JPS59161815A - Device for detecting rotational deflection of exposing equipment - Google Patents

Abstract

PURPOSE:To perform an excellent exposure even when there exists a rotational deflection present on an optical image by a method wherein, when a material to be exposed whereon a mask pattern is be transferred is placed on a stage is exposed on a two- dimensional movable stage, a local part of the optical image of the pattern is detected by a microscopic aperture part, the surface of which is aligned with the material to be exposed, and the stage is controlled by the result of said detection. CONSTITUTION:The light outputted from an illumination light source 1 to be used for exposure is focussed by the first condenser lens 2, is made incident on the second condenser lens 3 through a shutter 4a, and the light emitted from the lens 3 is made to be incident on the test reticle 5 supported by a holder 15. At this point, light-transmissing marks RR and RL and provided on the reticle, and they are controlled by the light- shielding materials 4b and 4c which are provided on the lens 3. Then, the beam of light passed through a mark is projected on the microscopic aperture member 8, having the surface aligned with the semiconductor wafer 10 by a gap sensor 12 using a projection lens 6, while this light is converted into an electric signal using a photoelectric detector 9, and the stage 7 mounting the wafer 10 thereon is controlled.

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to a (b) dislocation detection device for exposure equipment, and in particular to a transfer mask pattern for a high-density integrated circuit on a semiconductor substrate (
The present invention relates to a rotational deviation detection device suitable for use in a positioning system for positioning a wafer without rotational deviation with respect to an optical image of a mask pattern in an apparatus for exposing a wafer.

The miniaturization of large-scale integrated circuit (LSI) patterns is progressing year by year, and reduction projection exposure apparatuses have become popular as devices for printing circuit patterns that meet the demands for miniaturization and are highly productive. In these conventionally used devices, a reticle pattern that is several times (for example, 5 times) the size of the pattern to be printed on the silicon wafer is reduced and projected by a projection lens, and printed in one exposure. is larger than a square with diagonal length 21 on the wafer.
This is a rather casual area. Therefore, in order to print a pattern on the entire surface of a wafer with a diameter of about 125 mm, a so-called step-and-repeat method is used in which the wafer is placed on a stage, moved a certain distance, and exposed repeatedly.

In LSI manufacturing, several layers or more of patterns are sequentially formed on a wafer, but if the overlay error (positional shift) of patterns between different J- is not kept below a certain value, conductivity between one layer or If the insulation state is not as intended, the LSI will no longer be able to perform its functions. For example, for a [5+ path with a minimum line width of 1 μm, at most 0.
Only a positional deviation of about 2 μm is allowed. The positional deviations caused by such seizure can be broadly classified into rotational deviations and parallel deviations. FIG. 1 is an exaggerated diagram showing the state in which the rotational deviation has occurred, and the solid line rectangular pattern area P1 is a circuit pattern area that has already been formed on the sea urchin.
A rectangular pattern area P2 shown by a broken line is exposed and printed while being rotated by a very small amount with respect to this area P+ and superimposed thereon. In this case, the center O of the diagonal line of area P1 is the area P2
Assuming that it completely coincides with the center of the diagonal line of
Dislocation will occur.

In conventional reduction projection type exposure mounting, when performing a step-and-repeat operation, the stage on which the wafer is placed via the wafer boulder is moved and positioned based on the origin of the orthogonal coordinates that form the movement plane. On the other hand, a reticle on which a circuit pattern is drawn as a mask is positioned and fixed so that the projected image of this reticle on the wafer has as little rotational deviation as possible with respect to the IU orthogonal coordinates. In this state, a series of step-and-repeat exposure printing processes are performed in which coordinate positions of the orthogonal coordinate system are given, the stage is moved, and the stage is positioned at a predetermined position and exposed. By the way, if the detection center of the position detector used to position the reticle (for example, a reticle alignment microscope) is off, the reticle positioned based on the position detector will have a positioning error with respect to the orthogonal coordinate system of the stage. Therefore, the pattern area printed on the sea urchin generally causes rotational deviation. Conventionally, this rotational deviation has been corrected by actually printing a reticle pattern repeatedly onto a wafer, observing the printed pattern on the wafer using an optical microscope, etc., and calculating the amount of deviation between adjacent printed patterns. Measure 7. The deviation of the detection center of the position detector was calculated from the measured value and corrected. However, in this method, a pattern is printed on the wafer, the pattern is developed, and then the amount of rotational deviation is observed and determined, which takes a lot of time and effort, and the accuracy of the correction is not high.
The positional deviation of the circuit pattern due to rotational deviation is reduced to 0.1
It was extremely difficult to reduce the thickness to below μm.

The present invention solves the above-mentioned problems and converts rotational deviation into short poems.
The purpose of this product is to provide a rotational deviation detection device that enables high-precision detection with high precision, and is designed for use with an exposure tool that enables exposure work for each mask pattern to be carried out with high overlay accuracy and high efficiency. This makes it possible to realize a positioning system.

