JPS63315903A - Pattern position detector - Google Patents

Abstract

PURPOSE:To eliminate the need for a spatial filter which need to be positioned with accuracy by interposing a cylindrical lens in the optical pat of a detection optical system and forming images of the pattern of an object of detection and its Fourier transformation surface on a two-dimensional image pickup element. CONSTITUTION:An axially asymmetric optical element such as a cylindrical lens 19 is interposed in the detection optical system and a pattern image of a surface crossing an alignment pattern on a wafer is formed in one picture array direction of the two-dimensional image pickup element 20. Further, an image of the Fourier transformation surface in the lengthwise direction of said alignment pattern is formed in the other picture element array direction of the two-dimensional image pickup element 20. Then, a pattern position is detected from an image of an optional desired range on the two-dimensional image pickup element 20.

Description

[Detailed Description of the Invention] [Field of Application of Tsutomu Higashi] The present invention relates to a pattern position 1t&' output device for a projection path light device, which is a high-resolution reduction projection device that transfers a tracing pattern at the same time. Pattern position y suitable for the light beam device if
It relates to a detection device.

[Conventional technology]

Conventionally, for semiconductor memory devices and other circuits, a resist (C photosensitive material) coated on a 5i substrate was used to print a circuit pattern of KF9r, which was then printed, and then subjected to chemical processing such as etching. produced in return. At this time, it is necessary to use a projection machine to accurately align and print the arrow pattern onto the circuit pattern already formed on the board. , by detecting the position of the time-saving pattern (alignment pattern) on the board.

As the degree of integration of integrated circuits increases, circuit patterns (
A finer pattern (0.7 to 1.6 μnL) has come to be used, and the original projection Ii#r device uses a reduction lens with a straw-dust image resolution to transfer the circuit pattern on the reticle (mask) to the substrate (wafer). Small projection exposure devices that route light to the surface became mainstream. Furthermore, with the miniaturization of circuit patterns, the accuracy of the alignment described above needs to be improved, so TTL (Thro
tL! Ih The Let's ) type pattern position detection devices have come into widespread use.

However, due to the characteristics of the reduction lens, this TTL pattern position detection device can only use monochromatic light for pattern detection, and the normal bright field detection method is easily affected by multi-page interference within the resist. It has characteristics. Due to this characteristic, in circuit patterns with many uneven resist coatings, the method of drawing the pattern on the urchin is detected as being distorted and the SIN ratio decreases, which increases the error in pattern position detection and causes deterioration of alignment accuracy. was.

For such a situation, for example, Japanese Patent Application Laid-Open No. 56-12729
A pattern position Ilf detection device as described in the above publication has been proposed. This pattern position detection device uses a pattern with a shape that diffracts light at a specific angle as an alignment pattern, and also provides a spatial filter with a specific shape in the detection optical system, and detects the alignment pattern position from this diffracted light component. This reduces the contrast of the detected image to improve alignment accuracy. Furthermore, in order to improve the accuracy of pattern alignment, a method is used to detect the pattern of the keratin wafer at the optical path position. In this path light position detection method, it is necessary to move the position of the detection optical system in accordance with the dimensions of the circuit pattern on which the path light travels. However, when performing this path light position detection, the relative position with the reduction lens changes as the detection optical system moves, so the position of the spatial filter inserted in the detection optical system must be precisely controlled. Furthermore, in order to be able to insert a spatial filter into the detection optical system, an optical system with a complicated configuration was required.

[Problem that the invention seeks to solve]

In the above conventional technology, it is necessary to precisely control the position of the spatial filter inserted in the detection optical system.
C. There was a problem in that an optical system with a special configuration was required to enable insertion of a spatial filter.

SUMMARY OF THE INVENTION An object of the present invention is to provide a highly accurate pattern position detection device that does not require a spatial filter that requires close alignment and can selectively detect specific diffracted light when detecting a pattern position.

