JP3460984B2 - Projection exposure method and exposure apparatus - Google Patents

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an exposure method and apparatus suitable for use in, for example, a photolithography process.

[0002]

2. Description of the Related Art A substrate on which a photosensitive material is applied with a pattern of a photomask or reticle (hereinafter referred to as "reticle") when manufacturing a semiconductor device, a liquid crystal display device, a thin film magnetic head, etc. by a photolithography process. A projection exposure apparatus for transferring onto a (wafer, glass plate, etc.) is used. A conventional projection exposure apparatus is a step-and-repeat type reduction projection type in which each shot area of a wafer is sequentially moved into the exposure field of a projection optical system to sequentially expose a pattern image of a reticle to each shot area. The exposure device (stepper) was often used.

FIG. 18 shows a main part of a conventional stepper,
In FIG. 18, the wafer 5 is placed on the wafer stage 4.
Is mounted, and the reference mark plate 57 is fixed on the wafer stage 4 near the wafer 5. Then, under exposure light from an illumination optical system (not shown), the reticle 1
The image of the pattern on 2 is projected and exposed on each shot area on the wafer 5 via the projection optical system 8. At this time, since the wafer stage 4 is driven along the wafer coordinate system, it is necessary to measure the position of the reticle 12 on the wafer coordinate system and the rotation angle of the reticle 12 with respect to the wafer coordinate system. Therefore, in the vicinity of the pattern area of the reticle 12, two alignment marks (reticle marks) 60R and 61R are formed so as to face each other, and the reference mark plate 5 is formed.
Two reference marks 60F and 61F are formed on the substrate 7 at intervals equal to the designed intervals of the reticle marks 60R and 61R on the wafer 5.

Also, the reticle mark 60 of the reticle 12
Reticle alignment microscopes 58 and 59 are arranged on R and 61R, respectively. The reticle alignment microscopes 58 and 59 respectively include an illumination light source that emits alignment light having the same wavelength as the exposure light and the reticle 12.
A sensor that can observe the upper reticle mark and the alignment mark (wafer mark) on the wafer 5 or the reference mark on the reference mark plate 57 at the same time is provided. FIG.
When the wafer 5 is exposed by the stepper, the image of the pattern on the reticle 12 is exposed in each shot area of the wafer 5 by sequentially moving only the wafer stage 4 by the step-and-repeat method. It

In such a stepper, the reticle 1 is further formed on the circuit pattern on the wafer 5 formed in the previous step.
In the case of exposing the second pattern image, the wafer coordinate system which defines the coordinates of each shot area on the wafer 5 and the reticle coordinate system which defines the coordinates of the pattern on the reticle 12 are made to correspond (that is, Need to be aligned). In the case of a stepper, the exposure field of the projection optical system 8 and the size of one shot area on the wafer 5 are equal,
Since it is not necessary to drive the reticle 12 when exposing,
The correspondence between the wafer coordinate system and the reticle coordinate system was taken as follows.

That is, after the wafer stage 4 is driven to move the reference mark plate 57 into the exposure field of the projection optical system 8, one of the reticle alignment microscopes 58 shifts the position of the reticle mark 60R from the reference mark 60F. Of the other reticle alignment microscope 59
The positional deviation amount between the reticle mark 61R and the reference mark 61F is detected by the method, and the position of the pattern of the reticle 12 on the wafer coordinate system is obtained from the positional deviation amount. Further, the reference mark 60F is moved to the position of the reference mark 61F, and the positional deviation amount between the reticle mark 61R and the reference mark 60F is detected by the reticle alignment microscope 59, whereby the rotation angle of the reticle 12 on the wafer coordinate system is detected. Was being measured. And reticle 12
Alternatively, the wafer coordinate system is finally associated with the reticle coordinate system by rotating the wafer stage 4 and correcting the rotation angle.

