JPH08264431A - Scanning projection aligner - Google Patents

Abstract

PURPOSE: To correct the nonlinear image deformation perpendicularly to the lengthwise direction of a slitted exposing light by correcting the magnification in such a manner that the displacement of formation position is corrected perpendicularly to or at a specified angle to the lengthwise direction of a nonlinear slitted exposing light. CONSTITUTION: A pattern on a mask 11 is irradiated with a slitted exposing light emitted from a lighting optical system, and the picture passes through a projection optical system 13 and parallel plane plate 14, then it is formed on a photosensitive substrate 10 at a magnification set in the system 13. While being synchronized with scanning of the mask 11 at a constant speed V in a direction of an arrow mark 15, the substrate 10 is scanned at a constant speed V/β (β: reducing magnification of the system 13) in a direction indicated by arrow 16. In addition, the plate 14 is bent in such a manner that the tilt of the plate 14 is proportional to its Y coordinate, so that the error of nonlinear element in a Y direction on the slit can be thoroughly corrected, and the tilt amount of the plate 14 is increased by increasing the driving quantity of pulse motors 19a and 19b.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a scanning projection exposure apparatus used when manufacturing a semiconductor integrated circuit, a liquid crystal display element or the like in a photolithography process.

[0002]

2. Description of the Related Art Conventionally, when a semiconductor element, a liquid crystal display element, or the like is manufactured by using a photolithography technique, a pattern of an original plate (hereinafter referred to as a mask) such as a photomask or a reticle is passed through a projection optical system. Although a projection exposure apparatus for exposing and transferring to a wafer or a light-transmitting plate coated with a photosensitive material such as photoresist is used, the pattern size of one element tends to increase in recent years. Therefore, a technique for transferring and exposing a pattern of a larger area on a mask onto a photosensitive substrate is required for the projection exposure apparatus.

In order to meet the demand for such a large area,
The mask and the photosensitive substrate are relatively synchronously scanned at a speed proportional to the projection system magnification with respect to a rectangular or arcuate illumination area having a constant width (hereinafter referred to as slit exposure area). The so-called scanning exposure type exposure apparatus has been developed in which a mask pattern having a larger area than the slit-shaped exposure area is exposed on the photosensitive substrate by scanning). It goes without saying that this scanning exposure type exposure apparatus also requires the high level of overlay accuracy required for the conventional exposure apparatus.
In general, the circuit structure of the semiconductor element and the liquid crystal display element is not a single layer, and it is necessary to perform overprinting with high dimensional accuracy using different masks for each layer based on the pattern formed in the first lithography process. Therefore, by forming at least two alignment marks on the photosensitive substrate and measuring the amount of deviation from the reference point provided for each device,
The photosensitive substrate is aligned and exposed.

However, at present, there is a demand for a technique for exposing a pattern having a larger area on a mask onto a photosensitive substrate. Due to a slight shift in image forming position, thermal expansion and thermal contraction of the photosensitive substrate, etc. Even if the matching is performed, a deviation occurs at other points. The occurrence of this phenomenon is referred to as the occurrence of distortion.

The distortion mode is as follows.
As shown in FIG. 8, a simple magnification error expressed as a linear function of XY coordinates provided on the photosensitive substrate 10 or a component of second or higher order of XY coordinate values called a non-linear component error. There is an error. Regarding the magnification error in the conventional scanning exposure apparatus, not only the distortion of uniform magnification but also the X direction and the Y
Various correction methods have been invented and disclosed for so-called bias magnification errors, which have different magnifications in each direction.

In the scanning exposure apparatus, the non-linear component error is expressed as a function of the position in the direction in which the mask and the photosensitive substrate are synchronously scanned at a speed difference proportional to the magnification of the projection optical system (hereinafter simply referred to as synchronous scanning). There is something to be done. Regarding this non-linear component error, when the mask and the photosensitive substrate are synchronously scanned, the mask or the photosensitive substrate is subjected to X, Y,
By driving a small amount corresponding to the non-linear component in the θ direction relative to the synchronous position, not only the magnification error but also the non-linear component can be corrected.