That is, in the rotational deviation detection device for an exposure apparatus of the present invention,
In order to sequentially expose a mask pattern onto an object to be exposed, such as a wafer, in a step-and-repeat method, two coordinates are set in the directions of both coordinate axes of an orthogonal coordinate system that forms a plane on which the object to be exposed is placed and moves.
A dimensionally movable stage is provided, and a predetermined local portion of the optical image of the mask pattern projected onto the stage, for example, an image of a mark attached to the mask pattern, is displayed on the stage. Photoelectric detection means are provided for detection. In order to detect the coordinate position of the local portion detected by this photoelectric detection means in the iIJ orthogonal coordinate system, for example, the stage moves to a position based on the origin of the orthogonal coordinate system of the stage. A position detecting means such as a laser interferometer is provided to measure when the photoelectric detecting means detects the local portion. Preferably, a control means constituted by a microcomputer or the like is combined so as to perform program control and various arithmetic processing in an integrated manner, and this control means controls the local portions of the optical image at different and complex points. The rotational deviation (inclination) of the mask pattern optical image with respect to the coordinate axes of the orthogonal coordinate system is calculated from the position information detected by the position detecting means. In a preferred application example of the present invention, another orthogonal coordinate system is defined in which the coordinate axes are rotated so as to offset this rotational deviation with respect to the orthogonal coordinate system, and the stage is moved in the direction of the coordinate axis of this another orthogonal coordinate thread. Positioning is performed by moving step by step.

According to the present invention, is the rotational deviation of the mask pattern optical image with respect to the movement coordinate axis of the stage itself? Instead of making it smaller, it is possible to control the position of the step movement of the stage so that even if this rotational deviation exists, it has no practical effect on the printing pattern.

The present invention will be explained in detail with reference to actual example drawings as follows.

FIG. 2 is a block diagram schematically showing a reduction projection type exposure apparatus as an example of the application of the present invention (displacement detection apparatus). Once converged, it reaches the second condenser lens 6.In the optical path, a shutter 4a is provided at a position where the light is converged to block the passage of the illumination light at a desired time.The second condenser The light beam passing through the lens 6 illuminates a test reticle (hereinafter simply referred to as "reticle") 5 serving as a mask.The light beam passing through this reticle 5 enters a projection lens 6 serving as an imaging optical system.

The projection lens 6 is an optical system in which the reticle 5 side, that is, the object side, is a non-telecenter link, and the image side is a telecenter link. The Sanno stage 7 of the projection lens 6 normally carries a semiconductor wafer 10 and moves two-dimensionally in the X and Y directions of the first orthogonal coordinate system, which forms the plane of movement of the semiconductor wafer 10. It is placed on a ueno holder 11 that moves two-dimensionally. The wafer holder 11 is provided so that it can be slightly rotated and moved up and down relative to the stage 7 . This Ueno holder 11 moves up and down so that the projected image of the circuit pattern (not shown) of the reticle 50 by the projection lens B is formed on the surface of the wafer 10, that is, so that focusing can be performed.

Now, on the bottom surface of the reticle 5, the above 1i:! In addition to the circuit pattern, light-transmissive marks RR and RL are drawn at predetermined positions (local portions) on both left and right ends. Luminous flux t transmitted through mark RR of reticle 5! Focusing on the luminous flux 1. is focused by the projection lens 6 and emitted as a light beam t2 from the image side, and is emitted from the small aperture member 8 provided on the stage 7.
An image of mark RR is formed thereon. The minute aperture member 8 has a combination of nine photoelectric detectors and five combinations of photoelectric detectors as optical direct conversion means for receiving the light sound that has passed therethrough and outputting a defective signal. surface (determined to approximately match the height of the surface of the wafer 10 on the d stage 7,
Therefore, the minute aperture member 8' and the photoelectric detector 9 are configured to move up and down together as the ueno holder 11 swings upward, as described above. For this kind of focusing,
A gap sensor 12 that measures the distance between the projection lens 6 and the surface of the Ueno 10 (or the aperture surface of the minute aperture member 8)
power: provided.

Automatic focus adjustment is possible with this gap sensor 12 and the vertical movement mechanism of the Ueno Holter 11.
When printing 50 circuit turns on the reticle 10, the height of the surface of the reticle 10 is detected so that a high contrast projection image can be transferred at all times.

On the other hand, the position of the stage 7 in the first orthogonal coordinate system However, in FIG. 2, only the laser interferometer 16 and the reflecting mirror 14 in the X-axis direction (horizontal direction in the paper) are shown; Needless to say, another combined force S of the laser interferometer and the reflecting mirror is similarly provided with respect to the direction (front and back directions of the page). These laser interferometers measure the positional coordinates of the stage 7 relative to the origin of the first orthogonal coordinate thread In this implementation @1, the origin of the coordinate system Power is distributed so that the intersection point of the measurement axis is located on the optical axis of the projection lens 60.

-i7' (The reticle holder 15 holds the reticle 5 and is movable two-dimensionally in the second orthogonal coordinate system xy in a plane parallel to the XY plane of the first orthogonal coordinate system.

The reticle 5 is positioned under drive control by a reticle alignment control system, which will be described later.

