JPH10144587A - Exposure method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To prevent generation of distortion of a shot region exposed on a wafer when an angle of inclination between the running direction of a wafer stage and a reference mark plate is changed. SOLUTION: A pattern image of a reticle 12 is exposed to a light on a shot region on a wafer 5, by scanning the wafer 5 on a Zθ axis driving stage 4 to an exposure region 62, synchronously with the scanning of the reticle 12 on a fine driving stage 11 of a reticle to an illumination region 61. A plurality of reference marks are formed on a reference mark plate 6, and corresponding alignment marks are formed on the reticle 12. Amount of position deviation of the reference marks and the alignment marks are measured by scanning the fine driving stage 11 for a reticle and the Zθ axis driving stage 4 in the corresponding scanning directions. The scanning direction of the reticle 12 is corrected by the amount of position deviation.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method for transferring a mask pattern onto a photosensitive substrate during a photolithography process for manufacturing, for example, a semiconductor device, an image pickup device (such as a CCD), a liquid crystal display device, or a thin film magnetic head. The exposure method is particularly suitable for use in performing exposure with a scanning exposure type exposure apparatus such as a step-and-scan method.

[0002]

2. Description of the Related Art For example, when manufacturing a semiconductor device, an exposure apparatus for transferring a pattern of a reticle as a mask to each shot area of a wafer (or a glass plate or the like) coated with a photoresist has conventionally been a stepper.・
A reduction projection type exposure apparatus (stepper) of an and repeat type (batch exposure type) has been frequently used. On the other hand, recently, in order to respond to the demand for transferring a large-area circuit pattern with high accuracy without significantly increasing the burden on the projection optical system, a part of the pattern on the reticle is transferred through the projection optical system. A so-called step-and-scan method in which the image of the pattern on the reticle is sequentially transferred to each shot area on the wafer by synchronously scanning the reticle and the wafer with respect to the projection optical system while being projected on the wafer. Projection exposure apparatuses have been developed.

A conventionally known aligner that transfers a pattern on the entire surface of a reticle onto the entire surface of a wafer with an equal-magnification normal image in a single scanning exposure using an integrated stage can be said to be a prototype of a scanning exposure apparatus. It is. On the other hand, in the step-and-scan method, since a projection optical system with a normal reduction magnification is used, it is necessary to drive the reticle stage and the wafer stage independently at a speed ratio according to the reduction magnification, and Since the movement between the shot areas is performed by the stepping method, the mechanism of the stage system is complicated and requires extremely high control (for example, Japanese Patent Laid-Open No. 7-1764).
No. 68).

Therefore, as shown in FIG. 9, the speed and the position of a stage of a step-and-scan type projection exposure apparatus have conventionally been controlled based on the measurement values of a laser interferometer. That is, in FIGS. 9A1 and 9A2, the wafer stage 51 on which the wafer W is mounted
The axis movable mirror 52X and the Y axis movable mirror 52Y are fixed,
An X-axis movable mirror 55X and a Y-axis movable mirror 55Y are also fixed to the reticle stage 54 on which the reticle R is mounted. Then, the orthogonal coordinate system of the plane on which the wafer W moves is represented by X
Assuming that the scanning direction at the time of scanning exposure is a direction along the Y axis (Y direction) as the axis and the Y axis, conventionally, two measurement lines are respectively provided in parallel with the Y-axis movable mirrors 52Y and 55Y for the scanning direction. Laser beams 53Y1, 53Y2, and 56Y
1,56Y2 is irradiated, and the moving mirror 52X for the non-scanning direction
And 55X are irradiated with one measurement laser beam 53X and 56X, respectively, the position in the scanning direction (Y coordinate) is measured by a biaxial laser interferometer, and the position in the non-scanning direction (X coordinate) is It was measured with a one-axis laser interferometer.

At this time, one of the two-axis laser interferometers for the scanning direction is a laser interferometer for yawing measurement,
The rotation angles of the wafer stage 51 (wafer W) and the reticle stage 54 (reticle R) are measured from the difference between the two Y-coordinates. Synchronous scanning of the two stages 51 and 54 has been performed so that the positional relationship is adjusted and the relative rotation angle between the two stages is constant. Since the projection optical system of reverse projection is usually used, the scanning directions of the wafer stage 51 and the reticle stage 54 are opposite to each other. Are both in the −Y direction.

That is, if the reflecting surface of the movable mirror is accurately parallel to the X-axis and the Y-axis during scanning exposure, the wafer W moves in the −Y direction with respect to the slit-shaped exposure region 58 by the wafer stage 51. Synchronously, the reticle stage 54 moves the reticle R in the −Y direction with respect to the slit-shaped illumination area 57, and the pattern image of the reticle R is transferred to one shot area on the wafer W.
The shot area to be exposed as a result is an accurate rectangle like the shot area SAa enlarged in FIG. 9A3, and the shot array formed on the wafer W
As shown in 4), it is a grid arranged along the X axis and the Y axis.

On the other hand, the yaw of the wafer stage 51 causes the movement of the movable mirror 52 as shown in FIG.
When X and 52Y rotate clockwise by the angle θ, the scanning direction of the wafer W is changed to the moving mirror 52X as shown by an arrow 60b.
(The direction inclined by an angle θ with respect to the original Y axis), and the stepping direction of the wafer W in the non-scanning direction is determined by the moving mirror 52Y as shown by an arrow 61b.
In the direction along the reflection surface. In this case, the moving mirror 52Y
The rotation of the wafer stage 51 is detected by the inclination of
Since the reticle stage 54 is also rotated by the angle θ at the same time, as shown by an arrow 59b in FIG.
The reticle R is also scanned while being rotated by the angle θ and in the rotated direction. Therefore, the shot area (transfer area of the pattern image of the reticle R) exposed on the wafer W by the scanning exposure is the shot area SA in FIG.
As shown by b, it is a rotating but accurate rectangle,
The shot arrangement on the wafer W (see FIG. 9 (b4)) also has a lattice shape (hereinafter, referred to as an "orthogonal lattice shape") that rotates but has an orthogonal arrangement direction.

[0008]

As described above, in the conventional step-and-scan type projection exposure apparatus, the coordinate positions of the wafer stage and the reticle stage are measured by the laser interferometer, and the two axes for the laser interferometer are used. When the orthogonality of the moving mirror was good, the shape of the shot area exposed even if the wafer stage was rotated by yawing was rectangular, and the shot arrangement obtained was also an orthogonal grid.

However, the stage is thermally deformed due to a temperature change of the ambient gas or a temperature rise due to exposure light irradiation, or the movable mirror itself for the laser interferometer is thermally deformed. When the orthogonality of the mirror is deteriorated, the shape of the shot area to be exposed may not be rectangular, and the shot arrangement may not be orthogonal grid. This is mainly because, conventionally, the position of the stage in the scanning direction was measured by a two-axis laser interferometer, and the rotation angle of the stage was obtained from the difference between the obtained measurement values. That is, in the example of FIG. Moving mirrors 52Y, 5
This is considered to be because the laser beam from the laser interferometer for yawing measurement was irradiated on 5Y.

More specifically, FIG. 9 (c1) shows a state in which the movable mirror 52X for the non-scanning direction on the wafer stage side, ie, the yaw measurement is not performed, is inclined by the angle θ. In this case, the scanning direction of the wafer W is the direction along the reflecting surface of the movable mirror 52X as shown by the arrow 60c, but since the change in the rotation angle of the wafer stage is not detected, FIG.
As shown in (c2), the reticle R (reticle stage 54) is scanned in the −Y direction. Therefore, the shot area formed on the wafer W becomes a parallelogram as shown by the shot area SAc in FIG. 9C3, and the shot arrangement (see FIG. 9C4) also becomes a parallelogram.

FIG. 9 (d1) shows a state in which the movable mirror 52Y for the scanning direction on the wafer stage side, ie, the yaw measurement is being performed, is inclined by the angle θ. In this case, the scanning direction of the wafer W is the −Y direction, but since a change in the rotation angle of the wafer stage is detected, the reticle R is moved to the original Y axis as indicated by an arrow 59d in FIG. 9D2. On the other hand, scanning is performed in a direction rotated by the angle θ and in a direction inclined by the angle θ. Therefore, the shot area formed on the wafer W has a shape obtained by rotating the parallelogram by 90 ° as shown by the shot area SAd in FIG. 9D3, and the shot arrangement (see FIG. 9D4) is the same. It becomes the shape of.

The shot arrangement error shown in FIG. 9 (c4) or (d4) is a linear error (primary error).
When performing exposure on a layer thereabove, for example, a so-called enhanced global alignment (EGA) type alignment is performed to obtain a shot array by statistical processing, and stepping of the wafer stage is performed according to the obtained shot array. Thus, the arrangement error can be substantially corrected. However, it goes without saying that the smaller the arrangement error, the better. On the other hand, when the shot area is deformed as shown by the shot areas SAc and SAd, assuming that the width of the slit-shaped exposure area 58 in FIG. 9A1 in the Y direction is D and the rotation angle of the movable mirror is θ (rad). The image exposed on the wafer W becomes equivalent to an image laterally shifted by approximately D · θ in the non-scanning direction during the scanning exposure, and has a disadvantage that image deterioration occurs.

In order to reduce such image deterioration, a reference mark plate on which a predetermined reference mark is formed is conventionally fixed on a wafer stage, and the reference mark plate is periodically (for example, when the wafer is replaced). The relative rotation angle between the reference mark and the alignment mark on the reticle is measured, and the rotation angle of the reticle is corrected based on the measurement result. At this time, if the running direction of the wafer stage determined by the moving mirror of the laser interferometer and the angle between the reference stage and the reference mark plate change due to thermal deformation of the wafer stage, etc., the rotation angle of the reticle can be accurately corrected. And the shot area to be exposed is deformed.

In the case of performing EGA alignment, for example, as described above, it is necessary to detect the position of an alignment mark (wafer mark) attached to a predetermined shot area on a wafer using a predetermined alignment sensor. . At this time, in order to accurately align each shot area on the wafer with the pattern image of the reticle based on the detection result of the alignment sensor, for example, the reference mark of the alignment sensor is periodically used by using a reference mark plate. A baseline amount which is an interval between a point (detection center or the like) and a reference point (exposure center or the like) of an image to be exposed on the wafer is obtained and stored, and the detection result of the alignment sensor is corrected by the base line amount. You. The regular measurement of the baseline amount in this way requires an interval
This is called a baseline check. Again, in this case,
If the running direction of the wafer stage determined by the moving mirror of the laser interferometer and the inclination angle of the reference mark plate fluctuate due to thermal deformation or the like, the baseline amount fluctuates substantially and the overlay error increases. was there.

In view of the above, the present invention has been made in view of the above circumstances, even when the running direction of the wafer stage and the relative angle between the reference mark plate serving as a reference for the rotation angle of the reticle are changed, the distortion of the shot area exposed on the wafer. It is a first object of the present invention to provide an exposure method which can prevent the occurrence of an image or can arrange a shot array on a wafer in an orthogonal lattice shape. The present invention further provides an exposure method capable of measuring the baseline amount with high accuracy even when the running direction of the wafer stage and the relative angle between the reference mark plate for measuring the baseline amount of the alignment sensor change. Is a second purpose.

Further, according to the present invention, the relative angle between the running direction of the wafer stage and the reference mark for measuring the baseline amount of the alignment sensor is hardly changed, and as a result, the exposure for measuring the baseline amount with high accuracy can be achieved. A third object is to provide a method.