As shown in FIG. 2, the slit-shaped exposure area 1
When the center line in the width direction of 2 is the W coordinate axis, the direction of the synchronous scanning is the Y coordinate axis, and the coordinate axis orthogonal to the Y coordinate axis on the photosensitive substrate is the X coordinate axis, and the coordinate axis is set on the photosensitive substrate, the above-mentioned correctable image position shift The components ΔX, ΔY, and Δθ are expressed by the equation 1 on the entire surface of the photosensitive substrate. However, the Y coordinate value is assumed to be the absolute value of the distance from the W axis in the direction parallel to the Y coordinate axis.

[0008]

[Equation 1]

[0009]

However, the non-linear component of the projection optical system and the component expressed as a function of the position along the W coordinate axis, that is, the direction in which the exposure light is simultaneously exposed by the projection optical system and the slit-shaped exposure light, is the same as that of the projection optical system. Correction is impossible unless the distortion characteristics can be changed nonlinearly. Therefore, in order to reduce the maximum value of the positional deviation by distributing the non-linear component, the offset is obtained each time by measurement, or the offset is registered as a device constant in advance, and the overprinting is performed while the imaging position deviation is distributed. Is going.

In order to specifically explain the component represented by the function of the position along the W direction in FIG.
Expressed as a formula. However, the W coordinate value is assumed to be equal to the X coordinate value.

[0011]

[Equation 2] In the above equation, the terms D ₀ and E ₀ are mere positional deviations and can be corrected by relatively moving the mask or the photosensitive substrate before exposure. Also D ₁ · W + D ₂ · W ² + · ·
.. + term D _n · W ⁿ are those which can be thought of as broad lateral magnification, for this, there are known various correcting method.

Further, the term of E ₁ · W represents the θ component, which can be expressed by the equation 2 if the distance between the rotation center and the printing position is L after the θ component is corrected. By correcting the X position during the scanning exposure by the correction amount, it is possible to correct by a simple θ rotation.

[0013]

(Equation 3) Further, when the X position is closed by the interferometer or the like, the X-direction distance measuring mirror is rotated by θ rotation so that the laser beam is extended on the extension line of the slit-shaped exposure area. If the optical axis is placed, the correction in the X direction is not necessary. That is, conventionally, the component that cannot be corrected is the component represented by the equation 4 of the equation 2.

[0014]

[Equation 4] It is an object of the present invention to provide a scanning exposure apparatus having means for correcting an even function component of this component.

[0015]

To achieve this object, in the present invention, an illumination optical system for illuminating an original plate on which a transfer pattern is formed by exposure light having a slit shape whose cross section is not linear. And a projection optical system for projecting the image of the transfer pattern illuminated thereby on the photosensitive substrate at a predetermined magnification, and in the scanning direction forming a right angle or a constant angle with respect to the longitudinal direction of the illumination area by the exposure light. A scanning projection exposure apparatus including a moving unit that synchronously moves the original plate and the photosensitive substrate at a speed ratio according to the projection magnification, and a magnification correction unit that corrects the magnification of the projection optical system at least in the scanning direction. Of the image forming position shift in the scanning direction when the image is transferred by being superposed on the image transferred and transferred onto the photosensitive substrate. In the correction of the relative position and corrects an inability correction component.

[0016]

First, a parallel plane plate having an optical thickness that does not substantially affect the image forming performance other than the deviation of the image forming position is inserted in the optical path of the projection optical system to set the lateral magnification in a broad sense. A method of correcting the deviation of the image forming position in the correcting scanning exposure apparatus will be described.

In FIG. 3, the thickness of the inserted optical plane-parallel plate 14 is d, the refractive index is n ', and the X-axis is the direction in which the image formation position on the photosensitive substrate 10 is to be laterally displaced, and is perpendicular to this. When the direction of the chief ray is the Z axis, the amount of bending of the inserted plane-parallel plate 14 in the Z direction is expressed as a function of X.
The lateral shift amount ΔX of the print image in the X direction obtained at this time is expressed by the equation (5).

