JPH06151277A - Aligner - Google Patents

Abstract

PURPOSE:To enable an aligner to carry out an accurate exposure operation for a long term without deteriorating in throughput by a method wherein factors which produce errors in horizontal position measurements when stages are driven are detected, and an X-Y stage and a micro-theta stage are controlled taking the detection results into consideration. CONSTITUTION:Factors, which produce errors in distances nux and nuy in a direction of Z from a light exposure reference plane to measurement axes Lx and Ly, angles betax and betay which the measurement axes Lx and Ly make with the normal lines of reflecting planes MX and MY, a parallelism kappa of the beam of a differential interferometer LZQ to an XY plane and a squareness tau of the beam to the reflecting planes MX and MY, a height difference delta between a measuring spot of the two beams of the differential interferometer LZQ on the reflecting planes MX and MY and the moving plane of an X-Y stage when the Y coordinates of the X-Y stage is zero, and an angle which the two beams of the differential interferometer LZQ make with each other when the beams are projected onto a plane which contains LZY and vertical to the moving plane of the X-Y stage, are detected. The X-Y stage and a micro-theta stage are controlled corresponding to the detection result.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an exposure apparatus for manufacturing a semiconductor integrated circuit, and more particularly, to tilt a wafer (photosensitive substrate) to provide a predetermined exposure reference plane, for example, a projection image plane of a mask pattern in a projection type exposure apparatus. The present invention relates to an exposure apparatus having a positioning mechanism that accurately matches the surface of a wafer and performs two-dimensional movement that minimizes horizontal rotation of the wafer.

[0002]

2. Description of the Related Art In a lithography process for manufacturing a semiconductor integrated circuit, a step-and-repeat type reduction projection exposure apparatus, a so-called stepper, plays a central role. In this stepper, it is necessary to increase the resolution limit of the projection lens year by year in response to the reduced line width of the circuit formed in the sub-micron range, and a large numerical aperture (NA) and a wide exposure field are required. At the same time, the demand for satisfaction is increasing. However, since a projection lens having a large numerical aperture (NA) and a wide exposure field inevitably has a shallow depth of focus, if a tilt occurs with respect to a predetermined exposure reference plane in a part of an arbitrary shot area on the wafer. However, it becomes difficult to always perform accurate focusing on the entire surface in the exposure field. Therefore, for example, Japanese Patent Laid-Open No. 3-24
Using the substrate surface position detection system disclosed in Japanese Patent No. 6411 etc., the inclination with respect to the exposure reference plane is detected for each shot area on the wafer. Then, in the same manner, JP-A-4-13429.
Stage device (wafer stage) disclosed in No. 3
Using a predetermined multiple points of the leveling stage (for example,
By driving the three operating points), the tilt angle of the leveling stage is controlled so that the tilt of the shot area becomes zero. As a result, by precisely matching the exposure reference plane, that is, the image plane and the surface of the shot area,
A projected image of a circuit pattern of a mask or reticle (hereinafter, referred to as a reticle) is exposed on the wafer with high resolution without causing a partial defocus in the exposure field.

The above-mentioned leveling mechanism has become an indispensable mechanism for coping with the future increase in resolution limit. Further, according to Japanese Patent Laid-Open No. 232321/1988 such as an alignment method using information on the entire surface of the wafer, the projection image of the circuit pattern of the reticle and the two-dimensional position of the chip of the wafer are aligned to perform high-precision exposure. In this alignment method, particularly in the case of off-axis alignment (the projection lens optical axis and the alignment system are separated), X
-It is important that the Y stage can be driven with an accurate grid in order to cope with the increase in wafer alignment accuracy. Therefore, the rotation angle of the plane mirror is measured using a differential interferometer, and the fine θ stage is controlled based on the difference from the desired value to control the accuracy of the XY stage, changes with time, history, etc. Instead, a mechanism is added to keep the rotation angle of the wafer constant.