[0017]

According to a first exposure method of the present invention, a part of a pattern on a mask (12) is formed by exposing a photosensitive substrate (1-4) on a substrate stage (1-4) under exposure light. 5)
The mask (12) and the substrate (5) are synchronously scanned in the corresponding scanning direction in a state of being projected on the upper surface, so that the pattern on the mask (12) is sequentially transferred to each shot area on the substrate (5). In the exposure method, a plurality of measurement marks (29) are formed on the mask (12) along the scanning direction.
A, 29D), and a reference mark member (6) on which a plurality of reference marks (35A, 35D) having substantially the same positional relationship as the plurality of measurement marks are formed.
Is placed on the substrate stage, and the mask (1) is moved by moving the substrate stage in the scanning direction.
2) The positional deviation amount between one of the plurality of measurement marks (29A) and the plurality of reference marks (35A, 35D) on the reference mark member (6) is sequentially determined. Measuring, and detecting a relative rotation angle θ1 between the arrangement direction of the plurality of reference marks (35A, 35D) and the running direction of the substrate stage from the measurement result;
The mask (12) and the substrate (5) are synchronously moved in the corresponding scanning direction, and each of the plurality of measurement marks (29A, 29D) on the mask (12) and the reference mark member (6) are moved. Corresponding reference mark (35A, 35D)
And a second step of detecting a relative rotation angle θ2 between the scanning direction of the mask (12) and the scanning direction of the substrate stage based on the measurement result. The stepping direction of the substrate stage is determined based on the information of the angle θ1, and the scanning direction of the mask (12) is determined based on the information of the relative rotation angle θ2.

According to the present invention, the relative rotation angle θ1 detected while the mask (12) is fixed and the substrate stage is scanned is determined by the arrangement direction of the reference marks (35A, 35D), that is, the reference mark. This is a relative angle between the angle of the member (6) and the running direction of the substrate stage during scanning exposure.
If the substrate stage is scanned along the reference mark member (6), the relative rotation angle θ2 measured when the mask (12) is moved in synchronization with the corresponding scanning direction.
Is a rotation error of both stages in the scanning direction. Therefore,
Even when the angle between the substrate stage and the reference mark member (6) changes due to thermal deformation of the stage or the like, the stepping direction between the shot areas of the substrate stage is determined by, for example, the arrangement of the reference marks on the reference mark member (6). By setting the direction or the direction perpendicular thereto, the shot arrangement formed on the substrate (5) has an orthogonal lattice shape. Further, by scanning the mask (12) along the reference mark member (6) with the scanning direction of the substrate stage as the direction along the reference mark member (6), the shot area becomes rectangular.

The second exposure method according to the present invention comprises the first exposure method of the present invention, the first step thereof, and the second exposure method.
It is the same until the process is executed. Thereafter, in the second exposure method, the rotation angle of the mask (12) is corrected based on the difference between the relative rotation angle θ1 and the relative rotation angle θ2 obtained in the first step and the second step. is there. According to the present invention, for example, assuming that a relative rotation error of the mask (12) with respect to the reference mark member (6) is Δθ, when the mask (12) and the substrate (5) are synchronously moved in the corresponding scanning direction. Is calculated as (Δθ +
θ1). That is, the relative rotation error Δθ of the mask (12) with respect to the reference mark member (6) is (θ2−θ1). Therefore, even when the angle between the substrate stage and the reference mark member (6) changes due to thermal deformation of the stage or the like, the relative rotation error Δθ according to the present invention is measured, and the rotation angle of the mask (12) is set to, for example, By correcting by −Δθ, the rotation angle of the mask (12) can be adjusted to the reference mark member (6). This makes the shot area exposed on the substrate (5) more rectangular.

The third exposure method according to the present invention is the same as the above-described first exposure method of the present invention until the first step is executed. Then, in this third exposure method,
By scanning the mask (12) in the scanning direction, one of the plurality of reference marks (35A) on the reference mark member (6) and the mask (12).
The displacement amount between each of the plurality of measurement marks (29A, 29D) is sequentially measured, and based on the measurement result, the arrangement direction of the plurality of measurement marks (29A, 29D) and the mask (12) A second step of detecting a relative rotation angle θ3 with respect to the running direction is performed, and the position of the substrate stage at the time of scanning exposure is corrected based on the information of the relative rotation angle θ1 detected in the first step. The position of the mask (12) at the time of scanning exposure is corrected based on the information of the relative rotation angle θ3.

According to the present invention, the relative rotation angle θ1 detected in the first step is the inclination angle of the traveling direction of the substrate stage with respect to the reference mark member (6), and is detected in the second step. The relative rotation angle θ3 is determined by the mask (1
The inclination angle of the pattern of the mask (12) with respect to the running direction of (2). Therefore, at the time of scanning exposure, the position of the substrate stage is gradually shifted so that the scanning direction of the substrate stage is in the direction along the reference mark member (6).
By gradually shifting the position of the mask (12) so that the scanning direction of the mask (12) is in the direction along the pattern of the mask (12), distortion of a shot area exposed on the substrate (5) is reduced. You.

In a fourth exposure method according to the present invention, an image of a part of a pattern on a mask (12) is exposed to light from a substrate stage (1-4) via a projection optical system (8) under exposure light. The mask (12) and the substrate (5) are synchronously scanned in a corresponding scanning direction while being projected onto the upper photosensitive substrate (5), so that the pattern on the mask (12) is projected onto the substrate (5). In an exposure method for sequentially transferring the shot marks to the respective shot areas, an off-axis alignment system (34) for detecting the position of an alignment mark on the substrate (5) near the projection optical system (8). And a plurality of measurement marks (29A, 29A) on the mask (12) along the scanning direction.
D) is formed and the first and second reference marks (35A, 37) are formed at intervals corresponding to the intervals between the reference points in the exposure field of the projection optical system (8) and the reference points of the alignment system (34).
The reference mark member (6) on which A) is formed is placed on the substrate stage, and the second reference mark (37A) on the reference mark member (6) is observed by the alignment system (34). The mask (12) is moved in the scanning direction, and a plurality of measurement marks (29) on the mask (12) are moved.
A, 29D) and the first reference mark (35A) on the reference mark member (6) are sequentially measured, and each of the plurality of measurement marks (29A, 29D) and the first reference mark (35D) are measured. The average value of the amount of misalignment with the reference mark (35A), and the mask (1
2) the relative rotation error with respect to the scanning direction, and the second reference mark (37) observed by the alignment system (34).
The distance (baseline amount) between the reference point in the exposure field of the projection optical system (8) and the reference point of the off-axis alignment system (34) is obtained from the displacement amount in (A).

According to the present invention, the arrangement direction of the measurement marks (29A, 29D) on the mask (12), that is, the direction of the transfer image of the mask pattern and the mask ( In parallel with the measurement of the relative rotation error θ3 with the running direction in 12), the displacement ΔB2 of the second reference mark (37A) observed by the alignment system (34) is measured, and these measured values are measured. Thus, a baseline amount of the alignment system (34) is obtained. Therefore, even when the relative angle of the reference mark member (6) to the substrate stage changes due to deformation of the substrate stage or the like, the baseline amount is measured with high accuracy based on the reference mark member (6).

In a fifth exposure method according to the present invention, an image of a part of a pattern on a mask (12) is exposed to light from a substrate stage (1-4) via a projection optical system (8) under exposure light. The mask (12) and the substrate (5) are synchronously scanned in a corresponding scanning direction while being projected onto the upper photosensitive substrate (5), so that the pattern on the mask (12) is projected onto the substrate (5). In the exposure method of sequentially transferring the shot marks to the respective shot areas, an off-axis alignment system (34) for detecting the position of the alignment mark on the substrate (5) near the projection optical system (8). ), And a plurality of measurement marks (29A, 2A) are arranged on the mask (12) along the scanning direction.
9D), and a first reference mark (35A, 3A) corresponding to the plurality of measurement marks on the mask (12).
5D) and a plurality of these first
The second reference marks (37A, 3A) are arranged at intervals corresponding to the intervals between the reference points in the exposure field of the projection optical system (8) and the reference points of the alignment system (34).
The reference mark member (6) on which a plurality of reference marks 7D) are formed is arranged on the substrate stages (1 to 4), and the mask (12) and the substrate (5) are synchronously moved in their corresponding scanning directions. A plurality of measurement marks (29A,
29D) and the second fiducial mark (35A, 35D) on the fiducial mark member (6) corresponding to the second fiducial mark (35D) by the alignment system (34). 37A, 37D) is repeated for each of the plurality of first reference marks (35A, 35D) on the reference mark member (6), and the plurality of first reference marks (35A, 35D) are measured. And correcting the relative rotation error between the mask (12) and the substrate (5) between their corresponding scanning directions on the basis of the positional shift amount obtained for each of the second reference mark and the reference mark member ( 6) It corrects the relative rotation angle between the arrangement direction of the first and second reference marks and the scanning direction of the substrate stage.

According to the present invention, substantially simultaneously with the execution of the second step of the first exposure method of the present invention,
The baseline amount of the alignment system (34) is measured. Then, based on the measurement results, the scanning direction of the substrate stage is corrected with reference to the reference mark member (6), and the scanning direction of the mask (12) is corrected with respect to the scanning direction of the substrate stage. . Therefore, even when the running direction of the substrate stage, the rotation angle of the mask (12), and the relative angle between the reference mark plate for measuring the baseline amount of the alignment system (34) change,
The distortion of the shot area to be exposed is reduced, and the baseline amount of the alignment system (34) can be measured with high accuracy.

In a sixth exposure method according to the present invention, an image of a part of a pattern on a mask (12) is exposed through exposure optical system via a projection optical system (8). 4
A) The pattern on the mask (12) is projected onto the photosensitive substrate (5) on the substrate (5) by synchronously scanning the mask (12) and the substrate (5) in the corresponding scanning direction. (5) In an exposure method for sequentially transferring images to each of the shot areas, an off-axis type alignment system for detecting the position of an alignment mark on the substrate (5) near the projection optical system (8). 34), a movable mirror (41X) for measuring coordinate position is fixed on the substrate stage, and a plurality of measurement marks (29A, 29D) are formed on the mask (12) along the scanning direction. And a moving mirror (4
1X), a plurality of first reference marks (35A, 35D) are formed corresponding to the plurality of measurement marks (29A, 29D) on the mask (12), and the plurality of first marks are formed. From the reference mark of each projection optical system (8)
Reference point in the exposure field and alignment system (34)
A plurality of second reference marks (37A, 37D) are formed at intervals corresponding to the interval with the reference point, and the mask (12) and the substrate (5) are moved in synchronization with the corresponding scanning direction, A plurality of measurement marks (29) on the mask (12)
A, 29D) and the second reference mark by the alignment system (34) in parallel with the measurement of the amount of displacement between the first reference mark (35A, 35D) on the movable mirror (41X) corresponding thereto. The step of measuring the displacement amount of (37A, 37D) is repeatedly performed for each of the plurality of first reference marks (35A, 35D) on the movable mirror (41X), and the plurality of first reference marks (35A, 35D) are measured. And a relative rotation error between the mask (12) and the substrate (5) between their corresponding scanning directions based on the amount of positional deviation obtained for each of the second reference mark and the projection optical system ( The interval (baseline amount) between the reference point in the exposure field of 8) and the reference point of the alignment system (34) is corrected.

According to the present invention, when measuring the baseline amount, the mask (12) is scanned based on the first reference marks (35A, 35D) on the movable mirror (41X) also serving as the reference mark plate. The direction is corrected. In addition, a moving mirror (4
1X) also serves as a reference mark plate, and the running direction of the substrate stage is along the reflection surface of the movable mirror (41X), so that the inclination angle of the reference mark with respect to the running direction of the substrate stage is It is hard to change.
Therefore, the measurement of the baseline amount can be performed with high accuracy, and distortion of the shot area to be exposed hardly occurs.