[0018]

(Equation 5) That is, the shift amount of the image forming position is proportional to the inclination of each part of the plane parallel plate 14. Therefore, when a general lateral magnification is generated, the inclination of the plane-parallel plate 14 is corrected by bending the plane-parallel plate 14 in proportion to the X coordinate. In other words, the deflection of the plane-parallel plate 14 is corrected. The amount may be bent so that it is proportional to the square of X.

When the image forming position is displaced in proportion to the nth power of X as a lateral magnification in a broad sense, the plane parallel plate 14 is (n +) of X.
1) The shape of the plane-parallel plate 14 or the holder shape of the plane-parallel plate 14 may be devised in terms of material mechanics so as to have a bending proportional to the power.

In the scanning exposure type exposure apparatus, it is necessary to bend the plane-parallel plate 14 around the axis of the mask and the photosensitive substrate in the synchronous scanning direction indicated by the arrow 36 as shown in FIG. As described above, it is effective as a means for correcting the image distortion represented by the equation (2), and this fact is well known. However, bending the plane-parallel plate 14 around the X coordinate axis as shown in FIG. 5 has no effect on the correction of the image distortion and is not considered.

In the scanning projection exposure apparatus of the present invention, with respect to the Y-direction shift represented by the function of the position in the direction (X direction) in which the projection optical system and the slit-shaped exposure light are simultaneously exposed and irradiated, the slit-shaped exposure light The shape is Y = F
By changing to a shape represented by an expression in which (X) and F (X) is an even function, the even function component is corrected by bending the optical parallel plane plate 14 around the X axis.

The shape of the slit-shaped exposure light is represented by the equation (6), and the slit shape is represented by Y = F (X) in calculation.

[0023]

(Equation 6) Here, it is assumed that the Y-direction image distortion of the projection system with respect to the slit-shaped exposure light is ΔY = G (X) with respect to the absolute reference grating. That is, a point on the slit represented by Y = F (X) is imaged on the exposure substrate in a state shifted by ΔY = G (X) from the regular image formation position.

In the scanning exposure type exposure apparatus, among the image distortions caused by the projection optical system corresponding to the slit-shaped exposure light, the non-linear component represented by the equation 4 is a non-uniform deformation of the photosensitive substrate and the mask. When the deformation of the device structure due to scanning is negligible, the distortion does not become a function of the position of the mask and the photosensitive substrate in the synchronous scanning direction, but becomes constant, and the distortion is represented by FIG.

When the distortion as shown in FIG. 6 occurs in the actual process, the alignment target positions provided in the previous photolithography process of the photosensitive substrate are measured at multiple points, and the deviations are normally distributed. Although the distortion will be as usual, the variation itself does not become small, and it is necessary to correct this variation.

Since the odd function component of the Y direction non-linear component expressed by the equation (4) can be approximated by the θ component, the error component can be approximately corrected by the θ correction. However, the variation of the even function component cannot be corrected by θ correction.

Next, a method of determining the bending shape of the optical plane-parallel plate for correcting the even function component of the equation (4) will be described. To determine the bending shape of an optically parallel plane plate,
A function approximating the Y direction non-linear component must be determined first. Now, the Y direction nonlinear displacement on the slit is

[0028]

(Equation 7) Is approximated by. At this time, the slit shape

[0029]

(Equation 8) The shape represented by And

[0030]

[Equation 9] F (X) or G (X) is determined so that That is, by substituting the equation 8 into the equation 9 and further substituting the equation 7
Substituting into the formula, the formula 10 is obtained.

[0031]

[Equation 10] Equation 10 shows that the correction is realized by shifting the image forming position in the Y direction in proportion to the Y coordinate value on the slit. Therefore, when the slit-shaped exposure light has a shape as in Expression 9, correction is realized by inclining the plane-parallel plate 14 in proportion to the Y coordinate. Therefore, by substituting the ΔY of the expression 10 into the ΔX of the expression 5 and rewriting dZ / dX into dZ / dY, the expression 11 is obtained.

[0032]

[Equation 11] It suffices to bend the plane-parallel plate 14 in the shape of the formula (11).
In order to realize this bending, the plane-parallel plate 14 or its holding member is made sufficiently larger than the rigidity of the plane-parallel plate 14, and the shape of the plane-parallel plate 14 is a cantilevered equal stress in terms of material mechanics. The beam may be applied with a concentrated load or forced displacement at the free end, or an equal stress beam with both ends supported and a beam center load or forced displacement may be applied. However, if the correction error is not a practical problem,
It may be a simple cantilever beam or a beam supporting both ends.