[Failure to solve the problem]

The above purpose is to provide a pattern position detection device that detects the position of an alignment pattern on an object through an optical system that projects a pattern on an original plate such as a mask onto an object such as a wafer. Using the above pattern, a non-axisymmetric optical element such as a cylindrical lens is inserted into the detection optical system, and the pattern image is formed in one pixel arrangement direction of the two-dimensional element, while the above image is formed in the other pixel arrangement direction. This is achieved by a pattern position detection device provided with a processing circuit that forms an image of the Fourier transform plane of the pattern image and detects the pattern position from an image within an arbitrary 79r range of the two-dimensional imaging device.

[Effect]

In the above pattern position detection device, the above-mentioned periodic broken-line alignment pattern is used, and a non-axisymmetric optical element such as a cylindrical lens is inserted into the detection optical system to form one side of the two-dimensional Shitsuka element. In the pixel arrangement direction, a pattern image of a plane perpendicular to the alignment pattern is displayed, and in the other pixel arrangement direction of the two-dimensional firing element, an image of the Fourier transform plane of the longitudinal plane of the alignment pattern is displayed. imaged, and the above 2
Select and extract I-Q of the pixel row (single row or parallel row) that detects the diffracted light at a specific angle generated from the alignment pattern using the detection image signal of the first element, and extract the extracted 1g. By detecting the center position of the alignment pattern using the 2D image sensor, the selection of the diffracted light at the specific angle can be performed by selecting the pixel row to be used for pattern detection. Since it can be easily distinguished from the entire detected image and the selection range can be electrically set arbitrarily, there is no need to worry about precise positioning of the spatial filter, which was indispensable in the past.

[Actual example]

An actual JM example of the present invention will be explained below with reference to FIGS. 1 to 10.

FIG. 1 shows a pattern position detection device according to an embodiment of the present invention.
It is a perspective view showing #4 formation. In FIG. 1, 2 is a reticle (mask), 5 is a pattern position detection device, 6a is a mirror, 7 is a mirror, 8 is a galvanometer mirror, 19 is an illumination lens, 10 is a mirror, 11 is an illumination optical system, 12 is a mirror,
13 is an objective lens, 14 is a wavelength plate (1/4λ), 15 is a polarizing beam splitter, 16 is an imaging lens, 17 is a mirror, 18 is a secondary magnifying lens, 19 is a cylindrical lens (non-axisymmetric original tuition fee line) , 20 is a two-dimensional imaging element, 21
is a stage, and 22 is a 4R laser beam. The Ar laser source (S polarized light) 22 from the Ar laser 6 (Fig. 3) is reflected by the mirror 6α, then reflected by the mirror 7 of the illumination optical system 11 and the galvano mirror 8, and then traces the illumination lens 9, and then passes through the mirror. 11 and the polarizing beam splitter 15, the wavelength plate 14, objective lens 13, and mirror 12
and the pattern surface of the reticle 2, and turn the reduction lens 5 (third
) and reaches the surface of the wafer 1 (FIG. 3).

Further, the reflected light from the 9eva 1 passes through the reduction lens 6, is reflected by the reticle 2 and the mirror 12, passes through the objective lens 13, the wavelength plate 14, the polarizing beam splitter 15, and the imaging lens 16, and is then reflected by the mirror 17. , further follows the secondary magnifying lens 18 and the 7-lindrical lens 19 of the detection optical system to reach the two-dimensional carrier element 20.

Furthermore, the entire pattern position detection device [5] is mounted on a stage 21, and the pattern position detection device [5] is mounted on a stage 21, and the mirror 6 is removed in the radial direction (direction of arrow R) and circumferential direction (direction of arrow T) with respect to the field of the reduction lens 3 (Fig. 5). It is configured to be movable. and an illumination optical system 11 consisting of a mirror 7, a galvanometer mirror 8, and an illumination lens 9.
is configured to be movable in the direction of the arrow inside the pattern position detecting device 5 in accordance with the movement of the pattern position detecting device 5 in the T direction. Note that the same reference numerals indicate the same or corresponding parts throughout the drawings.