[0028]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of the present invention will be described below with reference to FIGS. In this example, the present invention is applied to the case where exposure is performed by a step-and-scan type projection exposure apparatus. FIG. 1 shows a projection exposure apparatus of the present embodiment. In FIG. 1, a rectangular illumination area (hereinafter, referred to as a “slit illumination area”) by exposure light EL from an illumination optical system (not shown) is provided on a reticle 12. Is illuminated, and an image of the pattern is projected through a projection optical system 8 onto a wafer 5 coated with a photoresist.
Projected above. In this state, the wafer 5 is scanned in synchronization with the reticle 12 being scanned forward (or backward) at a constant speed V with respect to the paper surface of FIG. Is scanned at a constant speed V / M (1 / M is the projection magnification of the projection optical system 8) backward (or forward) with respect to the paper surface of FIG. The projection magnification (1 / M) is, for example, 1/4, 1/5, or the like. Hereinafter, the optical axis A of the projection optical system 8
The Z axis is taken in parallel with X, and the Y axis is set in a design scanning direction of the reticle 12 and the wafer 5 in a plane perpendicular to the Z axis (that is, a direction perpendicular to the plane of FIG. 1), and is orthogonal to this scanning direction. The description will be made by taking the X axis in the non-scanning direction (that is, the direction along the plane of FIG. 1). However, the actual scanning direction of the reticle and the wafer is parallel to the Y axis (Y direction) due to the inclination of the moving mirror of the interferometer for coordinate measurement of the stage system as described later.
May deviate from

Next, the reticle 12 and the stage system of the wafer 5 in this embodiment will be described. First, the reticle support 9
Reticle Y-axis drive stage 10 that can be driven upward in the Y direction
The reticle minute drive stage 11 is placed on the reticle Y-axis drive stage 10, and the reticle 12 is held on the reticle minute drive stage 11 by a vacuum chuck or the like. Reticle micro drive stage 11
Performs position control of the reticle 12 in a small amount and with high accuracy in each of the X direction, the Y direction, and the rotation direction (θ direction). Reticle support 9 and reticle Y-axis drive stage 1
A reticle stage is constituted by 0 and the reticle minute drive stage 11. Reticle micro drive stage 1
A movable mirror 21 is disposed on the reticle 1, and the positions of the reticle minute drive stage 11 in the X, Y, and θ directions are constantly monitored by an interferometer body 14 disposed on the reticle support 9. That is, the interferometer body 14 is actually
The four-axis interferometer bodies 14X1 and 14X shown in FIG.
2, 14Y1 and 14Y2. Interferometer body 1
4 is supplied to the main control system 22A that controls the overall operation of the apparatus. Main control system 22
A controls the operations of the reticle Y-axis driving stage 10 and the reticle minute driving stage 11 via the reticle driving device 22D.

On the other hand, a wafer Y-axis drive stage 2 is mounted on the wafer support table 1 so as to be drivable in the Y direction, and a wafer X-axis drive stage 3 is mounted thereon so as to be drivable in the X direction. A Zθ axis driving stage 4 is provided thereon, and a wafer 5 is held on the Zθ driving stage 4 by vacuum suction. The Zθ axis drive stage 4
It controls the position of the direction, the tilt angle, and the minute rotation angle. The wafer stage is composed of the wafer support 1, the wafer Y-axis drive stage 2, the wafer X-axis drive stage 3, and the Zθ-axis drive stage 4. The movable mirror 7 is also fixed on the Zθ axis drive stage 4, and the interferometer body 1 disposed outside
3, the position of the Zθ axis drive stage 4 in the X, Y and θ directions is monitored, and the position information obtained by the interferometer main body 13 is also supplied to the main control system 22A. That is, the interferometer body 13 is also actually a five-axis interferometer body 13X1, 13X2, 13FX, 13Y shown in FIG.
1,13Y2 is a general term. The main control system 22A is connected to the wafer Y-axis driving stage 2 via the wafer driving device 22B,
Wafer X axis drive stage 3 and Zθ axis drive stage 4
Control the positioning operation.

As will be described later, the correspondence between the coordinate system of the wafer stage defined by the coordinates measured by the interferometer body 13 and the coordinate system of the reticle stage defined by the coordinates measured by the interferometer body 14. For this purpose, a reference mark plate 6 on which a predetermined reference mark is formed is fixed near the wafer 5 on the Zθ axis drive stage 4. Among the reference marks, there are also reference marks that are illuminated from the bottom side with illumination light of the same wavelength range as the exposure light IL guided into the Zθ axis drive stage 4, that is, a luminescent reference mark.

A reticle alignment microscope (hereinafter referred to as "RA microscope") 19 for observing the reference mark on the reference mark plate 6 and the alignment mark on the reticle 12 simultaneously above the reticle 12 of the present embodiment. 20 are arranged. In this case, deflecting mirrors 15 and 16 for guiding the detection light from the reticle 12 to the RA microscopes 19 and 20, respectively, are movably arranged, and when the exposure sequence is started, under an instruction from the main control system 22A. Then, the deflection mirrors 15 and 16 are retracted by the mirror driving devices 17 and 18, respectively. Further, an off-axis type alignment sensor 34 for detecting the position of an alignment mark (wafer mark) on the wafer 5 is disposed on a side surface of the projection optical system 8 in the Y direction. As an example, the alignment sensor 34 of the present embodiment detects the amount of misalignment of the wafer mark with respect to the internal index mark by an imaging method, and supplies the detected amount to the main control system 22A. As the alignment sensor 34, a so-called two-beam interference method (LIA method) of irradiating two coherent light beams to a diffraction grating wafer mark, or a laser beam irradiated in a slit shape and a dot-row-shaped wafer mark is used. A laser step alignment method (LSA method) or the like that relatively scans can be used.
A console 22C for inputting commands from an operator and displaying measurement data is connected to the main control system 22A.

Next, the configuration of the laser interferometer (laser light wave interferometer) for the stage system of this embodiment will be described with reference to FIG. FIG. 2A is a plan view showing the reticle minute drive stage 11 in FIG. 1, and FIG. 2B is a plan view showing the Zθ axis drive stage 4 on the wafer stage side in FIG. ), The reticle minute drive stage 1
Reticle 12 is held on 1 by vacuum suction or the like,
Exposure light is applied to a slit-shaped illumination area 61 elongated in the X direction on the reticle 12.

Since the reticle 12 (reticle minute drive stage 11) is scanned in the Y direction, the reticle minute drive stage 11 is formed of parallel flat glass at the end in the + X direction of the reticle minute drive stage 11 extending in the scanning direction (Y direction). The movable mirror 21X is installed, and the movable mirror 2 is moved from the interferometer bodies 14X1 and 14X2.
The 1X reflection surface is irradiated with measurement laser beams (measurement beams) LRX1 and LRX2 in parallel in the Y direction at an interval L1. The interferometer bodies 14X1 and 14X2 each include a reference mirror, a movable mirror 21X and a receiver for receiving interference light of a laser beam from the reference mirror, and a signal processing unit for processing a photoelectric conversion signal from the receiver. The X-coordinate of the reflection surface of the movable mirror 21X can be detected by processing the photoelectric conversion signal from the receiver. The movable mirror 21X accelerates the reticle 12,
At the time of deceleration, both measurement beams LRX1 and LRX2 are
It is formed long enough not to deviate from 1X. Further, the measurement beams LRX1 and LRX2 are at the center of the slit-shaped illumination area 61 (the optical axis AX of the projection optical system 8).
, The measurement values of the interferometer bodies 14X1 and 14X2 are XR1 and XR2, respectively, and the position of the reticle 12 in the non-scanning direction as an average value of these measurement values (XR1 and XR2) X coordinate) XR
Is detected, and the difference between the measured values is divided by the interval L1 to detect the rotation angle θ _{RX of} the reticle 12 viewed from the non-scanning direction. That is, the following equation holds.

[0035]

XR = (XR1 + XR2) / 2

[0036]

Θ _RX = (XR1−XR2) / L1 Corner cubes 21Y1 and 21Y2 as movable mirrors are fixed at the end of the reticle minute drive stage 11 in the + Y direction at an interval L2 in the X direction. Of the interferometer body 14 with respect to the corner cubes 21Y1 and 21Y2.
From Y1 and 14Y2, measurement beams LRY1 and LRY2 are emitted in parallel along the scanning direction. In addition, the measurement beams LRY1 and LRY2 reflected by the corner cubes 21Y1 and 21Y2 are respectively reflected to interferometer main body 1
Fixed plane mirrors 14M1 and 14 returning to the 4Y1, 14Y2 side
M2 is arranged, and the interferometer main bodies 14Y1 and 14Y2 are respectively provided with corner cubes 21Y1 and 21Y2 by a double-pass interference method.
The Y coordinate of 1Y2 is detected. The reticle 12 has a narrow moving range in the X direction, and has a corner cube 21Y1,
Since the position can be accurately detected in a range in which the measurement beams LRY1 and LRY2 fall on the incident surface of 21Y2, the corner cube 21 is used as a movable mirror in the scanning direction.
Y1, 21Y2 can be used.

The measurement beams LRY1 and LRY2 are also distributed symmetrically in the X direction with respect to the center (optical axis AX) of the illumination area 61, and the interferometer bodies 14Y1 and 14Y2
Of the reticle 12 by the average of the measured values YR1 and YR2 of
Is detected in the scanning direction (Y coordinate) YR. Also,
The rotation angle θ _{RY of the} reticle 12 viewed from the scanning direction is detected by dividing the difference between the measured values by the interval L2. That is, the following equation holds.

[0038]

## EQU3 ## YR = (YR1 + YR2) / 2

[0039]

The difference between the rotation angle θ _RY viewed from the scanning direction and the rotation angle θ _RX viewed from the non-scanning direction is as follows: θ _RY = (YR1−YR2) / L2 Orthogonality error ΔωR with cubes 21Y1, 21Y2
Becomes

[0040]

ΔωR = (YR1-YR2) / L2- (XR1
−XR2) / L1 In this example, at the time of normal exposure, the rotation angle (the amount of yawing) of the reticle 12 is corrected based on the rotation angle θ _RX viewed from the non-scanning direction (Equation 2), and the correction is performed when viewed from the scanning direction. The rotation angle _θRY is not used. However, for example, when measuring the amount of bending of the movable mirror 21X, the rotation angle θ viewed from the scanning direction is used.
The rotation angle of the reticle 12 is controlled by _RY . That is,
When measuring the amount of bending of the movable mirror 21X, the reticle minute drive stage 11 is scanned in the Y direction so that the rotation angle θ _RY seen from the scanning direction becomes a constant value, and the interferometer main bodies 14X1 and 14X2 are moved at predetermined intervals. The difference between the measured values (XR1-XR
Sample 2). Assuming that the n-th (n = 1, 2,...) Sampled difference is ΔXDn, since the yawing of the reticle micro-drive stage 11 is sequentially corrected, the difference ΔXDn is purely the movable mirror 21.
X bending information, reticle minute drive stage 11
In each of the Y coordinates, the difference ΔXDn up to that point is integrated, and at the intermediate Y coordinate, the amount of bending before and after is interpolated, whereby the amount of bending of the movable mirror 21X is obtained as a function of the Y coordinate. The bending amount of the movable mirror 21X may be obtained as a function of the Y coordinate using a spline function or the like.

Thereafter, when scanning exposure is performed, the amount of bending of the movable mirror 21X is determined, for example, with respect to the X coordinate XR (= (XR1 + XR2) / 2) actually measured by the movable mirror 21X of the reticle minute drive stage 11. By adding
An accurate X coordinate of the reticle minute drive stage 11 in which the bending amount of the movable mirror 21X has been corrected is obtained. By using the corrected X coordinate, the reticle minute drive stage 11 (reticle 12) is linearly scanned, so that the shape of the shot area exposed on the wafer becomes a more accurate rectangle.

In FIG. 2A, a two-axis laser interferometer for the non-scanning direction is constituted by the movable mirror 21X and the interferometer main bodies 14X1 and 14X2.
The 1Y2, the fixed plane mirrors 14M1 and 14M2, and the interferometer main bodies 14Y1 and 14Y2 constitute a two-axis laser interferometer for the scanning direction. Then, the interferometer body 14X1,
A coordinate system consisting of an X coordinate XR measured by 14X2 and a Y coordinate YR measured by the interferometer bodies 14Y1 and 14Y2 is defined as a coordinate system (XR, Y) of the reticle stage.
R). Although this coordinate system may be somewhat different from an ideal rectangular coordinate system in design consisting of the X axis and the Y axis, the reticle 12 is based on the coordinate system (XR, YR) of the reticle stage. Driven.