Further, as described above, if the magnification in at least one direction can be changed in the projection optical system instead of performing the correction by causing the image shift in the plane-parallel plate 14, the slit-like exposure which is not linear is possible. Similar correction is possible by combining with light.

[0034]

Embodiments of the present invention will be described below with reference to the drawings. FIG. 1 is a perspective view showing a scanning projection exposure apparatus according to an embodiment of the present invention. In FIG. 1, 10 is a photosensitive substrate,
11 is a mask, 12a to 12d are slit exposure light, 1
3 is a projection optical system, and 14 is an optical plane-parallel plate. A pattern on the mask 11 is illuminated by slit-shaped exposure light from an illumination optical system (not shown), and its image passes through the projection optical system 13 and the plane-parallel plate 14, and is projected on the photosensitive substrate 10 at a magnification of the projection optical system 13. It is designed to form an image of a size that suits you.

Then, in synchronization with the scanning of the mask 11 in the direction of arrow 15 at a constant speed V, the photosensitive substrate 10 is moved in the direction of arrow 1.
Scanning is performed in six directions at a constant speed V / β (β is a reduction magnification of the projection optical system 13). The mask 11 and the photosensitive substrate 10 are respectively held by a mask stage and a photosensitive substrate stage (not shown), and the X, Y, θ positions of both stages are monitored by a laser interferometer (not shown). Based on the closed loop feedback servo, the position coordinates are managed.

The closed loop feedback servo may be performed only on the photosensitive substrate stage, and the servo position of the photosensitive substrate stage may be corrected based on the position information of the mask stage. FIG. 9 shows a reference mark for alignment with the photosensitive substrate 10 (hereinafter, referred to as a reference mark) arranged on the mask 11.
An arrangement of mask side AA marks) 25a to 33a is shown. On the photosensitive substrate 10 side as well, the alignment target mark (hereinafter referred to as the plate side AA mark) 25b corresponding to the mask side AA mark is provided at the common position via the projection optical system 13.
33b is provided in the previous photolithography process. Normally, these marks are not allowed to be arranged inside the element pattern 34 provided on the mask, so that they are arranged so as to surround the periphery of the element pattern 34.

Reference numerals 24a and 24b in FIG. 1 are microscopes in which the span of the observation position can be changed along the area irradiated by the slit exposure light 12a, and the observation image can be sent to the image processing apparatus by the built-in camera. Has become.
The image processing apparatus calculates a shift amount between the mask-side AA mark image and the plate-side AA mark image which are captured in an overlapping manner.

In scanning exposure, after the mask 11 is positioned with respect to a mask position reference mark (not shown) provided on the mask stage, two mask side AA marks 2 are provided.
5a moves so as to be positioned substantially directly under the slit-shaped exposure light 12a at the same time.

Next, the photosensitive substrate 10 on which the plate-side AA mark is formed is positioned so that the projected image of the plate-side AA mark 25b onto the mask 11 by the projection optical system 13 overlaps the mask-side AA mark 25a. Is adjusted to form an overlapping image on the microscopes 24a and 24b on the camera, and automatic alignment is executed by the image processing device. Synchronous scanning in proportion to the magnification of the projection optical system is performed with the positions of the mask stage and the photosensitive substrate stage after completion of the alignment as initial positions.

AA mark 26 on the mask side and the photosensitive substrate side
A-29a and 26b-29b are also sequentially aligned, the offset amount from the initial position of the stage at that time is stored in the memory, and then the span VS between the microscopes 24a and 24b is reduced along the slit-shaped exposure region. In this case, as shown in FIG.
Since it is necessary to move 5a and 35b in the scanning direction by VY, the mask 11 and the photosensitive substrate 10 are moved by an amount corresponding to VY. The movement amount VY is given by the equation Y = F (X) representing the shape of the slit, and the AA mark 25
If the span VS at the time of executing the alignment of a to 29a and 25b to 29b is VS0, the formula 12 is obtained.