[0004]

However, in a stepper having a leveling mechanism and a rotation control mechanism of this kind, the interferometer beam for measuring the position of the XY stage and the interferometer beam for measuring the rotation angle of the wafer are: It was necessary to align the height and direction so as to pass through the exposure reference plane configured to be located in the plane parallel to the running plane. Furthermore, X-
The mirror for two-dimensional position measurement on the Y stage is XY
It was necessary to align the laser beam of the interferometer vertically with the running plane of the stage. Furthermore, it was necessary to match the parallelism of the two beams for measuring the rotation angle. This is because if the above beam matching is not sufficient, an error will occur in the measurement value of the interferometer as follows.

(1) If the height of the beam of the XY interferometer is not aligned with the exposure reference plane, when the leveling stage is tilted by, for example, an angle θ to perform leveling of an arbitrary shot area, this exposure reference plane is used. A deviation of the center of the shot area with respect to the XY coordinates due to a height deviation ν from the center of the interferometer beam, a so-called sine error ΔSe occurs (ΔSe
= Ν · sin θ).

(2) If the beam of the XY interferometer does not strike the mirror on the XY stage perpendicularly, when the leveling stage is tilted by an angle θ, for example, to perform the leveling of an arbitrary shot area, for example, the exposure reference When the distance from the center of the surface to the mirror surface on the stage is Xm, and the distance from the center of the exposure reference surface to the interference surface of the interferometer is Lo, the angle β between the interferometer beam and the normal to the mirror surface is A shift in the XY coordinates of the center of the shot area, which is a factor, a so-called cosine error ΔCe occurs (ΔCe = (2Lo−X
m) (1-cos (θ + β))).

(3) The angle formed by two beams of a differential interferometer, which measures the rotation of the mirror surface to keep the rotation angle of the wafer constant, projected onto a plane parallel to the running surface of the XY stage. When the XY stage is moved in the beam direction (Y direction in the embodiment) with (parallelism) being κ and the squareness between the mirror and the beam being τ, the parallelism κ between the two differential interferometer beams. The measured value of the wafer rotation angle due to the squareness τ between the mirror and the mirror is deceived with respect to the actual rotation angle, resulting in ΔYe (ΔYe = Y · (1 / (cosκ−tan (τ + θ
q) · sin κ) −1 / L; where L is the span of the differential interferometer beam and θq is the driving position of the fine θ stage).

(4) The two beams of the differential interferometer measuring the rotation of the mirror surface to keep the rotation angle of the wafer constant, the laser beams emitted from the laser interferometer for detecting the position in the Y direction. The angle (parallelism) formed on each other when projected onto a plane perpendicular to the running surface of the XY stage is η, and the two beams when the Y coordinate of the XY stage is 0 measure the mirror surface. When the leveling stage is tilted by, for example, an angle θ and the leveling of an arbitrary shot area is performed in a state where the height difference from the running surface of the XY stage is δ, 2
Due to the height difference δ and parallelism η of the book differential interferometer,
The measured value of the wafer rotation angle is deceived (so-called sign error) ΔHe with respect to the actual rotation angle (ΔHe = (δ + Y)
-Tan η) -sin θ / L; where L is the span of the differential interferometer beam).

Therefore, for example, the tilt angle θ of the wafer is ±
1 second (300ppm), height deviation ν 0.5mm, | 2
Lo-Xm | 800 mm, β 1000 ppm, δ 0.5 mm, τ 1000 ppm, κ 1000 pp
m, η is 1000 ppm, Y is 200 mm, L is 70 m
m, ΔSe = 0.150 μm, ΔCe = 0.2
76 μm, ΔYe = 4.29 ppm, ΔHe = 2.14
An error of ppm occurs. The deviation amount including these errors is too large for the positioning accuracy (0.02 μm, 0.1 ppm) required for this type of wafer stage.
However, in order to correct these errors, it is necessary to perform leveling in each shot area and then again perform alignment between the projected image of the reticle circuit pattern and the circuit pattern already formed on the wafer. However, there was a problem in that Also,
It is extremely difficult to completely remove these factors by adjustment, and it is a work that requires a lot of skilled man-hours in manufacturing, which causes a problem of increasing the cost of the stepper.