Further, shutters 4b and 4C arranged on both sides of the incident side of the second condenser lens 60 block only the incident light to the marks RR and KL of the reticle 5 in the illumination optical path from the illumination light source 1 to the reticle 5 when desired. It is intended for use only, and its placement position is not limited to the illustrated position.

As mentioned above, the reticle 5 has optically transparent marks RR and R.
Specifically, as shown in FIG. 6, these marks RR and RL are placed in the light shielding area around the ID'l path pattern area 5a of the reticle 5, for example, in the X-axis direction. This is a transparent slit provided. In the example shown in FIG. 6, the marks RR and RL are provided at two locations apart from each other on the X-axis of the second orthogonal coordinate system xy whose origin is the center of the reticle 5, but A similar translucent mark may also be provided.

Incidentally, since the image projected by the projection lens B of the pattern of the reticle 5 is an inverted image with respect to the xy coordinates, the directions of the coordinate system xy of the reticle 5 and the coordinate system XY of the stage 7 are reversed in FIG.

Therefore, when the reticle 5 is placed and fixed on the ocular holder 15, these Y-axes and the optical axis of the projection lens 6,
A plane including 6 axes including the Y axis can be defined, and the y axis and
If the optical axis of the projection lens 6 can be defined as another plane including six axes including the Y axis, the rotational position of the reticle 50 with respect to the stage 7 will be zero.

The writing stage 7 is provided with the minute aperture member 8 that detects the projected images of the marks RR and RL on the reticle 5, as described above. As shown in FIG. 4(a), this micro-aperture member 8 is made by depositing a chromium layer or the like on the entire surface of a circular glass plate, and forming a slit opening 8a in a part of the chromium layer. Figure 4(b) is A-AIfl in Figure 4(a).
The longitudinal direction of the slit opening 8a is determined to coincide with the X-axis direction of the first orthogonal coordinate system XY of the stage 7, and the width dimension of the slit opening 8a in the Y-axis direction is Mark RR, taking into consideration the reduction ratio of the projection optical system, etc.
Alternatively, RL is determined to be approximately equal to the width dimension in the Y-axis direction of the projected image of the mark when the image is formed on the aperture surface of the member 8.

As mentioned above, the origin of the first orthogonal coordinate system XY is on the optical axis of the projection lens B. The measurement of the position coordinates of the stage 7 by the laser interferometer in this actual m fil is the same as the normally known method, and is carried out in the X-axis direction and in the Y#i direction.
! The measurement is carried out in a digital manner by a digital counter in each laser interferometer in the direction. This counter increases or decreases its count value as the stage 7 moves.When turning on the power of the device, etc., the origin position of the stage 7 is determined and the stiffness counter is reset to zero at that origin position, or the counter value is reset to zero at that origin position. However, it is possible to orient the stage 7 to the origin position by using the slit aperture 8a. In other words, in this case, the output of the photoelectric detector 9 marks the slit aperture 8a. It is fully detected that the image of RR (or RL) coincides with the image of RR (or RL), and at this time, the counter (X
counter) to zero. More precisely, taking into account the rotational deviation of the reticle 5, the mark RRO image and the mark RL
X when the images of and respectively coincide with the slit opening 8a
From the count values Yr and Yt of the counter, the X counter (Y
After positioning the stage 7 in the Y-axis direction so that r + Yz)/2, reset the X counter to zero,
Performs the origin positioning of the Y-axis.

The same goes for the X-axis direction; although not shown, a pair of similar transparent marks are provided on the X-axis of the reticle 5, and another slit is provided in the minute aperture member 8 extending in a direction perpendicular to the suline 1-8a. An opening may be provided and the counter (X counter) in the laser interferometer in the X-axis direction may be reset to zero using the same procedure. In this way, the origin of the coordinate thread XY is set, and thereafter the position of the stage 7 is successively measured during two-dimensional movement using the position coordinates of the first orthogonal coordinate system XY with this origin as a reference.

Now, the apparatus of this embodiment includes an alignment microscope (WAM) used for positioning the wafer 10.
r Alignment Microscope)
are provided, and their arrangement is as shown in FIG. That is, this microscope detects a specific shape of the Ueno mark for alignment formed on the Uni... 10 as described later, and as shown in FIG. Six of the WL, WX and WR are fixedly arranged off axis. The first alignment microscope WL is for detecting the position of the sea urchin 10 in the Y-axis direction, and is arranged so that its detection center (observation center) is located on the Y-axis in the reference state. The second alignment a fine dWX is
110 in the X-axis direction, and is arranged so that its detection center (observation center) is located on the X-axis in the reference state. In this way, both alignment microscope holes 1 and W
The reason why the X detection center is made to coincide with the Y axis and the X axis is to eliminate Atsube errors when detecting the position of Ueno 10. The fifth alignment microscope WR is paired with the first alignment microscope WL to detect rotational deviation of the wafer 10, and in the reference state, the detection center (observation center) and the second alignment microscope) Th micro 4WL The device is arranged so that the line segment connecting the detection center of

These six alignment microscopes are used to print the first layer pattern and transfer it onto the street line inside or near the circuit pattern on the wafer for alignment of the second layer and subsequent exposure. A photoelectron microscope detects a short linear wafer mark by scanning it with a vibrating slit or vibrating beam of laser light.The detection center is set to coincide with the vibrating center of the vibrating slit or beam, and is particularly The optical system of the sixth alignment microscope WR is provided with a rotatable parallel plate glass or the like in order to slightly displace the detection center in the Y-axis direction.

