.

The alignment mark imaged with this reference mark is enlarged and imaged by an enlarging optical system 10. The enlarged image is scanned by a slit, and the brightness and darkness of the light passing through the slit is photoelectrically converted by a photomultiplier to obtain a detection signal. The relative position of the reference mark and the alignment mark of the wafer is determined from the detection signal, and the reticle is moved according to the amount of deviation of the relative position to perform alignment.

[Problem to be solved by the invention]

In the above-mentioned prior art, when detecting a mark position via the projection lens 3, since the projection lens 3 has aberrations corrected with respect to the exposure wavelength, monochromatic light having only the exposure wavelength is normally used as the detection light wavelength.

Therefore, there is a drawback that equal five interference fringes are generated in the photoresist 19 applied on the wafer, leading to a decrease in the contrast of the detection signal depending on the thickness of the photoresist, and pattern alignment accuracy is degraded. In addition, reflection of the exposure wavelength from the wafer hinders the formation of fine patterns.
An antireflection film may be provided on the wafer, or a light absorber at the exposure wavelength may be mixed into the photoresist. At times like this,
Detection using the exposure wavelength does not provide sufficient reflected light intensity from the wafer for detection. In order to eliminate the above drawbacks, it is necessary to detect marks using a wavelength different from the exposure wavelength. At this time, a chromatic aberration correcting lens is usually provided between the projection lens 3 and the reticle 2 to enable detection of the relative position between the reticle and the wafer.

However, this has the disadvantage that mark position detection errors occur due to positional fluctuations of the chromatic aberration correcting lens, and highly accurate mark detection cannot be performed.

An object of the present invention is to provide a means for improving mark detection accuracy without using a chromatic aberration correction lens or the like when detecting marks using a wavelength different from the exposure wavelength, and achieving high precision in pattern alignment. It is in.

[Means to solve the problem]

In order to achieve the above object, in the present invention, as shown in Figs. Two grating patterns 12 are provided on the wafer as alignment marks so as to correspond to the two FZPs on the reticle. The alignment mark is illuminated with light 6 having a wavelength different from the exposure wavelength. In addition, the 0th order light component of the diffracted light from the wafer is removed within the detection optical system, and ±1
A spatial filter was provided for extracting the 7-1, 7-2 order light components or the diffracted light components of the 1st ± 1st order or more.

[Effect]

In the above configuration, as shown in FIG. 1, the alignment mark 12 on the wafer is illuminated via the projection lens 3 with light 6 having a wavelength different from the exposure wavelength. The image of the alignment mark 12 is back-projected by the projection lens 3, and if the FZP 2 were not provided, an image J would be formed at a position considerably shifted from the reticle 1 in the optical axis direction due to the chromatic aberration of the projection lens 3. Therefore, since the FXP 2 placed on the reticle 1 functions as a chromatic aberration correcting lens, it becomes possible to form the image (2) of the alignment mark 12 at an arbitrary position. Note that the FZP may be of a transmission type or a reflection type.

Further, the FXP 2 is of a linear type and is arranged perpendicularly to the alignment mark 12 on the wafer.

Also, since the alignment mark 12 is a grid pattern,
If only the ±1st-order diffracted lights 7-1 and 7-2 from the grating pattern are selectively taken out, the imaged registration mark image becomes an interference pattern consisting of a pitch of - of the grating pattern pitch. In this relationship, the alignment mark image ■ is FZP,
In other words, if the reticle is displaced in the Y direction, it is displaced in the Y direction, but there is no change even if it is displaced in the X direction. On the other hand, if the wafer is displaced in the X direction, the machine will be displaced in the X direction, but the Y
Even if the machine is displaced in the Y direction, the machine will not be displaced in the Y direction. From the above, the displacement of the reticle and wafer in either of the X and Y directions can be determined using a set of FXP and alignment marks. Therefore, as shown in FIG. 2-A, by using two sets of FZPs and alignment marks, it is possible to simultaneously detect the displacements of the reticle and wafer in the XY force directions. In other words, one set is FZP2
a and the alignment mark 12a, and the other set is a combination of the FZP 2b and the alignment mark 12b arranged orthogonally to this. By doing this, the alignment mark image r is
Images are formed as shown in Ia and Ib. Now, if the positional relationship between the reticle and the wafer is misaligned, the images of the alignment marks are ■ and L.