Further, in order to prevent the mirror from moving due to the operation of the leveling mechanism, a stage is provided in which the mirror is directly provided on the XY stage and the leveling mechanism is arranged inside the stage to drive so as to reduce mechanically generated errors. However, as the accuracy required for the stepper becomes more severe, other components due to the operation of the leveling mechanism are not observed, and it is difficult to maintain high accuracy for a long period of time.

The present invention has been made in consideration of the above points. Rough adjustment is performed between the exposure reference plane and the XY interferometer beam and the differential interferometer beam, and an error factor remains. However, it is an object of the present invention to provide a stepper type exposure apparatus capable of performing stable and highly accurate alignment and exposure for a long period of time without deteriorating the positioning accuracy and throughput of the wafer stage.

[0012]

In order to achieve this object, the present invention illuminates an original plate on which a pattern is formed to form an image of the pattern on a predetermined exposure reference plane, for example, an image forming plane of the pattern image. Optical means for exposing and transferring onto the matched photosensitive substrate, substrate surface position detecting means for detecting the distance and inclination of the photosensitive substrate surface with respect to the exposure reference plane, substrate holding means for holding the photosensitive substrate, and exposing the substrate holding means. An XY stage that moves in the X and Y directions substantially parallel to the reference plane;
A leveling stage for inclining the substrate holding means in an arbitrary direction with respect to the exposure reference plane; a fine θ stage for rotating the substrate holding means in the θ direction around the optical axis of the optical means substantially perpendicular to the exposure reference plane; A horizontal position measuring means for measuring the X, Y, θ positions of the substrate holding means by a reflection plane fixed to the holding means and a laser interferometer for measuring the position of the reflection plane, and a substrate surface position detecting means and a horizontal position measuring means. Based on the output, each stage is driven and positioned, and the surface of the photosensitive substrate and the exposure reference surface are aligned,
In an exposure apparatus comprising a drive control means for exposing and transferring the pattern image of the original onto the photosensitive substrate by the optical means,
The driving means includes error factor detecting means for detecting an error factor that causes an error in the measurement value of the horizontal position measuring means when the respective stages are driven, and the control means considers the detection result, and the XY stage and The drive control of the fine θ stage is performed.

The error factor detecting means includes, for example, a deviation amount detecting means for detecting a deviation amount of the photosensitive substrate in the X and Y directions, and thereby detects a deviation amount when each stage is driven by a constant amount. By doing so, the error factor is detected.

The error factor detecting means detects the error factor by detecting, for example, a plurality of marks provided at three or more positions which are fixed in position with respect to the substrate holding means and are separated from each other. The plurality of marks are formed on the surface of the baseline measurement photochromic plate by exposure, for example.

The deviation amount detecting means is arranged, for example, at a different position away from the exposure optical axis of the optical means. An alignment detection system between the original plate and the photosensitive substrate can be used as the deviation detection means.

[0016]

In this structure, as shown in FIGS. 4 to 6, the distances νx, νy in the Z direction from the exposure reference plane of the measurement axes Lx, Ly in the horizontal position measuring means and the measurement axes Lx, Ly are the reflection planes MX, MY. Βx, βy formed by the normal line of the differential interferometer, the parallelism κ of the beam of the differential interferometer LZQ with respect to the XY plane, and the squareness τ with the reflection plane, and the differential interferometer LZQ when the Y coordinate of the XY stage is 0. The height difference δ from the running surface of the XY stage at the point where the two beams measure the reflection plane, and the two beams of the differential interferometer LZQ including XY
When each stage is moved, an error occurs in the measurement value of the horizontal position measuring means due to an error factor such as an angle (parallelism) formed by each other when projected onto a plane perpendicular to the running surface of the stage. However, these error factors are detected, and the drive control of the XY stage and the fine θ stage is performed in consideration of the detection result. Therefore, regardless of the error factors, the accurate XY stage and fine θ stage are always controlled. Positioning is performed.