Next, in FIG. 2B, a wafer 5 is held on the Zθ axis drive stage 4 by vacuum suction or the like, and a reference mark plate 6 is fixed near the wafer 5. Further, an image of a part of the pattern of the reticle 12 is projected on a slit-shaped exposure area 62 conjugate with an illumination area 61 on the reticle on the wafer 5, and the wafer 5 is scanned in the Y direction with respect to the exposure area 62. Thereby, the pattern of the reticle 12 is transferred to one shot area on the wafer 5. A movable mirror 7X made of a parallel plate glass extending in the scanning direction (Y direction) is provided at an end of the Zθ axis drive stage 4 in the −X direction, and moves to an end of the Zθ axis drive stage 4 in the + Y direction. A movable mirror 7Y made of a parallel plate glass extending in the non-scanning direction is fixed to be orthogonal to the mirror 7X. The measurement surfaces LWX1 and LWX2 are radiated in parallel to the reflecting surface of the movable mirror 7X at intervals L3 in the Y direction from the interferometer bodies 13X1 and 13X2.
The measurement beams LWY1 and LWY2 are radiated in parallel to the reflection surface of the movable mirror 7Y at intervals L4 in the X direction from 1 and 13Y2. The interferometer body 13FX is arranged at a position symmetrical to the interferometer body 13X2 with respect to the interferometer body 13X1, and the measurement beam LWX1 is transmitted from the interferometer body 13FX.
The movable mirror 7X is irradiated with the measurement beam BFX at an interval L3 in the Y direction.

The movable mirrors 7X and 7Y scan and expose the wafer 5,
Alternatively, it is formed long enough so that the corresponding measurement beam does not come off during stepping. Further, the measurement beams LWX1 and LWX2 are emitted from the slit-shaped exposure region 6.
2 are arranged in the Y direction with respect to the center of 2 (optical axis AX).
The position (X coordinate) XW of the wafer 5 in the non-scanning direction is detected based on the average value of the measurement values XW1 and XW2 of 13X2 and the measurement values YW1 of the interferometer bodies 13Y1 and 13Y2.
According to the average value of YW2, the position (Y
(Coordinates) YW are detected. Also, the measured values XW1, XW2
Is divided by the interval L3, the yawing amount (rotation angle) θ _{WX of the} wafer 5 is detected, and the measured value YW
1, the rotation angle θ _WY obtained by dividing the difference of YW2 by the interval L4;
The moving mirrors 7X and 7X are determined by the difference from the yawing amount θ _WX.
An orthogonality error ΔωW of Y is detected. That is, the following equation holds.

[0045]

XW = (XW1 + XW2) / 2

[0046]

## EQU7 ## YW = (YW1 + YW2) / 2

[0047]

Equation 8 θ _WX = (XW1-XW2) / L3

[0048]

## EQU9 ## ΔωW = (YW1-YW2) / L4- (XW1
−XW2) / L3 Further, in this example, the measurement beams LWX1 and BFX are arranged so as to be distributed in the Y direction with respect to the optical axis that is the detection center of the off-axis type alignment sensor 34. When the alignment is performed using the sensor 34, the position (X coordinate) XW ′ (the following equation) of the wafer 5 in the non-scanning direction is detected based on the average value of the measured values XW1 and XWF of the interferometer bodies 13X1 and 13FX. The wafer stage is driven based on the X coordinate XW '.

[0049]

XW ′ = (XW1 + XWF) / 2 At this time, since the interval between the measurement beams LWX1 and LWX2 is equal to the interval between the measurement beams BFX and LWX1, even if the movable mirror 7X is bent, the wafer stage Is switched from (Equation 6) to (Equation 10), the effect of the bending of the movable mirror 7X is reduced.

The center of the exposure area 62 (projection optical system 8
The distance IL between the optical axis AX of the alignment sensor 34 and the optical axis of the alignment sensor 34 in the Y direction is a baseline amount of the alignment sensor 34 in the Y direction, and the distance between the centers of these in the X direction (0 in this example) is the alignment. This is the base line amount of the sensor 34 in the X direction. Further, the RA microscopes 19 and 2 of FIG. 1 are provided at both ends in the X direction of the slit-shaped illumination area 61 of FIG.
0 observation regions 19R and 20R are set, and observation regions 19W and 20W conjugate to those observation regions are set at both ends in the X direction of the slit-shaped exposure region 62 in FIG. 2B.

In FIG. 2B, a three-axis laser interferometer for the non-scanning direction is constituted by the moving mirror 7X and the interferometer bodies 13X1, 13X2, and 13FX, and is scanned by the moving mirror 7Y and the interferometer bodies 13Y1 and 13Y2. A two-axis laser interferometer for directions is configured. Thus, the interferometer body 13
X coordinate XW (XW 'at the time of alignment) measured by X1, 13X2, and interferometer main bodies 13Y1, 13
A coordinate system including the Y coordinate YW measured by Y2 is referred to as a coordinate system (XW, YW) of the wafer stage. This coordinate system may be somewhat different from an ideal rectangular coordinate system in design consisting of the X axis and the Y axis. However, the scanning and stepping of the wafer 5 are performed by the coordinate system (XW , YW). For example, when the correction of the orthogonality error ΔωW in (Equation 9) is not performed, the stepping direction of the Zθ axis drive stage 4 (wafer 5) in the scanning direction is the direction along the reflection surface of the movable mirror 7X (the X coordinate XW is In the non-scanning direction, the direction is the direction along the reflecting surface of the movable mirror 7Y (the direction in which the Y coordinate YW does not change).

The correspondence between the coordinate system of the reticle stage and the coordinate system of the wafer stage is determined via the reference mark plate 6. Therefore, a plurality of alignment marks are formed on the reticle 12, and corresponding reference marks and reference marks for baseline measurement are formed on the reference mark plate 6. FIG. 4A is a plan view of the reticle 12. In FIG. Y on both sides in the scanning direction (± X direction)
Linear rough alignment marks 27 and 28 are formed along the direction. Rough alignment mark 2
Both ends of 7 and 28 are cross-shaped, respectively. Further, the light-shielding band 31 and the rough alignment marks 27 and 2
8, fine alignment marks 29A to 29D and 30A to 30D are formed along the Y direction, respectively. In this case, the fine alignment marks 29A, 29D, 30A, and 30D are arranged at the vertices of an accurate rectangle, and the fine alignment marks 29B, 2B
9C, 30B and 30C are symmetrically arranged on the sides of the rectangle. And fine alignment mark 29A
4D to 30D and 30A to 30D have a shape in which three light-shielding patterns are commonly arranged symmetrically at four positions in a frame shape as shown in FIG. 4C. The horizontal lines of the cross-shaped pattern at both ends of the rough alignment marks 27 and 28 pass through the centers of the fine alignment marks 29A and 29D and the centers of the fine alignment marks 30A and 30D, respectively. Although the reticle 12 of this example moves in a wide range in the Y direction, the rough alignment mark 2
By using 7 and 28, fine alignment marks 29A to 29D and 30A to 30D can be easily detected.

FIG. 4 (b) shows the RA microscopes 19 and 2 of FIG.
FIG. 4 shows the observation areas 19R and 20R according to FIG.
In (b), a slit-shaped illumination area 61 is set so as to be inscribed in a circular area 33R conjugate with the effective exposure field of the projection optical system 8 in FIG. In the observation area 19R, 2
0R is set. In FIG. 4B, the reticle 12
Are moved in the Y direction, whereby the fine alignment marks 29D and 30D on the reticle 12 are located substantially at the centers of the observation regions 19R and 20R, respectively.
By moving the reticle 12 in the Y direction in this state, the fine alignment marks 29C to 29A, 30C to 30C are sequentially arranged substantially at the centers of the observation regions 19R and 20R.
30A is set.

FIG. 6A shows the reticle 12 shown in FIG.
Is shown on the reference mark plate 6 shown in FIG. 2B, and a reticle image 12W obtained by the projection is shown in FIG.
(A) Fine alignment marks 29A to 29D
Mark images 29AW to 29DW conjugate to the fine alignment marks 30A to 30D
AW to 30 DW are shown. Each mark image 29AW
29DW and 30AW-30DW are respectively shown in FIG.
As shown in (b), the shape is such that three linear patterns are arranged in a frame shape.

FIG. 6C shows the arrangement of fiducial marks on the fiducial mark plate 6. On the fiducial mark plate 6 of FIG. 6C, the mark images 29AW-29DW and FIG. 3
Reference marks 35A to 35D and 36A to 36D are formed in substantially the same arrangement as 0AW to 30DW. That is, the reference marks 35A, 35D, 36A, 36D
Are located at the vertices of the exact rectangle and the other fiducial marks 35
B, 35C, 36B and 36C are symmetrically arranged on the sides of the rectangle. The reference mark plate 6 is formed of a light-transmitting material having a very low coefficient of thermal expansion such as quartz or glass ceramic (for example, “Zerodur” manufactured by Schott), and these reference marks are formed on the reference mark plate 6. It is illuminated with illumination light having the same wavelength as the exposure light from the back surface or from an illumination system provided inside the reticle alignment microscopes 19 and 20 on the reticle 12. Further, on the reference mark plate 6, a reference mark 37A is formed at a position separated from the middle point of the reference marks 35A and 36A by an interval IL in the Y direction which is the scanning direction. The interval IL is shown in FIG.
Is equal to the baseline amount of the off-axis type alignment sensor 34 in the Y direction. Similarly, fiducial mark 35
The reference marks 37B, 37B are located at positions spaced apart from the midpoints of B, 36B, the midpoints of the reference marks 35C, 36C, and the midpoints of the reference marks 35D, 36D by the distance IL in the Y direction, respectively.
C and 37D are formed.

Reference marks 35A-35D, 36A-36
D is a linear pattern of 7 rows × 7 columns as shown in FIG.
A to 35D and 36A to 36D are mark images 2 in FIG.
It is a size that fits inside 9AW to 30DW. Also,
As shown in FIG. 6E, the reference marks 37A to 37D use corresponding grid points in a grid pattern formed at a predetermined pitch in the X direction and the Y direction.

Next, basic operations for performing alignment, scanning exposure, and stepping with the projection exposure apparatus of this embodiment will be described. First, reticle alignment is performed using the reference mark plate 6 shown in FIG. That is, by driving the wafer Y-axis drive stage 2 and the wafer X-axis drive stage 3 of FIG.
After moving 5A and 36A into the observation areas 19W and 20W on both sides of the exposure area 62 in FIG. 2B and stopping the movement, the reticle Y-axis drive stage 10 is driven and the reticle 12 in FIG.
Fine alignment marks 29A and 30A
Observation areas 19R, 2 on both sides of the illumination area 61 in FIG.
Move into 0R. Then, the RA microscopes 19 and 2 in FIG.
0, the amount of misalignment between the reference marks 35A, 36B and the corresponding fine alignment marks 29A, 30A is detected, and the reticle Y-axis driving stage 10 and the reticle minute driving stage 11 are driven to obtain the reference marks 35.
A, the center of the image of 36A matches the center of the fine alignment marks 29A, 30A, and the reference mark 35
The fine alignment marks 29A and 30A are aligned with respect to the images of A and 36A so that the positional deviation amounts are symmetrical. Thereby, the position and the rotation angle of the reticle 12 are aligned with the reference mark plate 6. In this state, for example, by resetting the measurement value of the 4-axis interferometer body 14 on the reticle stage side and the measurement value of the 5-axis interferometer body 13 on the wafer stage side, (Equation 1)
And the coordinate system of the reticle stage (X
R, YR) and the origin offset between the wafer stage coordinate system (XW, YW) determined by (Equation 6) and (Equation 7) are corrected.