It becomes like Ib'. Therefore, from the ideal images Ia and Ib, X rt + X w + YR in the Y direction.
This means that it is shifted by +Yw. In addition, two images Ia
, Iゎ, Xo, and Yo are values determined by the distance between the two FXPs 2 and the alignment mark 12, and the magnification of the projection lens d and D, respectively.

As described above, the relative position between the reticle and the wafer can be accurately detected without providing a chromatic aberration correction lens between the reticle and the reduction lens as in the conventional case.

Note that it is effective not only to form an image using only the ±1st-order diffracted light of the light reflected from the alignment mark 12 as described above, but also to form an image using components of ±1st order or higher by simply removing the 0th-order light. It acts on However, in this case, it is necessary to make the size of the alignment mark sufficiently larger than the FZP.

In addition, by passing through a spatial filter provided in the detection optical system, the image to be formed is formed of the ±1st-order light component of the diffracted light diffracted by the alignment mark on the wafer, or the diffracted light component of ±1st-order or higher order. is formed. Therefore, the detection signal is not affected by interference from the photoresist film on the wafer, and the contrast can be increased.

Further, the FZPs 2a, 2b and alignment marks 12a, 12b shown in Fig. 2-A may be set as shown in Figs. 2-B and 2-C with d=0°D=O, respectively. By doing so, the area of the FZP and alignment mark can be made smaller.

Furthermore, if the chromatic aberration of the reduction lens is very large, the contrast of an image formed by detection light different from the exposure wavelength will be low due to the aberration. In such a case, by using a spatial filter that passes only the ±1st-order diffraction components, the image formed becomes an interference pattern, so that a detection signal with very good contrast can be obtained.

Furthermore, in the present invention, the alignment marks are arranged at the periphery of the projection area of the projection lens, and the patterns are arranged orthogonally to each other, so it is conceivable that the astigmatism of the projection lens will be affected. To this end, if the FZP pattern is a pattern that corrects astigmatism, two orthogonal patterns can be detected at the same focal position without defocusing.

In this way, a detection signal with very good contrast can be obtained, making it possible to detect the pattern position with high precision.

〔Example〕

An embodiment of the present invention will be described below with reference to FIG. Third
The figure shows the present invention implemented in a reduction projection exposure apparatus using an excimer laser as a light source. Wavelength 248 as exposure light
．． 4 nm, F laser light is used, and the reduction lens 3 has a refractive index of 1.5083 with a focal length of 801 m and a magnification of 115.
A monochromatic type made of only synthetic quartz glass is used.

If the reduction lens 3 is telecentric to the wafer 4 side, the distance between the reticle 1 and the entrance hole is 4801f.
It becomes f11. The light source 5 for mark detection has a wavelength of 632
．． An 8 nm He-Ne laser is used.

An FZP pattern 2 is provided on the reticle 1.

A grating pattern 12 is provided on the wafer 4 as an alignment mark. Mark detection optical systems 9, 10, 1.1 are provided above the reticle 1.