The main structure of the control system for controlling the devices shown in FIGS. 2 and 6 is shown in the block diagram of FIG. The entire device is equipped with a microcomputer (including memory, etc.) to enable program control and various arithmetic processing.
It is centrally controlled by a CPU (CPU) 30. CPU30 is
A peripheral detection unit via an interface (IF) 31,
It collects and collects various information from the control section 11 or the drive section.

The shutter drive unit 62 opens and closes each shutter 4a+4b, 4c according to instructions from the CPU 50.

The reticle alignment control system (R-ALG) 33 C sets the reticle 5 at a predetermined position with respect to the optical axis of the projection lens B.
The reticle holder 15'f is moved so that the reticle holder 15'f is aligned. On the other hand, in order to measure the position value coordinates of the stage 7, the position information of the stage 7 in the X-axis direction read by the aforementioned Y-axis laser interferometer 16 and the position information read by the Y-axis laser interferometer 64 are used. Stage 7 Y
The axial position information is also sent to the CPU 60 via the interface 31'&. In addition, in order to move the stage 7 two-dimensionally, an X-axis drive unit (X-ACT) 35 that drives the stage 7 in the X-axis direction and an ) 56 is the CPU
A θ-axis rotation drive unit (θ-ACT) 37 for slightly rotating the wafer holder 11 on the stage 7, and a sea urchin...
A Z@ drive unit (Z-
ACT) 38 is provided and is configured to operate according to instructions from the CPU 30. The photoelectric output signal of the photoelectric detector 9 is also sent to the CCPU 60 via the interface 61.
In addition, the focus detection unit (AFD) 39 receives the signal from the gap sensor 12 shown in FIG. The information on the deviation (focus information) is provided to the CPU 60 via the interface 51. Note that an output terminal device 40 such as a CRT display for monitoring or a printer is also connected to the CPU 30 via an interface 61 to display measurement results, operating conditions, and the like.

In addition, as shown in Fig. 6, the off-axis alignment microscopes WL, WX, and WR are photoelectron microscopes or laser spot scanning type wafer mark detection means, and 6C1 is the detection center of each of them. When the mark matches a predetermined wafer mark on the wafer 10, a mark detection signal is provided to the CPU 30 via the interface 31. In addition, sea urchin and
- It goes without saying that the matching point information from the gap sensor 12 and the focus detection section 69 is input to the CPU 30 when detecting a mark.

In the alignment microscope wa, a parallel flat glass plate and a prism are controlled by the commands of the CPU 30 in order to shift the imaging center of the oscillating slit or the oscillating beam and the light transmission path of the laser beam by a minute amount in the Y-axis direction. It is rotatably or movably provided.

In the above configuration, the reticle 5 is placed on the reticle holter 15, and the reticle alignment control system 66 performs alignment of the reticle 5. When the pattern of the reticle 5 is projected onto the wafer 10, the optical image is approximately It will look like the figure 6.

FIG. 6 shows a projected image 5a' of the pattern area 5a of the reticle 5 with respect to the first orthogonal coordinate XY of the stage 7 to be measured with the laser interferometers 13 and 64.

The marks RR and RL on the reticle 5 are also projected as images RR' and RL' on both sides of the projection image 5a' along the X-axis, respectively. Note that here, the internal coordinate system x of the projected image of the reticle 5
The origin of y is made to coincide with the origin 0 of the orthogonal coordinate system XY of the stage 7, and the distortion caused by the projection lens B is treated as negligible. Normally, the second orthogonal coordinate thread xy has an overlay error due to rotational deviation with respect to the first orthogonal coordinate system It represents. In one embodiment of the present invention, instead of adjusting the angle C to be negligible/JQ small, the stage 7 is controlled to move so that the printed pattern is substantially unaffected by the angle e. This is intended to eliminate overlay errors due to rotational deviation.

In FIG. 6, mark images RL' and RR' are on the reticle 50.
It is assumed that the y-axis coordinate values of some of the projected images are the same. It is also assumed that the projection plane level of the reticle 5 coincides with the aperture plane level of the minute aperture member 8, and that the member 8 moves two-dimensionally within this projection plane together with the movement of the stage 7. Here, when discussing the Fuchi 1 law of angle 6, first,
The gap sensor 12 and focus detection units 69 and 2
Focusing is performed using the shaft drive section 68. Thereafter, the stage 7 is moved to scan the mark images RR' and RL' with the slit aperture 8a, and the output of the photoelectric detector 9 and the outputs of the laser interferometers 16 and 64 obtained during this scanning are used to
The positions of mark images RR' and RL' in the first orthogonal coordinate system XY are measured. For example, in FIG. 6, the slit opening 8a moves in the Y-axis direction and the mark image RR' or RL'
When scanning, the photoelectric detector 9 shows a peak value at the moment when the slit opening 8a and the mark image match, so the laser interferometer 16 and 3 detect the peak value when this peak value is obtained.
You can measure it with 4. The Y coordinate value of the mark image RR' thus measured is YRI, and the Y coordinate value of the mark image RL' is YL, and the mark images RR' and RL' obtained from the amount of movement of the stage 7 in the X-axis direction are If the interval is t, the angle ε
Since is usually a minute angle, it can be expressed as the following equation (1).