Therefore, even if the position of the exposure reference plane and the interferometer of the horizontal position measuring means or the differential interferometer is roughly adjusted and an error factor remains, X, By calculating the correction values in the Y and θ directions and driving the XY stage and the fine θ stage so that the error with the target is minimized, the sin error and cos error that accompany the leveling operation are combined. The lateral deviation amount and the rotation error are equal to or less than a predetermined allowable value (resolution of the measurement system), that is, the resolution of the laser interferometer (0.005 μm, 0.07 pp).
m) Below.

[0018]

Embodiments of the present invention will now be described in detail with reference to the drawings. 1 is a diagram showing a schematic configuration of a stepper having a control mechanism according to an embodiment of the present invention, and FIG. 2 is a diagram showing a schematic arrangement of an interferometer of a wafer stage in the apparatus of FIG. In FIG. 2, a telecentric projection lens LZ having an optical axis perpendicular to the XY movement plane (coordinate system XY) of the wafer W forms a projected image of a circuit pattern drawn on a reticle (not shown) on the apparatus. It is formed on a predetermined exposure reference plane, that is, the image plane IM. The wafer W to be exposed is held via the wafer holder WH by the fine θ stage FQ rotating around the optical axis AX of the projection lens LZ. A substrate surface position detection system FD for measuring the distance and inclination of the surface of the wafer W from the image plane IM is provided on the projection lens LZ side. The fine θ stage FQ is provided on a leveling stage LB that can be tilted in an arbitrary direction with respect to the image formation plane IM, and the leveling stage LB has an image formation plane I.
It is provided on an XY stage XY that moves in the XY directions along M.

A flat mirror MX for a laser interferometer LZX for detecting the position in the X direction is provided at the end of the fine θ stage FQ.
And a laser interferometer LZY for detecting the position in the Y direction and a differential laser interferometer L for detecting the rotation angle of the fine θ stage.
A plane mirror MY for ZQ is provided. The center lines of the laser beams emitted from the laser interferometers LZX and LZY are measurement axes Lx and Ly in the X and Y directions, respectively, and the measurement axes Lx and Ly intersect at a point Q and the point Q
Is adjusted so that the optical axis AX passes through. Since the rotation center of the fine θ stage FQ does not coincide with the optical axis AX, when the fine θ stage FQ is rotated, the measurement axis L
A displacement proportional to the rotation amount and the distance from the optical axis AX to the wafer center WC appears in x and Ly. XY stage XY
In order to return the displacement amount to the original position, rotating the fine θ stage FQ results in the rotation about the optical axis AX being given to the wafer.

2 and 3 show a state in which the center point of the wafer W (wafer center WC) is aligned with the point Q, and it is assumed that the wafer W has a thickness variation of t due to TTV. At this time, the measurement axis L as shown in FIG.
The distances νx, νy in the optical axis direction of the projection lens from the exposure reference plane IM of x, Ly and the measurement axes Lx, Ly are the plane mirror M.
The angles βx and βy formed with the normals to X and MY, the parallelism κ of the beam of the differential interferometer LZQ and the perpendicularity τ of the differential interferometer beam to the mirror MY, as shown in FIG. As shown, when the Y coordinate of the XY stage is 0, the height difference from the running surface of the XY stage at the point where the two beams measure the mirror surface, and the difference interferometer's two XY including a laser beam emitted from a laser interferometer LYZ
An error occurs due to the angle (parallelism) η between the two when they are projected on a plane perpendicular to the running surface of the stage. Therefore, the error caused by these error factors is measured in advance and the magnitude of the error factor is calculated.

A method of measuring the error factor will be described below. First, in the wafer holder WH, the alignment detection system A
A wafer W having a mark as shown in FIG. 9 in which the lateral shift amount can be detected by A is held, and this mark is
The XY stage XY is driven so as to be positioned within the detection range of the alignment detection system AA. At this time, a deviation in the distance and inclination between the exposure reference plane (imaging plane) IM and the surface of the wafer W may occur due to the thickness variation of the wafer W.
Therefore, the main controller MC uses the surface position detection system FD,
The leveling stage LB is driven so that the detection signal of the surface position detection system FD becomes zero, and the surface of the wafer W is matched with the image plane IM. At this time, the amount of lateral displacement of the mark is measured by the alignment detection system AA. From this state, the leveling stage LB is tilted by θx and θy respectively on the X-axis and the Y-axis, and the lateral deviation amount of the mark is measured again by the alignment detection system AA to obtain lateral deviations ΔLZX and ΔLZY after the tilt before and after the tilt. The amount is represented by the following formula.