In addition, the scanning direction of the Zθ-axis driving stage 4 on the wafer stage side during the scanning exposure is made parallel to the arrangement direction of the reference marks on the reference mark plate 6 in advance. For this purpose, as an example, the arrangement direction of the reference marks 35A to 35D (the direction of the reference mark plate 6) may be mechanically made parallel to the direction of the reflection surface of the movable mirror 7X (the running direction of the wafer stage). However, in this example, since the inclination angle of the running direction of the wafer stage with respect to the direction of the reference mark plate 6 can be easily measured as described later, when a mechanical adjustment error remains, the Y coordinate YW The scanning direction of the Zθ-axis driving stage 4 may be corrected by software so that the X coordinate XW changes by a corresponding amount every time. Hereinafter, the coordinate system in which the scanning direction corrected in this way is the Y axis is referred to as the coordinate system (XW, YW) of the wafer stage.

Next, without irradiating the exposure light, the Zθ axis driving stage 4
And the reticle minute drive stage 11 are scanned synchronously, and the relative positional deviation between the reference marks 35A to 35D on the reference mark plate 6 and the corresponding fine alignment marks 29A to 29D on the reticle 12 is determined by the RA microscope 1.
Detected sequentially by 9 and 20. From the average value of these relative displacement amounts, the inclination angle between the scanning direction of the reticle 12 and the scanning direction of the wafer 5, that is, the coordinate system (XR, YR) of the reticle stage and the coordinate system (XW,
YW) and the rotation angle of the axis in the scanning direction. Thereafter, when the reticle 12 is scanned, the X coordinate XR is laterally shifted by a corresponding amount while the Y coordinate YR changes by a predetermined interval via the reticle Y axis drive stage 10 and the reticle minute drive stage 11, thereby making the software The scanning direction of the reticle 12 is matched with the arrangement direction of the reference marks on the reference mark plate 6. Hereinafter, the coordinate system having the scanning direction corrected in this manner as the Y axis is referred to as the reticle stage coordinate system (XR, YR). As a result, the coordinate system (XW, YW) of the wafer stage and the coordinate system (X
(R, YR), the axes in the scanning direction are parallel to each other with reference to the reference mark plate 6, and the reticle 12 and the wafer 5 are scanned in parallel during scanning exposure.

In this case, the movement of each stage is based on the guide surface of each stage.
For example, the parallelism between the guide surface of the reticle Y-axis drive stage 10 and the guide surface of the wafer Y-axis drive stage 2 is several hundreds.
It is mechanically adjusted to about μrad or less. Further, by fixing the movable mirror and the reference mark plate 6 together on the guide surfaces, the amount of software correction by driving each stage also in the non-scanning direction during scanning exposure is reduced, and the control accuracy is reduced. Has been improved. When the reticle 12 is actually mounted on the reticle micro-drive stage 11 adjusted in this way, when the reticle 12 is set on the basis of the outer shape or the like, the fineness of the reticle 12 is adjusted with respect to each movable mirror and the reference mark plate 6.・ Alignment marks 29A-3
There is a possibility that only 0D is rotating greatly. this is,
This is because the amount of displacement between the outer shape of the reticle and the transfer pattern is about 0.5 mm when large.

If the positional deviation between the outer shape of the reticle 12 and the transfer pattern shown in FIG. 2A is large, the positions of the fine alignment marks 29A to 29D of the reticle 12 and the reference marks 35A to 35D of the reference mark plate 6 are changed. When the displacement amount is measured, the reticle 12 or the reference mark plate 6
Is measured to have a large rotation or a large offset. However, since the reference mark plate 6 is fixed according to the running direction of the movable mirrors 7X and 7Y,
The correction is performed by rotating or shifting the reticle minute drive stage 11. When the reticle minute drive stage 11 is rotated in this manner, the movable mirror 21X also rotates in the same manner, so that the movable mirror 2X moves in the running direction of the reticle 12.
1X will be tilted.
The alignment marks 29A to 29D are parallel to the reference marks 35A to 35D on the reference mark plate 6, and are controlled so that the running direction of the reticle 12 and the running direction of the wafer 5 are parallel during scanning exposure.

Next, wafer alignment is performed to determine the alignment of each shot area on the wafer 5 on the coordinate system (XW, YW) of the wafer stage. As an example, a predetermined number of shot areas (sample shots) selected from on the wafer 5 using the alignment sensor 34 of FIG.
The coordinates of all the shot areas on the wafer 5 are calculated by the EGA (Enhanced Global Alignment) method that measures the coordinates of the wafer mark of the above and statistically processes the measurement result. In addition, a baseline amount of the alignment sensor 34 is determined in advance by using a reference mark plate 6 in a later-described baseline measurement process (interval / baseline check) and stored in the main control system 22A. Therefore, based on the arrangement coordinates of each shot area on the wafer 5, the baseline amount of the alignment sensor 34, and the relationship between the coordinate system (XW, YW) of the wafer stage and the coordinate system (XR, YR) of the reticle stage. The exposure target shot area on the wafer 5 is positioned at the scanning start position, and the reticle 12 is also positioned at the corresponding position.

After that, scanning exposure is performed in accordance with the coordinate system (XW, YW) of the wafer stage and the coordinate system (XR, YR) of the reticle stage determined at the time of the previous reticle alignment. , 7Y, 2
This is corrected by software with reference to the reflection surfaces of the 1X and the corner cubes 21Y1, 21Y2. If the positions of these movable mirrors are relatively displaced with respect to the reticle 12 or the wafer 5, the shape of the shot area And the shot arrangement. In this example, scanning exposure and stepping are performed by the following method so that an accurate rectangular shot area and an orthogonal lattice shot arrangement are formed even in such a case.

That is, the coordinates of the reticle stage coordinate system (XR, YR) when the shot area to be exposed and the reticle are aligned by wafer alignment are (XR ₀ , YR ₀ ), and the coordinate system of the wafer stage. (XW, Y
If the coordinates of (W) are (XW ₀ , YW ₀ ), since the projection magnification of the projection optical system 8 is 1 / M, the reticle minute drive stage 11 (reticle 12) and the Zθ axis drive stage 4 (wafer) The synchronization errors ΔX and ΔY in the scanning direction with respect to 5) and in the non-scanning direction can be expressed as follows. However,
These synchronization errors are errors converted on the reticle 12. Although the projection optical system 8 in FIG. 1 is a reverse projection system,
As shown in FIG. 2, since the measurement directions of the interferometer on the reticle stage side and the interferometer on the wafer stage side are reversed in the non-scanning direction, the synchronization error is simply the difference between the magnification correction values of the movement amount. Is required. On the other hand, since the moving mirrors of both stages are in the same direction in the scanning direction, for example, in the interferometer main bodies 14Y1 and 14Y2 on the reticle side, the signs are determined so that the measured values decrease when the reticle minute drive stage 11 moves away. Shall be kept.

[0065]

ΔX = (XW−XW ₀ ) · M− (XR−XR ₀ )

[0066]

ΔY = (YW−YW ₀ ) · M− (YR−YR ₀ )
In this example, the rotation angle θ _{WX of the} Zθ axis driving stage 4 as viewed from the non-scanning direction expressed by (Equation 9) and the rotation angle θ _WX of the reticle minute driving stage 11 expressed by (Equation 2) are viewed. The difference from the rotation angle θ _RX is defined as a synchronization error Δθ in the rotation direction as follows.

[0067]

ΔX = θ _WX −θ _RX = (XW1−XW2) / L3− (XR1−XR2) / L1 At the time of scanning exposure, the reticle Y-axis drive stage 10 and wafer Y-axis drive stage 2 of FIG. Acceleration is started, and after each of them reaches a predetermined scanning speed, the reticle minute drive stage 11 is driven so that the above-mentioned synchronization errors ΔX, ΔY, Δθ become 0, respectively, to perform synchronization control. After a predetermined settling time has elapsed in this state, irradiation of exposure light on the illumination area 61 on the reticle 12 is started, and exposure is performed.

Thereafter, when the stepping of the wafer 5 is performed in order to expose the next shot area, even if the orthogonality of the movable mirrors 7X and 7Y on the wafer stage side is deteriorated, the shot array has an orthogonal lattice ( The stepping direction in the non-scanning direction of the Zθ-axis driving stage 4 is changed so that the orthogonality error Δω of (Equation 9) is maintained so as to maintain the shape of the grid in which the arrangement direction is orthogonal.
Correct by W.

Further, when the orthogonality error ΔωW of (Expression 9) or the orthogonality error ΔωR of (Expression 5) greatly changes beyond a predetermined allowable value, the other off-axis type alignment sensor 34 Problems may have occurred in the accuracy of the baseline amount and its mechanical stability. Therefore, when the orthogonality error ΔωW or ΔωR greatly changes beyond a predetermined allowable value, the above-described reticle alignment and baseline check are performed again when the wafer is replaced. Thereby, the overlay accuracy of the reticle pattern and each shot area of the wafer can be improved.

Next, an example of a baseline check (interval / baseline check) that is periodically performed by the projection exposure apparatus of this embodiment, for example, will be described with reference to the flowchart of FIG. In this example, so-called dynamic reticle alignment is also performed in parallel with the baseline check. First, in step 101 of FIG.
The fine alignment marks 29A and 30A on the reticle 12 in FIG. 4A are moved into the observation areas 19R and 20R of the RA microscopes 19 and 20 in FIG. Thereafter, in step 102, the reference marks 35A, 36A on the reference mark plate 6 in FIG. 6C are respectively set to the observation regions 19W, 20W on the wafer stage conjugate to the respective observation regions 19R, 20R (FIG. 2B). See).

FIG. 5A shows a state in which the mark of the reticle 12 and the reference mark on the reference mark plate 6 have been moved into the corresponding observation area, as shown in FIG. 5A. The mark image 29AW and the reference mark 35A can be observed simultaneously in the observation region 19W, and the mark image 30AW and the reference mark 36A can be observed simultaneously in the observation region 20W. As shown in FIG. 5C, the observation regions 19W and 20W are located at positions crossing the optical axis in the exposure field of the projection optical system 8, respectively, and are within the observation region of the off-axis type alignment sensor. The reference mark 37A is set. Thereafter, in step 103, the RA microscope 1
In 9 and 20, the observed image is converted into an image signal,
The image pickup signal is processed, and the displacement amount of the mark image 29AW with respect to the reference mark 35A and the reference mark 3
The position shift amounts of the mark image 30AW with respect to 6A are obtained, and these position shift amounts are supplied to the main control system 22A of FIG.
At the same time, the off-axis alignment sensor 34 also captures an image of the corresponding reference mark 37A, and the reference mark 37A obtained by processing this image signal is displaced from the detection center (the center of the index mark, etc.). The amount is supplied to the main control system 22A.

Thereafter, at step 104, the Zθ axis driving stage 4 shown in FIG.
The reticle minute drive stage 11 shown in FIG. 2A is moved in the −Y direction in synchronization with the movement in the Y direction. Actually, the moving direction of the Zθ axis driving stage 4 is a direction in which the value of the coordinate XW does not change in the coordinate system (XW, YW) of the wafer stage, and the moving direction of the reticle minute driving stage 11 is the coordinate system of the reticle stage. Coordinate X in (XR, YR)
This is the direction in which the value of R does not change. As a result, FIG.
As shown in (b), the reference mark plate 6 and the reticle image 12
Both W move in the + Y direction. At this time, since the observation regions 19W and 20W of the RA microscopes 19 and 20 and the observation region of the off-axis type alignment sensor 34 are fixed, the observation regions 19W and 20W and the observation region of the alignment sensor 34 are denoted by reference numerals. A mark group marked with A (the mark images 29AW, 30AW and the reference mark 35)
A, 36A, 37A), a mark group (mark images 29DW, 30DW, reference mark 35)
D, 36D, and 37D) move. At this time,
The mark group denoted by reference symbol B, the mark group denoted by reference symbol C, and the mark group denoted by reference symbol D are sequentially arranged in the observation region 19.
When entering the observation area of W, 20W and the alignment sensor 34, the Zθ axis drive stage 4 and the reticle minute drive stage 11 are stopped to detect the position of each mark.