a = jan-1(YRYL)/l'''torture CYR-
YL)/l ......A... (1) Also, if the distance between the marks RR and RL on the reticle 5 is known in advance, the X axis of the projected images RR' and RL' direction spacing t
′ is also known, the angle ε can also be expressed as in the following equation (2).

t = sin ” (YR-YL)/l'ζ (
YR-YL)/l'...Song...Song (2) In this way, the angle ε is obtained in advance, and the angle ε is calculated in advance for the lateral movement of stage 7 in the step-and-repeat exposure operation. Used as direction correction information.

In the above explanation, in FIG. 6, mark images RR' and R
The farther L' is from the circuit pattern area image 5a', the more accurate the measurement of the amount of rotational deviation (angle C) is. In addition, the slit opening 8a' of the minute opening member 8 and the mark image RR'
As the length of the slit shape of and RL increases, the amount of detected light increases, and the +141i determination accuracy improves. Here, if the length of t or t' is made longer, the mark image R
Since R' and RL' overlap the image of the adjacent circuit pattern area that is repeatedly exposed, in such a case, only the light incident on the mark images RR' and RL' from the illumination light source to the reticle 5 is transmitted by the shutters 4b and 40. The shutters 4b and 4c are opened only at regular times of the rotational deviation fff amount (angle ε), and are closed when the circuit pattern is exposed.

Incidentally, if the accuracy of suppressing the amount of rotational deviation does not need to be so high, the X-axis in the circuit pattern area image 58' can be A flat line, for example, both ends of the image of the boundary line between the pattern area and the surroundings (il in Figure 6, i2 position measurement is il,
This is carried out using the measurement value of the laser interferometer 16° 34 at the center of the rise or fall of the output signal of the photoelectric detector 9 when the slit opening 8a crosses the bright/dark boundary of i2.

Next, an exposure operation using the apparatus of this embodiment will be explained.

First, the first circuit pattern on the wafer 1o is
Exposure operations when printing layers will be explained below with reference to FIGS. 7, 8, and 9.

In this case, since neither a circuit pattern nor a wafer mark for alignment exists on the surface of the wafer 1o yet, the wafer 1o has a linear cutout partially constructed on its outer periphery as shown in FIG. facet (or franc)
1[]a as a reference, it is placed on the wafer holder 11 and fixed by suction.

In this embodiment, the facet lOa is positioned so that the @ line direction coincides with the X-axis direction of the stage 7. Next, the surface of the wafer 1o is focused on the projection image plane by the gap sensor 12, the focus detection section 69, and the two-axis drive section 68. After that, the laser interferometers 16 and 64, the X-axis moving part 6
5 and the Y-axis moving part 66, the shutter 4a is opened for a predetermined period of time, and the reduced projected image of the pattern area 5a of the reticle 5 is exposed and transferred onto the photoresist on the wafer 10. Repeat. In this case, the shutters 4b and 4c are closed and the marks RR and RL are closed.
image is prevented from being transferred onto the wafer 10. In this way, the reduced image of the pattern area 5a is transferred and printed in a matrix on almost the entire surface of the wafer 10.

Here, if step-and-repeat exposure transfer is performed while the reticle 5 has a rotational deviation of the above-mentioned angle, as the stage 7 moves in the X-axis direction,
As shown in FIG. 7, the line segment connecting the centers of pattern areas Po + PH, which are successively transferred to sea urchins I and 10, is X1i.
The individual pattern regions of those located on l have a deviation angle ε
The image is transferred in a rotated state with respect to the X axis (or Y axis).

Therefore, in this embodiment, another orthogonal coordinate system αβ (hereinafter referred to as the sixth orthogonal coordinate system) rotated by the angle ε with respect to the XY coordinates is created in the CPU 30 using the angle ε determined in advance as described above. are set on the wafer 10 with the origin coincident. Note that when transferring onto the wafer 10, it is not necessary to make the origins of both the orthogonal coordinate systems XY and αβ coincide with each other for the first layer, but for the second and subsequent layers, it is essential that the origins coincide. .

Now, after transferring one pattern area Po, when moving the stage 7 in the X-axis direction to transfer the next pattern area P1, by aligning the advancing direction of the stage 7 along the α-axis, the next pattern area The transfer pattern area is P2 on the 2nd floor.
This pattern area P2 and the pattern area Po
As can be seen from the arrangement, the rotational deviation (angle ε) is substantially canceled by using the sixth orthogonal coordinate thread αβ as a matrix array of the transfer pattern on the wafer 1o.