ΔLZX = νx · sin θx + (2Lox−
Xm) (1-cos (θx + βx)) ΔLZY = νy · sin θy + (2Loy-Ym) (1-
cos (θy + βy)) The lateral deviation amounts ΔLZX and ΔLZY are determined by the leveling stage L.
It depends on the tilt amount θ of B and the position x of the XY stage.
Here, when linear approximation is performed assuming that the inclination amount θ is sufficiently small, the lateral deviation amount ΔLZX is ΔLZX≈νx · θx + (2Lox−Xm) (βx + θx) ^ 2 / 2≈νx · θx + (Lox−Xm / 2) (βx ^ 2 + θx ^ 2 + 2 · βx · θx) = (νx + 2 · Lox · βx) · θx−Xm · βx · θx + Lox · (βx ^ 2 + θx ^ 2) ≈νx ′ · θx−Xm · βx · θx. Here, νx ′ = νx + 2 · Lox · βx, and the squared term is ignored. ΔLZY can be similarly approximated. Therefore, a plurality of XY stage positions are selected, and the lateral displacement amount due to the leveling stage drive at those positions is calculated,
A two-dimensional least squares approximation of the stage position and the leveling stage tilt amount is performed to obtain νx, βx, νy, βy.

There are two methods for obtaining the error factor of the rotation amount. The first method can be applied when the alignment system can measure marks corresponding to two marks on the reticle at the same time without moving the XY stage. That is, in the case of an alignment system that can directly measure the rotation amount of the mark, for example, a TTL alignment system that can measure the left and right marks, by driving the position of the XY stage to a plurality of points and measuring the rotation amount of the mark at that time, the error factor Can be asked. This is a method of measuring the rotation error of the stage by measuring the rotation of the mark with the rotation suppressed as much as possible in advance.

The second method can be applied when the alignment system cannot measure two marks corresponding to the two marks on the reticle without moving the XY stage. However,
In this case, the two marks on the reticle are exposed to the extent that the alignment system can measure at the same time, and the images are measured to indirectly measure the rotation amount of the stage. If the exposure target is a resist, it needs to be developed. In the case of an exposure target coated with a photochromic material, the afterimage is measured immediately. In this case, it is not necessary to prepare a wafer having a mark whose rotation is suppressed.

The rotation amount ΔLZQ is determined by the leveling stage L.
The tilt amount θy of B, the position Y of the XY stage XY, and the fine θ
It depends on the position θq of the stage FQ and is expressed by the following equation. ΔLZQ = ((δ + Y ・ tan η) ・ sin θy + Y ・ (1
/ (Cosκ-tan (τ + θq) · sinκ) -1)) / L ΔLZQ is the amount of wafer rotation obtained from the difference between the two marks when the alignment detection system AA measures the two marks. . When the linear approximation is performed assuming that the angle (parallelism) κ between the two beams of the differential interferometer and the squareness τ of the mirror and the beam are both small enough to be linearly approximated, the above equation yields ΔLZQ = (Δ + η ・ Y) / L ・ θy + Y ・ (κ ・ κ /
2+ (τ + θq) κ) / L. Therefore, after the movement of the XY stage XY, δ is obtained from the rotation amount when the tilt amount θy of the leveling stage LB is moved without moving the fine θ stage FQ. Also, select a plurality of Y coordinates of the XY stage XY,
Η is obtained from the change in the rotation amount by θy with respect to the Y coordinate.

Further, the XY stage position (Y in this embodiment) and a slight θ can be maintained without changing the inclination of the leveling stage LB.
A plurality of stage positions θq are selected, the amount of rotation at that position is obtained, and κ and τ are obtained from the amounts.