That is, assuming that the state shown in FIG. 5A is the first stationary position, the observation regions 19W and 20W at the second stationary position.
The mark group existing in the observation region of the alignment sensor 34 is a mark group denoted by a symbol B, that is, mark images 29BW and 30BW in FIG. 6A and reference marks 35B, 36B and 37B in FIG. 6C. is there. Then, the RA microscopes 19 and 20 determine the positional deviation amounts of the mark images 29BW and 30BW with respect to the reference marks 35B and 36B and supply them to the main control system 22A, and the alignment sensor 34 determines the positional deviation amounts of the corresponding reference marks 37B. Main control system 22
A. By repeating the above sequence with the third stationary position and the fourth stationary position (the state of FIG. 5B), the mark group (the mark image 29)
CW, 30CW, and fiducial marks 35C, 36C, 37
C) and the mark group denoted by the symbol D, the RA microscopes 19 and 20 and the alignment sensor 34, respectively.
Position measurement is performed. The amount of displacement in the coordinate system of the wafer stage measured by the RA microscopes 19 and 20 with respect to the eight mark images 29AW to 30DW is (ΔXn, Δ
Yn) (n = 1 to 8) as the alignment sensor 34
The positional deviation amount from the center of detection in the coordinate system of the wafer stage measured with respect to the four reference marks by (ΔAXi,
ΔAYi) (i = 1 to 4).

Next, at step 105, dynamic reticle alignment is performed by processing the obtained measurement data. At this time, in this example, in order to increase the measurement accuracy,
Reticle alignment is performed with reference to a coordinate system determined by the arrangement direction of the reference marks on the reference mark plate 6 in FIG. 6C (hereinafter, referred to as “reference mark plate coordinate system”). As an example, the coordinate system of the reference mark plate is such that a straight line passing through the reference marks 35A and 36A of the reference mark member 6 is represented by a horizontal axis (this is referred to as an XS axis), and the reference marks 35A and 35D.
Is a coordinate system (XS, YS) in which a straight line passing through is referred to as a vertical axis (this is called a YS axis). Also, scaling of the coordinate system (XR, YS) of the reticle stage onto the reference mark plate 6 in the XS and YS directions (linear expansion and contraction) with respect to the coordinate system (XS, YS) of the reference mark plate. ) To R
x, Ry, rotation (rotation) are θ, orthogonality error is ω, and offsets in the XS and YS directions are Ox and Oy. In this case, the rotation θ is the rotation angle of the axis on which the YR axis of the coordinate system of the reticle stage is projected with respect to the YS axis, that is, the rotation error in the scanning direction of the reticle stage, and the rotation θ is the relative rotation angle θ2 in the present invention.
Corresponding to

Further, the mark images 29AW-3A shown in FIG.
Design coordinates in the coordinate system obtained by projecting the coordinate system of the 0DW reticle stage onto the reference mark plate 6 are (Dxn, Dy
n) (n = 1 to 8) as the corresponding reference mark 35A
The coordinates on the coordinate system (XS, YS) of ~ 36D are represented by (Ex
n, Eyn). At this time, RA microscopes 19 and 20
Using the displacement amounts (ΔXn, ΔYn) measured at
Measured coordinates (Dxn ', Dy) of the mark images 29AW to 30DW in the coordinate system (XS, YS) of the reference mark plate 6
n ′) is approximately as follows.

[0076]

Dxn ′ = Exn + ΔXn Dyn ′ = Eyn + ΔYn At this time, the above-mentioned six conversion parameters (Rx, Ry,
θ, ω, Ox, Oy) and mark images 29AW-30D
From the design coordinates (Dxn, Dyn) of W, mark image 2
9AW to 30DW reference mark plate coordinate system (XS, Y
The calculated coordinates (Fxn, Fyn) on S) are expressed as follows.

[0077]

(Equation 15)

The XS of the mark images 29AW-30DW
Coordinates (Fxn, Fyn) in the YS and YS directions
And the difference between the actually measured coordinates (Dxn ′, Dyn ′), that is, the nonlinear error (εxn, εyn) are as follows.

[0079]

(Equation 16)

Then, the nonlinear errors (εxn, εy
n), the main control system 22A of FIG. 1 uses the least squares method to minimize the sum of squares of the eight mark images 29AW to 30DW.
Determine the values of y, θ, ω, Ox, Oy. Next, in step 106, the coordinates of the reticle stage are set so as to have the coordinates obtained by multiplying the determined scaling Rx, Ry by the coordinates of the coordinate system (XR, YR) of the reticle stage, and to set the determined rotation θ to zero. System (XR, Y
The coordinate system obtained by rotating R) is newly set as the coordinate system (XR, YR) of the reticle stage, and thereafter, the reticle minute drive stage 11 (reticle 12) is scanned along this new coordinate system. This corresponds to the fine alignment marks 29A to 29A in FIG. 4A along the arrangement direction of the reference marks 35A to 35D on the reference mark plate 6 in FIG.
This means that the reticle 12 is scanned so that 29D moves. Note that the offsets Ox and Oy are corrected by wafer alignment, and therefore need not be corrected here. Further, it is not necessary to particularly correct the orthogonality error ω.

As a result, the pattern of the reticle 12 has the reference marks 35A to 35A arranged in a rectangular shape on the reference mark plate 6.
As a result of scanning along 36D, the shape of the shot area exposed on the wafer becomes an accurate rectangle. Further, even when the inclination angle between the movable mirror 7X on the wafer stage side and the reference mark plate 6 changes, the reference mark on the reference mark plate 6 is used as a reference, so that it is not affected by the change in the inclination angle. There is also an advantage.

In the next step 107, the main control system 2
In 2A, the displacement amounts (ΔXn, ΔYn) (n = 1 to 8) measured by the RA microscopes 19 and 20 and the displacement amounts (ΔAXi, ΔA) measured by the alignment sensor 34
Yi) (i = 1 to 4) is processed to calculate the baseline amounts (BX, B) of the alignment sensor 34 in the X direction and the Y direction.
Y) is calculated. That is, the reference mark 35 shown in FIG.
A, 36A, the positional shift amounts of the mark images 29AW, 30AW are (ΔX1, ΔY1), (ΔX2, ΔY2).
Assuming that the displacement amount measured in step 4 is (ΔAX1, ΔAY1), the baseline amount (BX1, BY1) for the mark group with the symbol A is as follows.

[0083]

BX1 = (ΔX1 + ΔX2) / 2−ΔAX1 BY1 = IL + (ΔY1 + ΔY2) / 2−ΔAY1 Similarly, the base line amounts (BX2, BY3) to (BX4, BY4) for the other three mark groups are calculated. Is done. By averaging these four baseline amounts, the baseline amounts (BX, BY) of the alignment sensor 34 in the X and Y directions are calculated. As described above, in this example, the baseline data is obtained by averaging the measurement data for the four mark groups, so that the baseline data can be measured with high accuracy.

In the present embodiment, the baseline amount is obtained by using the reference mark plate 6 as the "length measurement reference", instead of using the measurement value of the laser interferometer as a reference. This reference mark plate 6
Is made of a material having a low expansion coefficient such as quartz or glass ceramics, there is an advantage that the baseline amount can be measured with high accuracy without being affected by air fluctuation unlike a laser interferometer.

In this example, in order to enhance the averaging effect, the alignment sensor 34 uses a plurality of reference marks 37A.
The position detection of the alignment sensor 34 is performed, but when the measurement accuracy of the alignment sensor 34 is good, the baseline amount of the alignment sensor 34 may be obtained from only the measurement data for one reference mark 37A. When the baseline amount is obtained from the measurement data of the plurality of reference marks 37A to 37D, when the inclination angle between the movable mirror 7X on the wafer stage side and the reference mark plate 6 changes in FIG. Since an error occurs in the measured value of the baseline amount BX, the inclination angle between the movable mirror 7X and the reference mark array direction of the reference mark plate 6 is measured in advance, and the baseline amount is corrected based on the inclination angle. It is desirable to do.

FIG. 7A shows a case where the movable mirror 7X on the wafer stage side is inclined with respect to the reference mark plate 6. In FIG. The direction of the reflecting surface of the movable mirror 7X (the running direction of the Zθ-axis driving stage 4) with respect to the arrangement direction is inclined clockwise by an inclination angle Δθ. This inclination angle Δθ corresponds to the relative rotation angle θ1 in the present invention. Further, the observation areas 19W and 20W of the RA microscope, and the alignment sensor 34
The reference marks 35A, 36A, 3
7A is observed. FIG. 7B shows that the Zθ-axis drive stage 4 is moved from the state of FIG.
In the direction by the distance WY, the observation areas 19W, 20
W and a state where the reference marks 35D, 36D, and 37D are observed in the observation region of the alignment sensor 34.

The displacement amount δX of the Zθ axis driving stage 4 in the X direction in the state of FIG. 7B is approximately as follows.

[0088]

ΔX = Δθ · WY At this time, since the reticle 12 is scanned along the direction of arrangement of the reference marks on the reference mark plate 6, the reference mark 35
The result (BX4-δX) obtained by subtracting δX from the baseline amount BX4 in the X direction obtained from the measurement data of D, 36D, and 37D is the actual baseline amount BX4 ′. Similarly, the other mark groups (35B, 36B) in FIG.
For B, 37B) and (35C, 36C, 37C), the baseline amount in the X direction is also corrected, and the final baseline amount is obtained by averaging the corrected baseline amounts.

Here, an example of a method of measuring the inclination angle Δθ of the movable mirror 7X with respect to the reference mark plate 6 will be described with reference to the flowchart of FIG. First, in step 111 of FIG. 3, the reticle minute drive stage 11 is driven to move the image 29AW of the fine alignment mark 29A on the reticle 12 into the observation area 19W of the RA microscope 19, as shown in FIG. After moving, the reticle minute drive stage 11 (reticle 12) is stopped. In this state, the Zθ axis drive stage 4 is driven to move the reference mark 35A on the reference mark plate 6 into the observation area 19W, and the RA microscope 19 shifts the reference mark 35A relative to the mark image 29AW in the X direction. The amount ΔXA1 is detected and supplied to the main control system 22A.

Next, at step 112, reticle 1
By moving the Zθ-axis drive stage 4 (reference mark plate 6) along the movable mirror 7X by the interval WY while the stage 2 is stationary, as shown in FIG.
The reference mark 35D is moved into W. Then, the RA microscope 19 detects the amount of displacement ΔXD1 of the reference mark 35D in the X direction with respect to the mark image 29AW, and the main control system 22
A. In response to this, in step 113, the main control system 22A determines the amount of displacement ΔXA1, ΔXD1,
And the interval WY, the inclination angle Δθ of the movable mirror 7X with respect to the reference mark plate 6 is calculated as follows, and the calculated inclination angle Δθ
Is stored in the storage unit in the main control system 22A. By using the inclination angle Δθ stored in this way in (Equation 18), an error due to the inclination angle Δθ is corrected.

[0091]

Δθ = (ΔXA1−ΔXD1) / WY Further, at the time of scanning exposure, the scanning direction of the Zθ axis driving stage 4 of the wafer stage is set to the reference mark on the reference mark plate 6 so as to cancel the inclination angle Δθ. In the array direction. In addition, the stepping direction in the scanning direction between the shot areas is the arrangement direction of the reference marks, and the stepping direction in the non-scanning direction is a direction orthogonal to the arrangement direction. As a result, even when the inclination angle between the movable mirror 7X and the reference mark plate 6 fluctuates, the obtained shot arrangement becomes an orthogonal lattice shape. In this case, the Zθ axis driving stage 4
In order to correct the scanning direction, the Zθ axis driving stage 4
The Zθ axis driving stage 4 is moved every time the
May be corrected by a corresponding amount. Such a driving method is easy to control.