Therefore, the array coordinates within the wafer 10 are set in the sixth orthogonal coordinate system αβ
The pattern area exposure position is determined along the coordinate axis αβ, and the stepping position is given to the stage 7 by coordinate transformation between the third orthogonal coordinate system and the first orthogonal coordinate system. Make it.

Now, as shown in FIG.
2, and the center of the pattern area P2 is 02. 1 is the center O of the pattern area Po in the 611th angular coordinate thread αβ
The coordinate value of (C0, β0) is the center 0 of the pattern area P2.
Let the coordinate values of 2 be (αl, βl), and the array coordinate system αβ on the sea urchin z-iQ with respect to the first orthogonal coordinate XY of the stage 7 whose origin is the optical axis of the projection lens 60, that is, the center of the pattern area Po. After rotating counterclockwise by angle #ε, the origin O of the coordinate thread αβ becomes the coordinate value (Xo, Yo
).

The exposure of the pattern area Po is completed, and the center 02 of the pattern area P2 and the origin O of the coordinate system XY (the light 11 of the projection lens 6
11), the stage 7 should be moved from the current position (Xo, Yo) by ΔX in the X-axis direction and ΔY in the Y-axis direction as shown in FIG. .

Here, ΔX and ΔY are ΔX= <C1-C0) possible ε-(β1-β0) self εΔ
It is expressed as Y=(C1-αo)sfng+(β1-IO)houseε, and if the angle C is small enough, ΔX=(C1-αG
)−(β!−β0)ε・・・・・・(3)ΔY=(C1
−α. )-(β1-β0)...(4) can be approximated.

Equations (3) and (4) are programmed into the CPU 30 shown in FIG. 5, and the center position of each pattern area to be exposed and transferred is stored in advance as the coordinate value of the array coordinate thread αβ. , using the angle e detected earlier (
3) Perform the calculation of equation (4), and control the stepping movement of the stage 7 so that the angle C is canceled out as a result.

This stepping is based on the current position of stage 7 (Xo =
This is done by moving the position of the stage 7 so that the counter measurements of the laser interferometers 16 and 64 are changed by ΔX and ΔY with respect to Ueno 1.
In order to expose and transfer a line parallel to the stone line of the facet 10a of 0 using the step-and-repeat method, set IO and βl to be equal values among the center coordinate values of the pattern area, and α
If l-α0 is the pitch Ep, from equations (3) and (4), △
Stepping in stage 7 may be performed so that X = Ep △y = Ep ε. For example, if 9 pattern areas are transferred in a row, 7
As shown in FIG. 9, this row of pattern areas E1 to E9
The respective centers Cl-C9 of are all carried on the α axis, and all of the pattern areas E1 to E9 are aligned without any rotational deviation with respect to the coordinate system αβ, and only the facets N[la The α-axis is only inclined by an angle C with respect to the straight line. In this way, by exposing and transferring the pattern areas in the second and third rows while arranging them on the coordinate system αβ in the same way, the entire surface of the wafer 10 is free from rotational deviation as shown in FIG. Transfer patterns aligned according to the array coordinate system αβ can be obtained. Furthermore, this first
As described above, by the #th transfer, the sea urchin marks for alignment of overlapping in the second and subsequent layers are transferred onto the street lines within or in the vicinity of each pattern area.

In this way, even if the reticle 5 is set under a setting condition that involves a rotational deviation of the angle ε, the pattern area of the first layer transferred to the sea urchin 10 is substantially affected by the angle C. I never receive it.

Next, the overlapping exposure transfer of patterns in the second and subsequent layers will be explained with reference to FIGS. 10 and 11.

FIG. 10 shows the arrangement relationship of the projection lens 6 and the alignment microscopes WL, WX, and WR shown in FIG.
In the explanatory diagram shown on the Y plane, the center of the projection area 6b of the reticle pattern area 5a within the maximum exposure area 6a of the projection lens 6, that is, the optical axis of the projection lens 6, coincides with the origin of the first orthogonal coordinate XY. ing. Each alignment microscope @WL, W distributed with power at a predetermined position around the projection lens 6
X, WR are indicated by dashed circles in each field of view), and each detection center is indicated by LCIXCs B-C.
The distance between LC and RCOX in the axial direction is shown as L.

FIG. 11 is a plan view schematically showing the wafer 10 on which the first layer pattern has already been transferred, and only depicts a row of pattern areas E, + to E6 in the wafer 10 and their vicinity. . Pattern area E.

Elongated wafer marks for alignment have already been formed on the upper, lower, left and right street lines of ~E6 along each axis of β in the array coordinate system. Here, among these wafer marks, two wafer marks AL and AR are spaced apart from each other by approximately L on one street line parallel to the α axis.

In addition, one wafer mark AX on one street line parallel to the β axis is used for alignment of the wafer 10 for overlapping exposure of the second and subsequent layers.