The operation for aligning the relative positions of the reticle 1 and the wafer 4 using the apparatus having the above configuration will be explained. A beam 6 emitted from the H e -Ne laser 5 installed between the reticle 1 and the reduction lens 3 is bent by the beam bender 8 and passes through the reduction lens 3 onto the wafer 4 at the alignment mark formed in the previous process. 12. The reflected light 7-1°7-2 from the alignment mark 12 passes through the reduction lens 3 again, and the image of the alignment mark 12 is back-projected onto the reticle side. Now, if we consider the case where there is no FZP2 on the reticle 1, the image of the alignment mark 12 will be formed at a position J, which is a distance L away from the reticle 1. In the case of the reduction lens used here,
Since a wavelength of 632.8 nm is used as the detection light, the refractive index of the quartz lens material for the detection light is 1.457.
1, and the focal length of the reduction lens also changes from 801m to 89mm.
It changes a lot. Therefore, the distance from the reticle 1 to the imaging position J of the alignment mark is 734 fields. Moreover,
Since the reduction lens 3 is a monochromatic type lens, not only chromatic aberration but also other aberrations occur, and the contrast of the formed alignment mark image becomes extremely low.

Therefore, in the present invention, an image of an alignment mark is formed on the reticle 1 at a position S near the reticle using FZ P2. The distance S from the reticle 1 to the imaging position is determined by the focal length of the FZP. In this example, the focal length of FZP2 is 1
Since it is set to ono, the distance S becomes 10.14 +a. F
The dimensions of ZP2 were 0.25 rrtn square, and the alignment marks 12 on the wafer were 40 tLm square with a pitch of 8 μm.

The alignment mark image (2) formed by the FZP2 is magnified by the detection optical system 9.10-1.10-3 and re-imaged on the imaging surface of the two-dimensional solid-state image sensor 11 as an image (2) s'.

A spatial filter 10-2 is provided in the detection optical system, and is configured to allow only the ±1st-order diffracted light among the reflected diffracted light from the alignment mark]2 to pass through. Therefore, the image of the alignment mark is formed as an interference pattern.

On the solid-state image sensor 11', as shown in FIG. 5(a), 1. /, a pattern of l is projected. The solid-state imaging device has M x N photoelectric conversion sections in two dimensions, and converts changes in brightness of the projected pattern into changes in voltage. In this example, an element of 512×500 picture elements was used. Signal processing is performed by the circuit shown in FIG. 5(d). In order to find the position on the element of the Ia' pattern currently projected onto the element, attention is paid to the rightmost pattern of Ta2.

The position error of the pattern Ia' is 1 (M
o , N o ). To obtain this position error, signal processing is performed only on the (11j)th to (i+m, j+n)th regions including (M o , N o ). That is, to obtain the position in the X direction, the light intensity from j to j + n is integrated in each X direction from i to i+m by the integrating circuit 20 and stored in the memory 21 . The optical intensity signal in the memory is shown in Figure 5 (
b). Similarly, for the position in the X direction, the light intensity from i to i+m is accumulated for each position in the X direction from j to j+n and stored in the memory 21. This light intensity signal becomes as shown in FIG. 5(c). Note that the regions (i, j) to (i+m, j+n) for which the light intensity signal is obtained are always at the same position. Since the wafer 4 and reticle 1 are roughly positioned in advance by ±5 μm using a different method, the pattern Ia' always falls within a predetermined area. After the optical intensity signal is stored in the memory using the above method,
In order to obtain the center position i + x or j + y of each pattern, it is sufficient to calculate the neutral point of the point that intersects the thresholds THX and THY.

Similarly, for the Ib' pattern, focusing on the upper end pattern, calculate the light intensity signal in the area from k, Q) to (k+m, Q+n) and set +η to the center position of the pattern.

Find Q+ξ.

To obtain the amount of deviation of each pattern from the center position of each pattern in the XY force direction obtained in this way.

This is performed by the arithmetic circuit 22 according to the following arithmetic expression.

Xw=Mo−(i+x) XR=(k+η)-Pa YR=(i+y)N.

Yw=Q CQ+ξ) Also, the relative position of Ia' and Ib' at the ideal position [X
o, Yo are X o = P o Mo, Yo
= Q o No.

Note that the magnification of the detection optical system is 40x, so the
The converted dimension on the surface of each pixel is 0.2 μm.