However, since the interval WY at the time of measuring the above-mentioned inclination angle Δθ is short, in order to improve the measurement accuracy of the inclination angle Δθ, the measurement of steps 111 to 113 is repeated a plurality of times, and these measurement results are averaged. Is desirable. Only when the orthogonality error ΔωW calculated from (Equation 9) is larger than a predetermined allowable value using the measurement data of the four-axis laser interferometer on the wafer stage side, the tilt angle of the movable mirror May be performed. In this case, in order to calculate the inclination angle Δθ with higher accuracy, measurement reference marks are formed at both ends of the reference mark plate 6 in the scanning direction, and the interval WY in (Equation 19) is made as long as possible. You may do so. Thus, the inclination angle Δθ can be obtained with high accuracy without performing a plurality of measurements.

As described above, according to this example, the interval
At the time of the baseline check, dynamic reticle alignment and measurement of the baseline amount can be performed at high speed. Further, for example, only when the orthogonality error ΔωW becomes large, that is, when there is a possibility that the tilt angle Δθ of the movable mirror 7X may have drifted, the measurement sequence of the tilt angle Δθ is executed to execute the exposure process. The baseline amount can be measured with high accuracy without lowering the throughput.

In the above-described embodiment, the tilt angle of the movable mirror 7X on the wafer stage side with respect to the reference mark plate 6 is measured, but the running direction of the reticle stage (along the reflecting surface of the movable mirror 21X). Reticle 1 for direction
2 may be measured.
For this purpose, FIG. 7 corresponds to step 111 in FIG.
As shown in (a), for example, the reference mark 35A on the reference mark plate 6 is moved into the observation area 19W of the RA microscope 19 to stop the reference mark plate 6, and then the reticle 12
Is moved into the observation area 19W, and the amount of displacement is measured by the RA microscope 19.

Thereafter, corresponding to step 112, FIG.
By driving the reticle micro-drive stage 11 along the movable mirror 21X at a predetermined interval (a), for example, the mark image 29DW of FIG. Is measured. afterwards,
By dividing the difference between the positional deviation amounts at the two locations by the predetermined interval, the fine alignment marks 29A and 29A with respect to the running direction of the reticle minute drive stage 11 can be obtained.
The inclination angle Δθ _{R in the} 9D arrangement direction (that is, the arrangement direction of the patterns on the reticle 12) is calculated. This inclination angle Δθ
_R corresponds to the relative rotation angle θ3 of the present invention.

In this case, the reticle minute drive stage 11 is rotated by the above-described dynamic reticle alignment so that the pattern arrangement direction on the reticle 12 is parallel to the reference mark arrangement direction on the reference mark plate 6. Correct the angle. Thereafter, at the time of scanning exposure, the rotation of each shot area on the wafer is performed by scanning the reticle minute drive stage 11 along the arrangement direction of the pattern of the reticle 12 so as to cancel the above-mentioned tilt angle Δθ _R. (Rotation) is eliminated, and a rectangular shot area is accurately exposed in an orthogonal lattice shot arrangement. Also in this case, in order to correct the scanning direction of the reticle micro-drive stage 11, the position of the reticle micro-drive stage 11 in the X-direction by a corresponding amount every time the reticle micro-drive stage 11 moves by a predetermined distance in the Y-direction. It may be corrected.

Next, another example of the embodiment of the present invention will be described with reference to FIG. In the present embodiment, the tilt angle Δθ of the above-described movable mirror 7X is prevented by integrating the movable mirror on the wafer stage side and the reference mark plate. Also in this example, a projection exposure apparatus similar to the projection exposure apparatus shown in FIGS. 1 and 2 is used, but the configuration of the Zθ axis drive stage of the wafer stage is different.

FIG. 8A is a plan view showing the configuration of the upper part of the wafer stage of the projection exposure apparatus of the present embodiment, and FIG. Zθ in FIG.
A wafer 5 is mounted on a Zθ axis driving stage 4A corresponding to the axis driving stage 4. Further, an X-axis movable mirror 41 made of a parallel plate glass extending in the Y direction is provided at each end of the Zθ axis drive stage 4A in the −X direction and the + Y direction.
A movable mirror 41Y on the X axis and the Y axis made of a parallel plate glass extending in the X direction is fixed. The moving mirror 41X of this example,
41Y is made of a low expansion coefficient glass ceramic (for example, a trade name of Zerodur can be used) and the outer surfaces of the movable mirrors 41X and 41Y are reflection surfaces, respectively. It is set to the same height as the surface. Furthermore, the moving mirror 41X
And 41Y, a plurality of reference marks are formed by chrome evaporation.

That is, on the movable mirror 41X, the linear patterns 42X parallel to the reflection surface and the reference patterns 35A, 35D and 36A having the same shape and arrangement as the reference patterns on the reference mark plate 6 in FIG. , 36D, 37A, 37D
Are formed on the movable mirror 41Y, and a linear pattern 42X parallel to the reflection surface, and the observation areas 19W and 2 of the RA microscopes 19 and 20 shown in FIG.
Cross-shaped reference patterns 43X and 44X arranged at the same interval as the interval of 0W are formed. Further, the three interferometer main bodies 13X shown in FIG.
Measurement beams LWX1, LWX2, and BFX are emitted from 1, 13X2, and 13FX (not shown in FIG. 8), and a measurement beam LTX is emitted from the added interferometer body 13TX. The measurement beam LTX passes through an intermediate position between the two measurement beams LWX1 and LWX2 in the Y direction. However, as shown in FIG. 8B, the heights (positions in the Z direction) of the three measurement beams LWX1, LWX2, and BFX are the same, whereas the height of the measurement beam LTX is With respect to the distance H.

Also, the Y-axis movable mirror 41X has the structure shown in FIG.
The measurement beams LWY1 and LWY2 are emitted from the two interferometer main bodies 13Y1 and 13Y2 (not shown in FIG. 8) shown in FIG. The beam LTY is being irradiated. The height of the measurement beam LTY is also measured by the measurement beam LTX.
Similarly to the above, the beam is shifted by an interval H as compared with the other measurement beams. Other configurations are the same as those in FIG. 1 and FIG.

When performing baseline measurement in this example,
At the same time that the reference marks 35A and 36A on the movable mirror 41X in FIG. 8A are observed by the RA microscopes 19 and 20 in FIG. 1, the reference marks 37A are observed by the alignment sensor 34 in FIG. , The Zθ axis drive stage 4A and the reticle minute drive stage 11 are moved, and the RA microscopes 19 and 20 and the alignment sensor 34 measure the amount of displacement corresponding to the reference marks 35D, 36D and 37D on the movable mirror 41X. do it. At this time, since the movable mirror 41X also serves as the reference mark plate, even if the wafer stage undergoes thermal deformation or the like, the direction of the reflection surface of the movable mirror 41X and the arrangement direction of the reference marks may not change. Absent. Therefore, the baseline amount can always be measured with high accuracy.

When measuring the amount of bending of the movable mirror 41X, for example, a linear pattern 42X on the movable mirror 41X is used.
Is moved to the projection position of the mark image 29AW of the reticle in FIG. 6A, and the amount of displacement of the linear pattern 42X with respect to the mark image 29AW is measured by the RA microscope 19 in FIG. Then, the Zθ axis driving stage 4A is
The change in the amount of displacement of the linear pattern 42X with respect to the mark image 29AW may be measured by the RA microscope 19 while moving along the X. In this case, the amount of bending of the movable mirror 41X can be measured directly and manually. In addition, the other RA microscope 20 or the alignment sensor 34 in FIG. 1 may be used for measuring the amount of bending of the movable mirror. Similarly,
The bending amount of the Y-axis movable mirror 41Y can be easily measured by using the linear pattern 42Y.

In this example, the moving mirror 41X and the moving mirror 41 are used.
A change in the degree of orthogonality with Y can also be measured. For this purpose, for example, at the time of assembly adjustment, the reference marks 43X and 44X on the movable mirror 41Y in FIG. One RA microscope 19 and 20 may be used to detect the amount of displacement from a predetermined reference point, and the difference between the amounts of displacement of the two marks may be obtained and stored. Thereafter, for example, every time a predetermined time elapses, the RA microscope 19, 2
By detecting the difference in the amount of displacement between the reference marks 43X and 44X at 0, and dividing the detection result by the interval between the observation areas of the RA microscopes 19 and 20, the orthogonality error caused by the change in the tilt angle of the movable mirror 41Y is obtained. Can be detected.

In this example, the height of the upper surfaces of the movable mirrors 41X and 41Y needs to be the same as the surface of the wafer 5, and the heights of the measurement beams LWX1 and LWY1 are set to the height of the surface of the wafer 5. Therefore, a measurement error due to a so-called Abbe error may occur due to pitching or rolling of the Zθ axis drive stage 4A. In order to correct this measurement error, in this example, interferometer main bodies 13TX and 13TY are provided as tilt interferometers. Specifically, the measured values of the X coordinate by the measurement beams LWX1, LWX2, and LTX are XW1, XW2, and XTW, and the measured value of the Y coordinate by the measurement beams LWY1, LWY2, and LTY is Y.
Assuming that W1, YW2, and YTW, a rolling angle (hereinafter, referred to as a “tilt angle”) Tθx of the Zθ-axis driving stage 4A around the Y-axis at the time of scanning exposure and a pitching angle (tilt angle) around the X-axis. ) Tθy is as follows.

[0105]

Tθx = {(XW1 + XW2) / 2−XTW} / H Tθy = {(YW1 + YW2) / 2−YTW} / H Therefore, the Zθ axis driving stage 4 before correction during normal exposure
The X coordinate of A is (XW1 + XW2) / 2, and for this coordinate, for example, a coordinate obtained by adding the value of (Tθx · H / 2) is set as a new coordinate. Abbe error can be corrected by adding the value of (Tθy · H / 2).

In the embodiment shown in FIG. 8, the moving mirror 41X
Also serves as a reference mark plate, but instead
When the movable mirror 7X and the reference mark plate 6 are used in (b), almost the same effect can be obtained by directly joining the movable mirror 7X and the reference mark plate 6. It should be noted that the present invention is not limited to the above-described embodiment, and it is needless to say that various configurations can be adopted without departing from the gist of the present invention.

[0107]

According to the first exposure method of the present invention, the relative direction between the arrangement direction of a plurality of reference marks on the reference mark plate (direction of the reference mark plate) and the running direction of the substrate stage (wafer stage) is determined. Since the stepping direction of the substrate stage is determined according to the rotation angle θ1, even when the running direction of the substrate stage and the relative angle between the reference mark plate and the running direction of the substrate stage change, the shot arrangement on the substrate is changed to an orthogonal grid. There is an advantage that can be formed. Further, since the relative rotation error of the mask is determined according to the relative rotation angle θ2 between the scanning direction of the mask and the scanning direction of the substrate stage, the exposure on the substrate is performed by correcting the relative rotation error. The shot area can be reduced.

According to the second exposure method of the present invention,
Since the rotation angle of the mask is corrected in accordance with the relative rotation angle θ1 and the relative rotation angle θ2, even when the running direction of the substrate stage and the relative angle with respect to the reference mark plate change, There is an advantage that distortion in a shot region exposed above can be reduced. Further, according to the third exposure method of the present invention, in addition to the relative rotation angle θ1, the relative rotation angle θ3 of the running direction of the mask with respect to the arrangement direction of the measurement marks is detected, and based on the information of the relative rotation angle θ1. Since the position of the substrate stage at the time of scanning exposure is corrected and the position of the mask at the time of scanning exposure is corrected based on the information of the relative rotation angle θ3, the relative movement between the running direction of the substrate stage and the reference mark plate is determined. Even when the angle changes, the distortion of the shot area exposed on the substrate can be reduced.