Figures 2 (α) and (b) are the Tsunino 1 of the embodiment shown in Figure 1.
1 is a plan view illustrating the alignment pattern shape on 1 and the detected image of the two-dimensional imaging element 20. Figure 2 (α
), Cb), 1α is an alignment pattern, 20α is a detected image, 30 is a specularly reflected light component, and 31 is a diffracted light component. In the embodiment shown in FIG. 1, as the alignment pattern on the wafer 1, an alignment pattern 1α having a shape (tentative box shape) in which minute elements are periodically arranged as shown in FIG. 2(α) is used. The detected image in FIG. 2(h) will be explained later. FIG. 3 is a perspective view illustrating the configuration of a reduction projection uniting device equipped with the pattern position t & horizontal extraction device M, 5 of the actual image example in FIG. 1. It is. In Fig. 3, 1 is a wafer, 2 is a reticle (mask), 3 is a reduction lens, 4 is a 9-eva stage, and 5 is a pattern position detection gff4-16.
It is @AT laser. The reduction projection light device in Figure 3 is
A reticle 2 is illuminated by an illumination system (not shown) onto a wafer 1 placed on a wafer stage 4 movable in the y direction.
It has a configuration in which the pattern is reduced (such as 175) by a reduction lens 3 and projected. Between the reticle 2 and the reduction lens 3, the pattern position detection device 5 of the embodiment shown in FIG. With this configuration, the wafer stage 4 is successively moved in accordance with the pattern arrangement of the wafer 1 in the m small projection film-lighting device, and at that position, the pattern position detection device 5 detects the alignment pattern 1α on the wafer 1 through the reduction lens 3. Detect location. Next, an alignment error is obtained from the detection result of the alignment pattern position, and this error is corrected by moving the wafer stage 4. Thereafter, the pattern of the reticle 2 is projected onto the wafer 1 by the reduction lens 5 and printed, thereby achieving a high degree of pattern alignment.

FIG. 4 is a block diagram illustrating the configuration of the image processing circuit of the embodiment shown in FIG. 1. In Figure 4, 20 is 2? ' is the original image sensor (solid-state element), 23 is the A/D conversion circuit 4 is the addition circuit, 25 and 26 are frame memories, 27
is a microcomputer. The pattern position detection device 5 of the embodiment of FIG. 1 is connected to the A/D converters 26 and 71 of FIG.
An image processing circuit consisting of a zero arithmetic circuit 24, frame memos 1725 and 26, and a computer 27 is provided inside to select and process an image within an arbitrary range of the two-dimensional image element 20, and to process the obtained position data. It has a function that can be output to a stage 210 control circuit (not shown).

Next, the basic operation of the embodiment shown in FIG. 1 will be explained. First, a first alignment pattern 1α on the wafer 1 having periodicity as shown in FIG.
In the figure, the 4R laser element (S polarization t) 22 is connected to the mirror 6, the mirror 7 of the illumination optical system 11, the galvanometer mirror 8, and the illumination lens 9.
The beam is incident on the polarizing beam splitter 15 from the mirror 10, reflected, and incident on the temporary plate (1/4λ) 14. Further, the 4R laser source 22 becomes circularly polarized light after passing through the wavelength plate 14, passes through the objective lens 13 and mirror 12, is reflected by the reticle 2, and then passes through the small lens 3 shown in FIG. 11 (coherent illumination). Further, the reflected light from the alignment pattern 1α on the wafer 1 passes through the reduction lens 6 in the opposite direction to the incident light, is reflected by the reticle 2, and forms an image of the alignment pattern 1a in the space below the reticle 2. Furthermore, this reflected light is reflected by the mirror 12 and the objective lens 13.
The light then enters the wavelength plate 14 and becomes P-polarized light. Then, the beam passes through the polarizing beam sinter 15 and follows the imaging lens 16, mirror 17, secondary magnifying lens 18, and cylindrical lens 19, and 2? : It reaches the former Naho image element 20 and is converted into an electric priest.