The error generating factor νx, obtained in this way,
νy, βx, βy, δ, κ, τ, η are input to the main controller MC, and at the time of exposure, the main controller MC first sucks the wafer W to be exposed to the wafer holder WH by vacuum suction. . Then, the lateral deviation amount of several points in the circuit pattern already formed on the surface of the wafer W is measured while driving the XY stage XY by the alignment detection system AA, and lateral deviation information of the entire wafer is obtained. At this time, a variation in the thickness of the wafer W may cause a deviation of the distance ν and a tilt θ between the exposure reference plane and the surface of the wafer W, as shown in FIG. Therefore, main controller MC uses surface position detection system FD, drives leveling stage LB so that the detection signal of this surface position detection system FD becomes zero, and makes the surface of wafer W coincide with image plane IM. At this time, the errors of LZX, LZY, and LZQ are calculated based on the drive angles θx and θy of the leveling stage LB, the XY stage position obtained from the lateral deviation information of the entire wafer, and the fine θ stage position, and XY Command the drive positions of the stage and the fine θ stage. The above operation is sequentially performed for all exposure regions on the wafer.

The formula for calculating the error is shown below. An approximate expression can be applied depending on the accuracy. ΔLZX = νx · sin θx + (2Lox−Xm) (1-
cos (θx + βx)) ≒ νx '· θx-Xm · θx · β
x ΔLZY = νy · sin θy + (2Loy−Ym) (1-
cos (θy + βy)) ≒ νy ′ ・ θy-Ym ・ θy ・ β
y ΔLZQ = ((δ + Y ・ tan η) ・ sin θy + Y ・ (1
/ (Cosκ-tan (τ + θq) ・ sinκ) -1)) / L
≒ (δ + Y ・ η) / L ・ θy + Y ・ (κ ・ κ / 2 + (τ
+ Θq) κ) / L Although the error factor is obtained by using a correctly patterned wafer here, instead of this, as shown in FIG. 10, a plurality of error factors provided in advance on the wafer stage or the wafer holder are used. You may make it require | calculate by detecting an accurate mark with an alignment detection means.
In this way, the error factor can be measured at any time.

Further, in order to measure the distance between the optical axis of the projection lens and the alignment detection system, that is, the so-called baseline, as shown in FIG. 11, an exposure is performed in which the photosensitive surface coated with the photochromic material PH is arranged on a wafer stage or a wafer holder. A device has been proposed. This material has the property of leaving an image of an exposed pattern or mark for a certain period of time. If this device is arranged, it is not necessary to prepare a pre-patterned wafer, and all five error factors can be measured.

Further, as shown in FIG. 7, the reproducibility between the exposure in which a normal resist-coated wafer is baked with the tilt of the leveling stage fixed and the exposure in which the leveling stage is forcibly tilted. Sought with vernier,
A two-dimensional least squares approximation of the stage position and the leveling stage tilt amount is performed from the data, and νx, βx, ν
y and βy may be obtained.

Similarly, as shown in FIG. 8, an adjacent shot may be overprinted on a wafer coated with a normal resist, and κ and τ may be obtained from the rotation data.

In the case of the off-axis type alignment system, the error factors δ and η can be obtained with higher accuracy. In the TTL alignment system, the rotation amount is measured by the difference appearing in the span corresponding to the two marks on the reticle, but in the off-axis method, the rotation amount can be measured by measuring only one mark. This is because the fine θ stage rotates about the optical axis of the projection lens. Assuming that the resolution of the alignment system is constant, as shown in FIG. 12, the distance (baseline) from the optical axis AX of the off-axis alignment system OA is usually larger than the mark span of the TTL alignment system TA. The resolution of the quantity ΔLZQ is higher in the off-axis method.

The partial differential equation for obtaining the error factor is shown in equation (1).

[0034]

[Equation 1]

[0035]

As described above, according to the present invention, even if the error factor remains due to the rough position adjustment between the exposure reference plane and the XY interferometer beam or the differential interferometer beam. ,
The lateral deviation amount including the sin error and the cos error and the rotation error are suppressed to a predetermined allowable value (resolution of the measurement system) or less, and the photosensitive substrate is accurately positioned at each exposure position without detection by the alignment optical system. Therefore, the throughput can be improved. Further, even if the error factor changes with time, high accuracy can be maintained without adjusting the device, only by re-measuring the error factor.