[0041]

(Equation 12) In consideration of the reduction ratio β of the projection optical system 13, the photosensitive substrate 10
The movement amount VY of the mask 11 is moved in synchronization with VY ′ = VY / β.

In this way, the span VS is reduced from VS0 to VS1 and VS2, and the AA marks 30a to 33a and 3 are generated.
The alignment marks 0b to 33b are not aligned this time, only the displacement amount is read, and the displacement amounts in the Y direction are set to 30aY to 32aY, respectively. Next, the correction of this shift amount will be described.

Equation 11 shows that the inclination of the plane parallel plate 14 is
By bending so as to be proportional to the Y coordinate, it is possible to correct the Y direction nonlinear component error on the slit over the entire surface. Therefore, consider an X and Y coordinate system fixed to the slit as shown in FIG. Microscopes 24a, 24
The initial span of b is VS0, and the Y coordinate at this time is 0. Increasing the correction amount is nothing but increasing the value of k _{1 in} the equation (11). This is a plane parallel plate 14
1 is to increase the amount of inclination of the pulse motors, which is equivalent to increasing the driving amounts of the pulse motors 19a and 19b as described later in FIG. When the plane-parallel plate 14 is curved as a whole, the image at the position of X = VS0 / 2 may also be displaced. Therefore, the equation 11 is rewritten to obtain the equation 1
Type 3

[0044]

(Equation 13) However, k ₁ and k ₂ are newly determined proportional constants, and P is the number of drive pulses of the pulse motors 19a and 19b. Therefore,
Observation microscope span and pulse motors 19a, 19b
By waving the number of pulses of (1) to examine how much the image shift occurs compared to when the drive pulse is zero. It is assumed that n pieces of data are obtained as a result. If the amount of deviation in the Y direction at this time is ΔY _i , the characteristics of the correction mechanism described later in FIG. 1 can be obtained by obtaining k ₁ and k ₂ that minimize the equation (14).

[0045]

[Equation 14] Specifically, k that minimizes F under the condition that the expression obtained by partially differentiating F with k ₁ and the expression obtained by partially differentiating F with k ₂ are 0
It is possible to determine ₁ and k ₂ .

At the time of actually measuring the deviation amount before exposure,

[0047]

(Equation 15) Then, the shift amounts 30aY to 33aY are substituted into the equation 15 to similarly obtain P which minimizes G, and the pulse motor 19
The drive pulse numbers of a and 19b are obtained. As a result, the bending correction amount of the plane parallel plate 14 is obtained.

Finally, the deviation amount of the image forming position when the span of the observing microscope is VS0 is obtained by bending the parallel plane plate 14 by the equation (13), and the mask stage or the photosensitive substrate stage is exposed at VS0 during exposure. Scanning exposure is performed by correcting in the Y direction by the amount of the image shift.

Although the curvature of the plane-parallel plate 14 in one exposure scan is constant here, when the driving pulse number P of the pulse motors 19a and 19b is obtained from the equation (15), the AA pattern of FIG. 9 is used. It is also possible to perform the exposure by linearly changing the pulse number during the exposure by using the values obtained by using 29a to 31a and the values obtained by using the AA patterns 27a, 32a, 33a. In this case, it is also necessary to linearly change the correction amount of the position corresponding to the case where the span of the observation microscope is VS0 according to the number of pulses.

Next, the structure and operation of the portion that bends the plane-parallel plate 14 will be described in detail with reference to FIG. The pulse motors 19a and 19b are driven so that the measurement values of the electrostatic sensors 20a and 20b become preset values by turning on the power source of the apparatus or by scanning by the operator after the power source is input. As a result, the parallel flat plate 14 is immediately initialized so that the initial state becomes a flat surface, the accuracy thereof is sufficient as compared with the accuracy of the photo sensor and the like, and the driving accuracy can be corrected.