Now, prior to the exposure of the 21st pattern, a reticle 5 with a circuit pattern different from that in the first layer is mounted on the reticle holder 15. Therefore, it is necessary to consider the rotational deviation due to the reticle again. Therefore, when performing pattern exposure for the second layer, stage 7 is first exposed.
is moved to detect the projected images of the transparent marks RR and RL on the reticle 5 using the minute aperture member 8 and the photoelectric detector 9;
Find its position coordinates using a laser interferometer 13.34,
The rotational deviation is detected as an angle ε' according to equation (1) or (2).

Next, the alignment microscope ψWL is the detection center LC and R of WR.
In this adjustment, as shown in Fig. 10, the rain detection center LC of alignment m fine sharp WL and WR is
The line segment connecting RC and RC is made to be inclined by the angle ε' with respect to the X axis on the XY coordinate plane. For this purpose, the detection center RC of the alignment microscope WR may be changed in the Y-axis direction from the rotational deviation angle ε' of the second layer reticle by the displacement ay determined by δy#L·ε'. this is,
Specifically, fine adjustment is possible by rotating or shifting optical components such as burping glass and prisms inside the optical system of the alignment microscope WR. In this case, to displace the detection center RC by a predetermined amount, another detection means for detecting the rotation angle of the burping glass or the amount of movement of the prism or the like may be provided and the detection signal may be used. , it is more convenient to use the laser interferometers 13 and 34. This laser interferometer 13
．． 34, the slit opening 8a of the positioning minute aperture member 8 provided around the stage 7 is used to locate the detection center LC of the first alignment microscope and QWL.
The stage 7 is positioned so that the center of the 111th slit h opening 8a in the YI11 direction coincides, and the laser interferometer 64 calculates the Y coordinate value y v-J at that time.
Measure f. Next, the stage 7 is moved in the X-axis direction by a distance so that the same slit opening 8a can be detected by the alignment microscope @WR. At this time, the stage is positioned so that the Y-axis coordinate measurement value by the laser interferometer 64 becomes (y-1+δy), and then the alignment microscope WR is moved so that the detection center RC and the Y-coordinate value of the center of the slit opening 8a match. Rotate or move optical components within an optical system.

When the adjustment of the alignment microscopes fiWL and WR is completed as described above, the way i o (on which the first layer transfer pattern is formed) is placed on the holder 11,
Fix by suction using a vacuum, etc. In this case, the wafer 10 has 7
10a to roughly position it on the holder 11. After that, two wafer marks AL and AR formed on one street line on the sea urchin 10 at a distance apart are aligned at the alignment fine boundaries WL and WR, respectively.
The stage 7 is positioned so that it is within the detection field of view.

In this positioning, since the position of the wafer 10 with respect to the minute aperture member 8 on the stage Z is approximately determined, the detection centers LC and RC of the alignment microscopes WL and WR are aligned with the slit opening 8a of the minute aperture member 8 in advance. This can be easily done by storing the coordinate values of the stage 7 when they match.

Now, the two kings on wafer 10... marks AL and AR.
is captured by the alignment microscopes WR and WL, the detection center LC of the alignment a micro-sharp WL matches the wafer mark AL, and the wafer mark is aligned so that the detection center RC of the alignment a micro-shape WR matches the wafer mark AR. - The holder 11 is rotated by the θ-axis rotation drive unit 67.

At this time, the rotation center of the wafer holder 11 is set at the wafer mark A.
If it is set near L, when the holder 11 is rotated,
Sea urchin from the detection center LC of alignment microscope dWL...
The amount of deviation of the mark AL becomes extremely small. In this case, the wafer holder 11 is slightly rotated or the stage 7 is slightly moved in the Y-axis direction so that the deviation of the Ueno mark AR in the Y-axis direction from the detection center RC of the alignment microscope WR becomes substantially zero. Just do it. When the rotational position of the sea urchin I/10 with respect to the stage 7 is determined in this manner, the holder 11 is fixed to the stage 7.

As described above, the array coordinate system αβ in the Ueno 10 is positioned at an angle ε' with respect to the XY coordinate axes of the stage 7 as shown in FIG. ' and sea urchin... Each of the pattern areas E1 to & of the first layer already existing on the 10 is set to be tilted by an angle ε' in the counterclockwise direction with respect to the XY coordinate axis. Stage 7 is when the positioning of the wafer 10 is completed and the wafer mark AL matches the detection center LC of the alignment microscope WL.
Based on the X coordinate value of CP
U30 is the pattern area P. Then, the positional relationship of the center with each pattern area El to E6 on the wafer 10 is determined, and the stage 7 is moved so that the pattern area PG' overlaps El, and the shutter 4a is opened for exposure. Next, the stage 7 is moved stepwise in order to sequentially overlap the pattern area Po' with each of the first layer pattern areas below E2. Once positioned, the rotational deviation of the angle ε' of the reticle 5 is canceled out in all pattern areas on the wafer 10, so that the pattern areas of the first layer and the second layer are overlapped with each other without any rotational deviation. become. below,
For the third and subsequent layers, alignment microscopes WL and W
By adjusting R, step-and-repeat printing is similarly performed without occurrence of rotational deviation.

In addition, in the above-mentioned example, the alignment microscope @l'L.