[Brief description of drawings]

FIG. 1 is a schematic configuration diagram of a stepper including a control mechanism according to an embodiment of the present invention.

2 is a side view showing a schematic configuration around a wafer stage in the apparatus of FIG.

FIG. 3 is a plan view of FIG.

FIG. 4 is an explanatory diagram showing factors of a sine error and a cosine error.

FIG. 5 is an explanatory diagram showing factors of a rotation error.

FIG. 6 is an explanatory diagram showing another factor of a rotation error.

FIG. 7 is a schematic diagram showing reproducibility when a wafer is forcibly tilted in another method of obtaining an error factor according to the present invention.

FIG. 8 is a schematic view showing a state of adjacent shot overlap burning in still another method for obtaining an error factor in the present invention.

9 is a plan view showing a wafer with patterned marks that may be used in the apparatus of FIG.

FIG. 10 is a plan view showing a plurality of marks provided on a stage as another example of marks for obtaining an error factor in the present invention.

FIG. 11 is a plan view showing afterimages of marks exposed on a plurality of photosensitive surfaces coated with a photochromic material provided on a wafer holder, as another example of marks for obtaining an error factor in the present invention. .

FIG. 12 is an explanatory diagram showing a state in which wafer rotation is measured by an off-axis alignment detection system in the present invention.

[Explanation of symbols]

MC: main controller, XY: XY stage, LB: leveling stage, FQ: fine θ stage, WH: wafer holder, W: wafer, LZ: projection lens, AX: projection lens optical axis, IM: image plane, FD: Substrate surface position detection system, M
X, MY: plane mirror, Q: fine θ stage rotation center, L
ZX, LZY, LZQ, LZ: Laser interferometer, Lx, L
y: measurement axis.

Claims (5)

[Claims] 1. Optical means for illuminating an original plate on which a pattern is formed and exposing and transferring an image of the pattern onto a photosensitive substrate whose surface is aligned with a predetermined exposure reference plane, and a photosensitive substrate surface for the exposure reference plane. Substrate position detecting means for detecting the distance and inclination of the substrate, substrate holding means for holding the photosensitive substrate, XY stage for moving the substrate holding means in the X and Y directions substantially parallel to the exposure reference plane, and substrate holding Leveling stage for inclining the means in an arbitrary direction with respect to the exposure reference plane, a fine θ stage for rotating the substrate holding means in the θ direction around the optical axis of the optical means substantially perpendicular to the exposure reference plane, and the substrate holding means. A reflection plane fixed to the substrate and a laser interferometer for measuring the position of the reflection plane, X of the substrate holding means,
The horizontal position measuring means for measuring the Y and θ positions and the respective stages are driven and positioned based on the outputs of the substrate surface position detecting means and the horizontal position measuring means so that the surface of the photosensitive substrate coincides with the exposure reference plane. Then, in the exposure apparatus having the drive control means for exposing and transferring the pattern image of the original onto the photosensitive substrate by the optical means, an error occurs in the measurement value of the horizontal position measuring means when the respective stages are driven. The exposure apparatus is provided with an error factor detection unit for detecting an error factor to be caused, and the control unit controls the drive of the XY stage and the fine θ stage in consideration of the detection result. 2. The error factor detecting means comprises X and Y on the photosensitive substrate.
A deviation amount detecting means for detecting a deviation amount in a direction is provided, whereby an error factor is detected by detecting a deviation amount when each stage is driven by a constant amount. 1. The exposure apparatus according to 1. 3. The error factor detecting means detects the error factor by detecting a plurality of marks provided at three or more positions which are fixed in position with respect to the substrate holding means and are separated from each other. The exposure apparatus according to claim 2, which is characterized in that. 4. The exposure apparatus according to claim 3, wherein the plurality of marks are formed on the surface of the baseline measurement photochromic plate by exposure. 5. The exposure apparatus according to claim 2, wherein the deviation amount detecting means is arranged at different positions apart from the exposure optical axis of the optical means.