5(a) and 5(b) are optical path diagrams illustrating the imaging relationship of the detection optical system of the embodiment shown in FIG. Figure (b) is a cross-sectional view in a direction orthogonal to the alignment pattern. Figure 5 (
α) In (b), α is the pupil of the small lens 5,
28 is a back projection image plane of the wafer 1, and 29 is an enlarged image of the wafer 1. The imaging relationship of the detection optical system of the embodiment shown in FIG.
This will be explained with reference to Figures (z) and Cb).

In FIGS. 5(a) and 5(b), each element is arranged in one direction to show the imaging relationship of the detection optical system. First, Figure 5 (
Regarding the plane that includes the central axis of the reduction lens 3 shown in α) and is parallel to the direction of the alignment pattern 1α on the wafer 1, in the direction of this alignment pattern 1α, the objective lens 13, the imaging lens 16, and the secondary magnification The lens 18 projects an image of the Fourier transform surface of the lens 3α of the reduction lens 5 behind the secondary enlarging lens 18, and this projected image of the recitation 5a of the reduction lens 3 is projected onto the cylindrical lens 19.
The image is re-imaged on the two-dimensional copying image element 20 by the following steps. Next, looking at the plane parallel to the central axis of the reduction lens 5 and perpendicular to the direction of the alignment pattern 1α shown in FIG. 5(b), the image of the alignment pattern 1α on the wafer 1 is First, a ti image is formed on the back projection image plane 28 in the space near the reticle 4, and this formed image of the alignment pattern 1α is further enlarged into an enlarged image 29 by the objective lens 13 and the imaging lens 16. This magnified image 29 is projected by the secondary magnifying lens 18 onto the two-dimensional imaging element 20 through the cylindrical lens 19. Due to such an imaging relationship, the two-dimensional image sensor 2
0 is a detected image 20α as shown in FIG. 2(b), for example.
The specularly reflected light component (0-dimensional) 30 and the diffracted light component (1st order, 2nd order)
31 separate images are formed.

FIG. 6 is a longitudinal cross-sectional view illustrating how the diffracted light 31 is generated in the alignment pattern 1α of the embodiment shown in FIG. Also, Figure 7 is the fruit of Figure 1! FIG. 7 is a perspective view illustrating an illumination source and reflection positions of reflected light and diffracted light on 20 surfaces of a reticle in example M; In FIG. 7, 2α is a circuit pattern portion of the reticle 2, 62 is a mirror pattern, and 53 is a range of diffracted light beams. Figure 8 (α).

(, 6) and (C) are explanatory diagrams illustrating the relationship between the detected IN mound 20 (Z) and the conventional waveform in the embodiment of FIG. 1, and FIG. Fig. cb> is a detection waveform diagram by conventional bright field detection, and Fig. 8(C) is a detection waveform diagram by an embodiment of the present invention. Fig. 8(b), (C
), the horizontal axis is the pixel, and the book axis shows the detection (G5 insect degree).

Next, the operation of the embodiment shown in FIG. 1 will be explained in more detail with reference to FIGS. 6 to 8 (α), Cb), and (C). Since the shape of the alignment pattern 1α on the wafer 1 shown in FIG. 1 has periodicity as shown in FIG.
Alignment pattern 1α on wafer 1 as shown in the figure
In addition to the specularly reflected light component (0-dimensional) 50, diffracted light components at specific angles (1st order, 2?X diffraction source, etc.) 31.31' are generated. Then, the specularly reflected light 3o of these reflected lights and diffracted lights passes through the reduction lens 6, and is also reflected by the reticle 20 provided adjacent to the circuit pattern 2α, as shown in FIG. The light is reflected by the mirror pattern 32 that is being held, and reaches the detection optical system having a pattern position detection length tit5. Similarly, the diffracted light component 61 follows the slightly small lens 3 and is reflected at the outer part of the mirror pattern 32 in the range ([M part) 33 where the reflected light and diffracted light flux returning from the wafer 1 reach the reticle 2. , reaches the detection optical system of the pattern position detection device 5.

Then, the detection optical system of the pattern position detection device 5 generates a detected image 2α of the two-dimensional image sensor 20 as shown in FIG. One house separated by 31 is detected.