The WR was adjusted using the minute aperture member 8 and the photoelectric detector 9, but if there is a wafer mark AL on the wafer 10, the same adjustment can be made using that mark. in this case,
First, the stage 7 is moved in the X-axis direction after aligning the wafer mark AL with the detection center LC of the alignment microscope WL, and when the alignment microscope WR captures the wafer mark AL within its detection field of view, the stage 7 is stopped and the detection The detection center RC is positioned so that the center RC and the wafer mark AL coincide. As a result, the line segment connecting the detection centers LC and EC of both alignment microscope boundaries WL and WR is
become parallel to the axis. After that, the previously determined angle 6'
The stage 7 is moved and positioned by δγζLε' in the Y-axis direction based on etc. may be displaced.

Furthermore, although the above-mentioned embodiment is applied to a projection exposure apparatus, the present invention is not limited to this; a mask and a wafer are made to face each other through a minute gap, and exposure light, X-rays, etc. are applied from the mask side. irradiate the mask, transfer the pattern image of the mask onto the wafer, and then move the mask a certain distance in parallel.
Needless to say, the present invention is similarly applicable to a so-called proximity type step-and-repeat exposure apparatus that repeats the process.

As described above, according to the present invention, the rotational deviation of the reticle relative to the wafer movement system can be determined quantitatively and accurately in a short time, and the direction of movement can be corrected prior to exposure. It can effectively prevent patterns from being misaligned with each other due to rotational deviation of the reticle, and by adjusting the alignment microscope only once for each reticle, rotational deviation of the reticle will not remain during subsequent stepping. Therefore, the prevention of overlay errors due to rotational deviation is extremely
This can be accomplished with a short amount of adjustment. In addition, in the present invention, the image of the reticle pattern projected by the projection optical system for exposure is directly detected by a photoelectric detector on the stage to determine the rotational deviation of the reticle pattern image with respect to the stage movement coordinate system. Since only the projection optical system, which inherently reduces field curvature and distortion, is used and no other detection optical system is required, the detection of rotational deviation itself is extremely accurate and can be done in alignment with the projection optical system. Therefore, the accuracy of the correction for canceling the rotational deviation based on the detected amount of rotational deviation can be checked to make it highly accurate.

[Brief explanation of the drawing]

FIG. 1 is an explanatory diagram showing the rotational deviation between superimposed patterns, and FIG. 2 is a configuration diagram showing the outline of a reduction projection type exposure apparatus as an example to which the rotational deviation detection apparatus of the present invention is applied. , Figure 6 shows the position of the transparent mark on the reticle and the arrangement of the six alignment microscopes around the projection lens.
A schematic perspective view shown in relation to the orthogonal coordinate thread XY and the second orthogonal coordinate thread xy, FIG. , FIG. 5 is a block diagram showing the configuration of the exposure apparatus control system according to the embodiment, and FIG. 6 shows the projected image of the reticle pattern on the wafer with the origins of the first and second orthogonal coordinate systems aligned. An enlarged view, FIG. 7 is an explanatory diagram showing the rotational deviation of adjacent transfer patterns and their cancellation when exposed one after another as the stage advances, and FIG. 8 shows the positioning coordinate transformation when the stage advances. An explanatory diagram, Fig. 9 is an explanatory diagram of the transfer pattern arrangement on the sea urchin showing that the rotational deviation of the first layer transfer pattern is canceled out, and Fig. 10 is an explanatory diagram of the transfer pattern arrangement on the surface of the projection lens and each alignment microscope boundary. An explanatory diagram showing the arrangement relationship on the coordinate XY plane, Figure 11 is a schematic plan view of the sea urchin to which the first layer pattern has already been transferred, and Figure 12 is the second layer pattern.
It is an explanatory view showing superposition of patterns of layers. 1: illumination optical system, 2.3: condenser lens, 4a,
4b, 4c: Shutter, 5th reticle, 5a
: pattern area, 6: projection lens, 7: stage, 8: minute aperture member, 8a: slit aperture, 9: photoelectric detector, 10: wafer, 11: wafer holder, 12: gap sensor, 16: laser interferometer , 14: Reflector, 15 Niticle Holter, WL, WX, WR Near Alignment) Microscope, 60
: Microcomputer (CPU), 61 coin interface, 62: Shunter drive unit, 66: Reticle alignment control system, 64: Laser interferometer, 3
5: X-axis drive section, 36: X-axis drive section, 67: θ-axis drive section, 68: Z-axis drive section, 69: Focus detection section, R
R%RL: Transparent mark% RR', RL': Mark image, AL A AX: Uni/) mark, Es~E9
; Transferred pattern area. Agent Saburo Kimura-7: Rice j figure l, +2 figure

Claims (1)

[Claims] a stage capable of two-dimensional movement in the directions of both coordinate axes of an orthogonal coordinate system forming a moving plane on which an exposed object to which a mask pattern is transferred; photoelectric detection means for detecting a predetermined local portion; position detection means for detecting the position of the optical g# local portion on the orthogonal coordinate thread detected by the photoelectric detection means as the stage moves; : A rotational deviation detecting device for an exposure apparatus based on positional information detected by the position detecting means for the optical image local portions at several different locations.