According to the fourth exposure method of the present invention,
The relative rotation error between the arrangement direction of the measurement marks on the mask and the running direction of the mask can be measured via the first reference mark on the reference mark member (reference mark plate). Further, since the baseline amount of the alignment system is measured via the first and second reference marks on the reference mark member (reference mark plate), the running direction of the substrate stage and the relative position between the reference direction and the reference mark member are measured. Even when the angle changes, there is an advantage that the baseline amount can be measured with high accuracy.

Further, according to the fifth exposure method of the present invention,
Based on the amount of positional deviation obtained for each of the plurality of first and second reference marks on the reference mark member, a relative rotation error between the corresponding scanning direction of the mask and the substrate is corrected, and Since the relative rotation angle between the arrangement direction of the first and second reference marks on the reference mark member and the scanning direction of the substrate stage is corrected, when the inclination angle between the substrate stage and the reference mark member changes. However, there is an advantage that distortion in a shot area exposed on the substrate can be reduced.

Further, according to the sixth exposure method of the present invention,
First and second fiducial marks are formed on the upper surface of the movable mirror on the substrate stage side, and the movable mirror also serves as a fiducial mark plate. There is an advantage that the relative angle hardly changes, and as a result, the baseline amount of the alignment system can be measured with high accuracy.

[Brief description of the drawings]

FIG. 1 is a configuration diagram showing a projection exposure apparatus used in an example of an embodiment of the present invention.

2A is a plan view showing a reticle stage of the projection exposure apparatus of FIG. 1, and FIG. 2B is a plan view showing a wafer stage of the projection exposure apparatus of FIG.

FIG. 3 is a flowchart illustrating an example of a baseline measurement operation and an operation of measuring a tilt angle of a running direction of a wafer stage according to the embodiment of the present invention.

4A is a plan view showing an alignment mark of the reticle 12 of FIG. 1, and FIG. 4B is a plan view showing a relationship between an observation area of the reticle alignment microscopes 19 and 20 of FIG. FIG. 3C is an enlarged plan view showing the configuration of the fine alignment marks 29A to 30D.

FIG. 5 is an explanatory diagram of a positional relationship when the reference mark plate 6 and the reticle 12 are relatively scanned.

6A is a plan view showing a projected image of a reticle 12 on a wafer stage, FIG. 6B is an enlarged plan view showing a projected image of a fine alignment mark on the reticle 12, FIG.
(C) is a plan view showing an array of fiducial marks on the fiducial mark plate 6, and (d) is a fiducial mark 35A-35D, 36A-3.
6D is an enlarged plan view showing 6D, and (e) is reference marks 37A to 3A.
It is an enlarged plan view which shows 7D.

FIG. 7 is an explanatory diagram of a method of measuring an inclination angle Δθ between the movable mirror 7X on the wafer stage side and the reference mark plate 6.

8A is a plan view showing a main part of a wafer stage of a projection exposure apparatus used in another example of the embodiment of the present invention, and FIG. 8B is a side view of FIG. .

FIG. 9 is an explanatory diagram when distortion of a shot area or distortion of a shot arrangement occurs in the related art.

[Explanation of symbols]

4 Zθ axis drive stage 5 Wafer 6 Reference mark plate 7X, 7Y Moving mirror on wafer stage side 8 Projection optical system 11 Reticle minute drive stage 12 Reticle 13X1, 13X2, 13FX, 13Y1, 13Y2
Interferometer main body 14X1, 14X2, 14Y1, 14Y2 Interferometer main body 19, 20 Reticle alignment microscope (RA microscope) 21X Moving mirror on reticle stage side 21Y1, 21Y2 Corner cube 22A Main control system 22D Reticle driving device 29A to 29D, 30A to 30D Fine alignment mark 35A-35D, 36A-36D Reference mark for RA microscope 37A-37D Reference mark for alignment sensor 41X, 41Y Moving mirror

Claims (6)

[Claims] 1. A mask and a substrate are synchronously scanned in a corresponding scanning direction while a part of a pattern on a mask is projected on a photosensitive substrate on a substrate stage under exposure light. An exposure method for sequentially transferring a pattern on the mask to each shot area on the substrate, wherein a plurality of measurement marks are formed on the mask along the scanning direction. A reference mark member on which a plurality of reference marks having substantially the same positional relationship as the mark are formed is arranged on the substrate stage, and the substrate stage is moved in the scanning direction, so that the mask Of the plurality of measurement marks
Sequentially measuring the amount of displacement between each of the two measurement marks and the plurality of reference marks on the reference mark member,
From the measurement result, the arrangement direction of the plurality of reference marks,
A first step of detecting a relative rotation angle θ1 with respect to a running direction of the substrate stage; and moving the mask and the substrate in synchronization with the corresponding scanning direction, thereby obtaining the plurality of measurement marks on the mask. Are sequentially measured with the corresponding reference mark on the reference mark member, and a relative rotation angle θ2 between the scanning direction of the mask and the scanning direction of the substrate stage is detected from the measurement result. A step of determining the stepping direction of the substrate stage based on the information of the relative rotation angle θ1, and determining the scanning direction of the mask based on the information of the relative rotation angle θ2. Exposure method characterized by the above-mentioned. 2. The mask and the substrate are synchronously scanned in a corresponding scanning direction while a part of a pattern on the mask is projected on a photosensitive substrate on a substrate stage under exposure light. An exposure method for sequentially transferring a pattern on the mask to each shot area on the substrate, wherein a plurality of measurement marks are formed on the mask along the scanning direction. A reference mark member on which a plurality of reference marks having substantially the same positional relationship as the mark are formed is arranged on the substrate stage, and the substrate stage is moved in the scanning direction, so that the mask Of the plurality of measurement marks
Sequentially measuring the amount of displacement between each of the two measurement marks and the plurality of reference marks on the reference mark member,
From the measurement result, the arrangement direction of the plurality of reference marks,
A first step of detecting a relative rotation angle θ1 with respect to a running direction of the substrate stage; and moving the mask and the substrate in synchronization with the corresponding scanning direction, thereby obtaining the plurality of measurement marks on the mask. Are sequentially measured with the corresponding reference mark on the reference mark member, and a relative rotation angle θ2 between the scanning direction of the mask and the scanning direction of the substrate stage is detected from the measurement result. An exposure method, comprising: correcting a rotation angle of the mask based on a difference between the relative rotation angle θ1 and the relative rotation angle θ2. 3. The mask and the substrate are synchronously scanned in a corresponding scanning direction while a part of a pattern on the mask is projected on a photosensitive substrate on a substrate stage under exposure light. An exposure method for sequentially transferring a pattern on the mask to each shot area on the substrate, wherein a plurality of measurement marks are formed on the mask along the scanning direction. A reference mark member on which a plurality of reference marks having substantially the same positional relationship as the mark are formed is arranged on the substrate stage, and the substrate stage is moved in the scanning direction, so that the mask Of the plurality of measurement marks
Sequentially measuring the amount of displacement between each of the two measurement marks and the plurality of reference marks on the reference mark member,
From the measurement result, the arrangement direction of the plurality of reference marks,
A first step of detecting a relative rotation angle θ1 with respect to a running direction of the substrate stage; and scanning the mask in the scanning direction, thereby obtaining one of the plurality of reference marks on the reference mark member.
The amount of displacement between each of the reference marks and each of the plurality of measurement marks on the mask is sequentially measured, and from the measurement result, the alignment direction of the plurality of measurement marks and the running direction of the mask are determined. A second step of detecting a relative rotation angle θ3, wherein the position of the substrate stage during scanning exposure is corrected based on the information on the relative rotation angle θ1, and scanning is performed based on the information on the relative rotation angle θ3. An exposure method, wherein a position of the mask at the time of exposure is corrected. 4. The mask and the substrate are synchronized in a state in which a partial image of the pattern on the mask is projected on a photosensitive substrate on a substrate stage via a projection optical system under exposure light. An exposure method for sequentially transferring a pattern on the mask to each shot area on the substrate by scanning in a corresponding scanning direction, wherein a position of an alignment mark on the substrate is located near the projection optical system. An alignment system of an off-axis type for detection is arranged, a plurality of measurement marks are formed on the mask along the scanning direction, and a reference point in an exposure field of the projection optical system and the alignment system are formed. A reference mark member on which first and second reference marks are formed is arranged on the substrate stage at an interval corresponding to an interval with the reference point, and the reference system is provided by the alignment system. The on-click member second
While observing the reference mark, the mask is moved in the scanning direction, and the displacement between each of the plurality of measurement marks on the mask and the first reference mark on the reference mark member is changed. Are sequentially measured, the average value of the positional deviation amounts between each of the plurality of measurement marks and the first reference mark, and the relative rotation error of the mask with respect to the scanning direction obtained from the respective positional deviation amounts. And determining a distance between a reference point in an exposure field of the projection optical system and a reference point of the off-axis type alignment system from a positional shift amount of the second reference mark observed by the alignment system. Characteristic exposure method. 5. The mask and the substrate are synchronized in a state where an image of a part of the pattern on the mask is projected on a photosensitive substrate on a substrate stage via a projection optical system under exposure light. An exposure method for sequentially transferring a pattern on the mask to each shot area on the substrate by scanning in a corresponding scanning direction, wherein a position of an alignment mark on the substrate near the projection optical system An alignment system of an off-axis system for detecting a mark is formed, a plurality of measurement marks are formed on the mask along the scanning direction, and the plurality of measurement marks are made to correspond to the plurality of measurement marks on the mask. A plurality of first fiducial marks, and a distance between a fiducial point in the exposure field of the projection optical system and a fiducial point of the alignment system from each of the plurality of first fiducial marks. A reference mark member on which a plurality of second reference marks are formed at corresponding intervals is arranged on the substrate stage, and the mask and the substrate are moved in synchronization with the corresponding scanning direction, and the In parallel with measuring the amount of positional deviation between the plurality of measurement marks and the corresponding first reference mark on the reference mark member, the amount of positional deviation of the second reference mark is determined by the alignment system. The step of measuring is repeated for each of the plurality of first reference marks on the reference mark member, and the positional deviation obtained for each of the plurality of first and second reference marks Correcting the relative rotation error between the mask and the substrate in the corresponding scanning direction based on the amount, and adjusting the first and second positions on the reference mark member.
An exposure method, wherein the relative rotation angle between the reference mark arrangement direction and the substrate stage scanning direction is corrected. 6. The mask and the substrate are synchronized in a state where an image of a part of the pattern on the mask is projected onto a photosensitive substrate on a substrate stage via a projection optical system under exposure light. An exposure method for sequentially transferring a pattern on the mask to each shot area on the substrate by scanning in a corresponding scanning direction, wherein a position of an alignment mark on the substrate is located near the projection optical system. An alignment system of an off-axis system for detection is arranged, a moving mirror for measuring a coordinate position is fixed on the substrate stage, and a plurality of measurement marks are formed on the mask along the scanning direction. Forming a plurality of first fiducial marks on the upper surface of the movable mirror in correspondence with the plurality of measurement marks on the mask, and projecting the plurality of first fiducial marks from the plurality of first fiducial marks, respectively; Forming a plurality of second fiducial marks at intervals corresponding to the intervals between the fiducial points in the exposure field of the optical system and the fiducial points of the alignment system; and synchronizing the mask and the substrate in the corresponding scanning direction. Moving the plurality of measurement marks on the mask and the corresponding first reference mark on the movable mirror in parallel with the measurement of the second reference mark by the alignment system. The step of measuring the displacement amount of the reference mark is repeatedly performed on each of the plurality of first reference marks on the movable mirror, and the step of measuring the amount of displacement of the reference mark is performed on each of the plurality of first and second reference marks. A relative rotation error between the mask and the substrate in the corresponding scanning direction is corrected based on the amount of positional deviation obtained with respect to the reference point, and a reference point in an exposure field of the projection optical system is aligned with the alignment. An exposure method, comprising correcting an interval between a reference point of a measurement system and a reference point.