Next, the operation of detecting the position of the realigned pattern 1α from the detected image 20α of the two-dimensional image sensor 20 will be explained. Masu above detection +1! The i-image 20α is compared with the smooth surface image stored in the frame memory 25 to correct illumination unevenness and sensitivity unevenness by the image processing circuit shown in FIG. 4, and is digitized by the AID'4;3 circuit 23. . Then, the digitized image data 25α is sent to the boost circuit 241C.
The data is then transferred within an appropriate range (window data 22α from the microcomputer 27) with the stroke data stored in the frame memory 26, and output as waveform data 24α. Next, this waveform data 24α is subjected to image processing such as 21 straightening using a threshold value or pattern matching in the microcomputer 27 to detect the position of the alignment pattern 1α, and the position data 27b is transferred to the control circuit of the stage 21. The pattern is aligned by outputting it to .

In this way, as shown in FIG. 8(b), in the conventional detection method using a bright field image, the specular reflection component 30 and the diffracted light component 31 of the detected image 20α as shown in FIG. 8(α) cannot be separated. Therefore, if there is a difference in resist thickness between the left and right sides of alignment pattern 1α due to uneven resist coating, the left and right side of alignment pattern 1α 8 (C ), in the method of the embodiment of the present invention, it is possible to select and detect only the diffracted light component 31 of the detection image 20α (adding a part including the diffracted light component 31 of the detection image 20α). Therefore, even under conditions where resist coating unevenness exists, it is possible to have a significant area of signal stiffness in the vicinity of alignment pattern 1α, while the detected waveform has a value close to 0 in other parts, and the high phase at the position of alignment pattern 1α can be observed. It has the effect of making frequent toileting easier. Furthermore, in this embodiment, the selection of the diffracted light component 31 can be performed by setting the window data in the 710 depth range in accordance with the detected image 20α of the two-dimensional image sensor 20, so that the selection of the diffracted light component 31 can be performed with precision, which is essential in the conventional device. This has the effect of eliminating the need for spatial filter alignment.

FIG. 9 is a longitudinal sectional view showing the structure of a pattern position tIL detection device according to another embodiment of the present invention. In Figure 9, 34
is a solid-state imaging camera. Similar to the embodiment in FIG. The wafer 1 is coherently illuminated through the reticle 2 through the lens 13 and the mirror 12, and then through the reduction lens 3. From this polishing, the beam of the 7-line pattern 34 of the wafer 1 is imaged below the reticle 2. The image is transmitted through a polarizing beam splitter 15 by a detection optical system consisting of a mirror 12, an objective lens 13, and a cylindrical lens 19, and is enlarged and projected onto a two-dimensional image sensor 2o of a solid-state imaging camera 350.

FIG. 10 is a longitudinal sectional view of the glaze showing the function of the cylindrical lens (non-axisymmetric optical system) 19 of the embodiment shown in FIG.

In FIG. 10, 13α is the position of the objective lens 13, and 35
is an image of the saliva 6α of the reduction lens 6. In this embodiment as well, as shown in FIG. 10, the image 35 of the Fourier transform plane of the pupil 3α of the reduction lens 3 formed slightly behind the pupil 13α of the objective lens 13 is cylindrical in the direction of the alignment pattern 1α of the Ueno S1. Solid 3 by lens 19
It is clear that the image is projected onto the two-dimensional image sensor 20 of the 'IEkiM camera 34, and that the same operation as in the previous embodiment of FIG. 1 is performed.

In this way, according to this embodiment, as in the practical example shown in FIG. 1
As shown in Figure 0, g! of the pupil 3a of the reduction lens 3! ! 35 is imaged into the configuration of the detection optical system (objective lens 15),
It is possible to achieve the same effect as inserting a spatial filter even in a general detection optical system where it is not possible to insert a spatial filter.Furthermore, the height exceeds the height of the two-dimensional imaging element 20 by several fis. Since the image in the range H of can be compressed and detected by the cylindrical lens 19, for example, j41
In processes where the wafer surface exhibits graininess, as in the process, the effect of averaging the underlying noise is great (if the underlying noise is reduced to 1
(can be reduced to 7 first τ).