When the shafts of the pulse motors 19a and 19b rotate, the feed screws 18a and 18b rotate and the connecting rod 21
Moves up and down. Connecting rod 21 and feed screws 18a, 18
The nuts of b are connected by a not-shown bose pin so that they do not twist or the like. As the connecting rod 21 moves up and down, the equal stress beams 17a and 17b holding the plane-parallel plate 14 bend so that their inclinations change in proportion to the Y coordinate. 17a and 17b have a greater rigidity than the plane-parallel plate 14, and the plane-parallel plate 14 is a member 2
Since it is fixed to the beams 17a and 17b by 2a and 22b, the plane-parallel plate 14 is curved following the equal stress beams 17a and 17b. Further, since the plane parallel plate 14 is also held by the members 22c and 22d at the center thereof,
As a whole, the equal stress beams 17a and 17b are uniformly curved.

In addition, instead of the members 22a and 22b, vacuum suction is used to move the plane-parallel plate 14 to the equal stress beams 17a and 17b.
It may be fixed to b. Also, the pulse motor 1
Instead of 9a and 19b, it is possible to bend the plane-parallel plate 14 by the pressure of the diaphragm and the servo valve, and it is also possible to use a servo motor or a piezo stack.

Further, in this embodiment, the plane-parallel plate 14 is arranged below the projection optical system 13, but the projection optical system 13
It is also possible to dispose the plane-parallel plate 14 between the mask and the mask. However, in this case, the driving direction for correction is different from the case where the plane parallel plate 14 is arranged below the projection optical system 13. That is, as shown in FIG. 1, a case where the plane-parallel plate is bent toward the photosensitive substrate side so as to be convex, and a case where the plane-parallel plate 14 is inserted at the mask side and bent so as to be convex toward the mask side are shown. Then, the shift direction of the image forming position is opposite.

Further, as an optical system for causing the image refraction means to be displaced in the correction of the magnification or the correction of the uniform magnification by moving the light refracting means in the projection optical system 13 in parallel,
In the former case, in the same manner as the above-described embodiment, in the latter case, the component orthogonal to the scanning exposure, which is generated at the same time as the nonlinear component correction in the scanning exposure direction, is performed in combination with the lateral magnification correcting means. Therefore, the same correction can be realized.

[0055]

As described above, according to the present invention, in the longitudinal direction of the exposure light having a slit shape which is not a straight line shape,
By driving the magnification correction means so as to correct the image forming position deviation in the direction orthogonal or forming a constant angle, the slit exposure light is orthogonal to the longitudinal direction, which is impossible in the conventional scanning projection exposure apparatus. Non-linear image distortion in the direction can be corrected without making complicated adjustments and changes to the projection optical system.

Further, the slit-shaped exposure area is formed into a shape that can be expressed by an equation representing an arc or an even function with respect to a point on a straight line showing the longitudinal direction of the slit.
It is possible to correct only the non-linear component of the Y-direction non-linear component excluding the component that can be corrected by θ rotation.

Further, by using an optical plane parallel plate as the magnification correction means, the magnification correction means can be constructed without designing a demagnification zoom optical system which requires complicated calculation.

Further, the holder that supports and bends the plane-parallel plate or the plane-parallel plate itself satisfies the deflection equation determined by the equation representing the slit-like exposure area shape and the equation representing the image-forming position deviation to be corrected. With such a shape, it is possible to mathematically uniquely and accurately calculate the bending drive amount of the plane-parallel plate when correcting the error in the Y-direction non-linear component, and also at one or both ends of the plane-parallel plate. A desired bent shape of the plane-parallel plate can be realized only by driving the surface.

Further, an electric actuator is provided as a means for bending the plane-parallel plate, and a sensor capable of detecting the amount of bending of the plane-parallel plate or its holder is provided, whereby the Y-direction nonlinear error component is automatically obtained. Can be corrected.

Before the exposure, the amount of deviation of the image forming position is measured,
The amount by which the plane-parallel plate is deflected is calculated from this shift amount, and based on this, the amount by which the plane-parallel plate is deflected before exposure scanning or during exposure scanning is changed, thereby causing drift in the projection optical system and exposure to light. It is also possible to accurately correct an image shift due to thermal expansion and thermal contraction of each substrate.

[Brief description of drawings]

FIG. 1 is a perspective view showing a scanning projection exposure apparatus according to an embodiment of the present invention.

FIG. 2 is a diagram showing that the field of view moves in the Y direction when the field of view of an observation microscope moves along exposure light that is not linear in the apparatus of FIG.