Furthermore, the detection resolution can be set to 0.02 μm by interpolating the distance between picture elements to 1/10.

With the above method, the relative positions of the reticle 1 and the wafer 4 can be obtained by determining XR+XW and YR+YW as shown in FIG. Next, set this value to zero,
Moreover, the fine movement base on which reticle 1 is mounted is slightly moved so that the positional relationship Xo and Yo between Ia' and ■b' becomes the predetermined value.
The relative positioning of the reticle 1 and the wafer 4 is performed.

On the other hand, in the above example, a pair of FZP and alignment marks
Alignment was performed in the Y force direction, but in order to align the reticle and wafer with higher precision, two sets of FZP2.2' were placed on the reticle 1 as shown in FIG.
At the same time, two sets of alignment marks 12 and 12' may be placed on each chip 15 on the wafer 4. By doing so, the rotational positional relationship between the reticle and the chip on the wafer can be detected accurately, so that highly accurate alignment can be achieved.

Although the above description has been made with respect to an embodiment of a reduction projection exposure apparatus using an excimer laser as a light source, the present invention can also be applied to an exposure apparatus using as a light source the g-line, r lt i-line, etc. of the emission spectrum of a mercury lamp. Furthermore, the present invention is not limited to a reduction projection type exposure apparatus, but can also be applied to a projection exposure apparatus of equal magnification, a proximity exposure apparatus that does not use a projection optical system, and the like.

〔Effect of the invention〕

According to the present invention, when mark position detection is performed using light of a wavelength different from the fi light wavelength, the FZP on the retake functions to correct chromatic aberration, so even if chromatic aberration occurs, new optical Since there is no need to add a system, there is an advantage that position detection and alignment can be performed with extremely high precision.

[Brief explanation of the drawing]

1 and 2 are diagrams showing the principle of the present invention, FIG. 3 is a schematic diagram of a reduction projection exposure apparatus which is an embodiment of the present invention, and FIG.
The figure shows the alignment mark reticle. An example of arrangement on a wafer, FIG. 5 is a diagram showing a signal processing method and processing circuit, and FIG. 6 is a schematic diagram of the configuration of a conventional example. DESCRIPTION OF SYMBOLS 1... Reticle, 2... Fresnel zone plate, 3... Projection lens, 4... Wafer, 6... Mark detection light, 9, 10... Detection optical system, 11... 2
Dimensional Photoelectric Inspection No. 1 Fig. HA - /2-B Fig.

Claims (1)

[Claims] 1. In a device for detecting the relative position of a first object and a second object, at least one set of linear Fresnel zone plates provided on the first object and whose band directions intersect with each other. an alignment mark having at least one set of mutually intersecting lattice patterns corresponding to the linear Fresnel zone plate provided on a second object; and an image of the alignment mark formed by forming an image on the linear Fresnel zone plate. A pattern position detection device comprising a detection optical system that forms an image and a detection means that detects a two-dimensional position of an image of the alignment mark. 2. Claim 1, wherein the one set of Fresnel zone plates has strip directions perpendicular to each other.
The pattern position detection device described in . 3. The set of alignment marks is provided in a direction such that the grids of the grid-like pattern of alignment marks on the bands of the corresponding linear Fresnel zone plate intersect perpendicularly. Or the pattern position detection device according to item 2. 4. Any one of the claims set forth in claim 3, wherein the detection optical system has a spatial filter that passes diffracted light components of ±1st order or higher of the light forming the image of the alignment mark. A pattern position detection device. 5. A projection lens is provided between the first object and the second object, the alignment mark on the second object is projected onto the first object side by the projection lens, and the alignment mark on the second object is projected onto the first object side; When an image is formed by the upper linear Fresnel zone plate, the pattern position according to any one of claims 1 to 4, wherein the linear Fresnel zone plate has a function of correcting aberrations of the projection lens. Detection device.