In the above two embodiments, a cylindrical lens is used as the non-axisymmetric optical element, but there are other types of cylinders 11 [I theory, gratings, refractive index distribution type temporary glass, etc. W cables can be used in the same way.

〔Effect of the invention〕

According to the present invention, a symmetrical element (cylindrical lens) is arranged in the optical path of the detection optical system, and an image of the pattern to be detected is formed in the direction in which one of the 1I111 elements of the two-dimensional focusing element is arranged. At the same time, an image of the Fourier transform plane of the pattern to be detected is formed in the direction of the other pixel arrangement, thereby forming an alignment pattern south of the processing of the detected image of the two-dimensional image carrier. By selecting an image in a range that matches the angle (diffracted light), the position of the alignment pattern can be detected from the information trap of this selected image. Therefore, it is not necessary to insert a spatial filter (only light with a specific diffraction angle can be used for detection), so QSIN is difficult to be affected by uneven resist coating.
In addition to achieving highly accurate alignment pattern positioning performance by inputting the ratio detection waveform, it also has the advantage of eliminating the need for precise spatial filter positioning and allowing settings to be made within the processing circuit ('findator). be.

Furthermore, according to the present invention, since there is no need to insert a spatial filter, the same effect as the insertion of a spatial filter can be obtained even in the 7IJ conversion surface of the pattern to be detected or in the optical system existing within the configuration of the detection optical system. Because it is difficult to detect images in a range that is wider than the two-dimensional image element, the background noise is smoothed and low i
There is an effect of (171,4~1/2).

[Brief explanation of the drawing]

FIG. 1 is a perspective view showing the configuration of a pattern position @L detection device according to an embodiment of the present invention, and FIG. 2 (α) and (, 6) show the alignment pattern shape and detection lj! of FIG. 1. FIG. 5 is a perspective view showing the configuration of the reduction projection path optical arrangement equipped with the pattern position I shown in FIG. 1, and FIG.
A block diagram showing the configuration of the processing circuit shown in FIG. 5 (α),
(A) is an optical path diagram of a cross section in the direction of the alignment pattern and a cross section in the direction perpendicular to the alignment pattern showing the imaging relationship in FIG. Figure 7 is a perspective view showing the reflection positions of the illumination light, reflected light, and diffracted light in the reticle surroundings in Figure 1, and Figures 8 (α), (b), and (C) are the extruded outline shapes in Figure 1. FIG. 9 is a longitudinal cross-sectional view showing the configuration of another pattern position detection device of the present invention, and FIG. FIG. 1 is a detailed longitudinal sectional view showing the steering wheel of the steering wheel. 1... Wafer 2... Reticle (mask) 3... Reducing lens 4... Wafer stay 75... Pattern position adjustment 6..., 4r laser 13... Objective lens 16...・Imaging lens 18... 2-arrow large lens 19... Cylindrical lens (asymmetric optical ☆ cable)
20... Two-dimensional Hachizuka element 25... 74/D conversion circuit 24... Addition circuit 30... Specularly reflected light 31... Diffraction source 32... Mirror pattern 65... Small lens Figure 20-n: 45Figure 25 Otsusu porridge F5 figure

Claims (1)

[Claims] 1. In a pattern position detection device that detects the position of a pattern on an object such as a wafer through an optical system that projects a pattern on an original plate such as a mask onto the object such as a wafer, A non-axisymmetric optical element such as a cylindrical lens is inserted into the optical path of an optical system that detects the position of the pattern on the object, using a pattern in which microelements are arranged periodically to detect the pattern on the object. A two-dimensional image sensor is used as the image detection element of the object, and a processing circuit is provided to select and process a detection image of a desired range of the two-dimensional image sensor to obtain positional information of the pattern on the object. pattern position detection device. 2. Form an image of the pattern of the object in one pixel arrangement direction of the two-dimensional image sensor, and form an image of the pattern of the object on or near the Fourier transform plane of the object pattern in the other pixel arrangement direction of the two-dimensional image sensor. 2. The pattern position detection device according to claim 1, wherein an image of the surface is formed.