FIG. 3 is a schematic diagram showing an image shift when an optical plane parallel plate is curved in the apparatus of FIG.

FIG. 4 is a diagram showing a bending direction of an optically parallel plane plate when performing lateral magnification correction in a broad sense.

FIG. 5 is a diagram for intuitively explaining that the Y-direction nonlinear error component cannot be corrected with the slit shape in the conventional example.

FIG. 6 is a vector diagram showing an image shift state when a Y-direction nonlinear error component occurs.

FIG. 7 is a shift state diagram when the maximum shift value is reduced by distribution when the image shift in FIG. 6 is performed.

FIG. 8 is an image shift diagram when the photosensitive substrate is uniformly thermally expanded.

9 is a diagram showing an arrangement example of AA patterns on a mask pattern when the scanning exposure apparatus of FIG. 1 is used.

10 is a diagram showing an arrangement example of AA patterns on a photosensitive substrate when the scanning exposure apparatus of FIG. 1 is used.

11 is a view showing a coordinate system fixed on slit-shaped exposure light which is not linear in the apparatus of FIG.

[Explanation of symbols showing main parts of the drawing]

10: Photosensitive substrate, 11: Mask or reticle, 12
a, 12b, 12c, 12d: slit-shaped exposure light, 1
3: Projection optical system, 14: Optically parallel plane plate, 15: Arrow showing scanning direction of mask, 16: Arrow showing scanning direction of photosensitive substrate, 17a, 17b: Beam holding optical plane parallel plate, 18a , 18b: feed screw, 19a, 19b: pulse motor, 20a, 20b: electrostatic sensor, 21: connecting rod, 22a, 22b, 22c, 22d: optical parallel flat plate holding metal fittings, 23a, 23b: folding mirror, 24
a, 24b: observation microscope, 25a to 33a, 25b to 3
3b: alignment mark, 34: actual element pattern, 35
a, 35b: visual field position of observation microscope, 36: arrow indicating scanning exposure direction of optically parallel plane plate, 37: displacement amount, 3
8: Reference grid.

Claims (6)

[Claims] 1. An illumination optical system for illuminating an original plate on which a transfer pattern is formed by exposure light having a slit shape whose cross section is not linear, and an image of the transfer pattern illuminated by the illumination optical system. A projection optical system for projecting onto the substrate at a predetermined magnification, and a speed ratio corresponding to the projection magnification of the original plate and the photosensitive substrate in a scanning direction that is at a right angle or at a constant angle with respect to the longitudinal direction of the illumination area by the exposure light. In the scanning projection exposure apparatus provided with the moving unit that moves in synchronization with each other, a magnification correcting unit that corrects the magnification of the projection optical system at least in the scanning direction is provided, and
Among the image-forming position shifts in the scanning direction in the case where the image is exposed and transferred onto the photosensitive substrate in an overlapping manner, a component that cannot be corrected by the correction of the relative position between the original plate and the photosensitive substrate is corrected. A scanning projection exposure apparatus characterized by the above. 2. The scanning projection exposure apparatus according to claim 1, wherein the slit shape can be expressed by an equation representing an arc or an even function. 3. The scanning projection exposure apparatus according to claim 1, wherein the magnification correction means includes a plane parallel plate arranged between the original plate and the photosensitive substrate, and a bending means for bending the parallel plane plate. 4. The bending means bends the plane-parallel plate so as to have a shape that satisfies the equation showing the slit shape and the equation showing the deflection determined by the equation showing the image forming position shift. 4. The scanning projection exposure apparatus according to claim 3, wherein the scanning projection exposure apparatus is one. 5. The bending means includes an actuator that applies a force for bending the parallel plane plate, and a sensor that detects a bending amount of the optical plane plate or its holder. Item 3. The scanning projection exposure apparatus according to Item 3. 6. A means for detecting a shift amount of the image forming position prior to the exposure, and a means for calculating a bending amount of the parallel flat plate from the shift amount, wherein the bending means is based on the result. 6. The scanning projection exposure apparatus according to claim 3, wherein the curvature amount is changed before or during the exposure scanning.