Further, in FIG. 18, in order to detect the position of each alignment mark (wafer mark) formed corresponding to each shot area on the wafer 5, the off-axis is provided on the side surface of the projection optical system 8. A system alignment microscope 34 is provided. In this case, based on the position of the wafer mark detected by the alignment microscope 34, the corresponding shot area on the wafer 5 is projected onto the projection optical system 8.
Exposure field. Therefore, a reference point (for example, the center of exposure) in the exposure field of the projection optical system 8 is previously set.
And the so-called baseline amount, which is the interval between the reference point 62 of the observation region of the off-axis type alignment microscope 34 and the reference point 62.

In the conventional stepper, when measuring such a baseline amount, the reticle mark 60 is used.
After measuring the amount of positional deviation between the R, 61R and the conjugate image of the reference marks 60F, 61F, the wafer stage 4 is moved by an amount equal to the design value of the baseline amount, and the reference point 62 is set by the alignment microscope 34. The amount of positional deviation from the corresponding reference mark on the reference mark plate 57 has been measured. The baseline amount has been obtained from the amount of displacement.

[0009]

In recent years, patterns are becoming finer in semiconductor elements and the like, so that it is required to increase the resolution of the projection optical system. Methods for increasing the resolution include shortening the wavelength of the exposure light or increasing the numerical aperture of the projection optical system.However, regardless of which method is used, the exposure field is similar to that of the conventional example. However, it is becoming difficult to maintain the imaging performance (distortion, curvature of field, etc.) at a predetermined accuracy over the entire exposure field. Therefore, what is currently being reviewed is a so-called slit scan exposure type projection exposure apparatus.

In this slit scan exposure type projection exposure apparatus, a reticle and a wafer are scanned relatively synchronously with respect to a rectangular or arcuate illumination area (hereinafter referred to as "slit illumination area"). Meanwhile, the pattern of the reticle is exposed on the wafer. Therefore, if a pattern having the same area as that of the stepper method is exposed on the wafer, the slit scan exposure method can make the exposure field of the projection optical system smaller than that of the stepper method. The accuracy of may improve.

The conventional mainstream size of a reticle is 6
Inch size, and the mainstream of projection magnification of projection optical system is 1
Although it was / 5 times, the size of the reticle under the magnification of ⅕ is not enough for the 6 inch size due to the increase in the area of the circuit pattern of the semiconductor element or the like. Therefore, it is necessary to design the projection exposure apparatus in which the projection magnification of the projection optical system is changed to, for example, 1/4. The slit scan exposure method is also advantageous in order to meet such a large area of the transferred pattern.

In such a slit scan exposure type projection exposure apparatus, when the method of associating the reticle coordinate system and the wafer coordinate system used in the conventional stepper is applied, the projection magnification becomes 1/4. Therefore, there is an inconvenience that the alignment accuracy is deteriorated by the drawing error of the circuit pattern on the reticle. Furthermore, Japanese Patent Application No. 3-1
No. 69781 proposes a technique for measuring the rotation angle of a reticle by simultaneously measuring the positional deviation amounts of a plurality of measurement marks without moving the wafer stage in a stepper. However, the concept of measuring the rotation angle by simultaneously measuring a plurality of measurement marks cannot be used in the scanning direction of the slit scan exposure type projection exposure apparatus, and the rotation angle between the reticle coordinate system and the wafer coordinate system and those There is an inconvenience that the coordinate orthogonality of the coordinate system cannot be measured accurately.

Regarding the method of measuring the baseline amount, which is the distance between the reference position in the exposure field of the projection optical system and the reference position of the off-axis alignment system, one position on the reticle of a conventional stepper ( Two
If the measurement method using () marks is directly applied to the projection exposure apparatus of the slit scan exposure method, there is a disadvantage that it is greatly affected by the reticle drawing error.

SUMMARY OF THE INVENTION In view of the above problems, an object of the present invention is to provide an exposure method and apparatus capable of accurately projecting and exposing a mask pattern image on a substrate. In addition, the present invention is a reticle
Clu coordinate system (mask coordinate system) and wafer coordinate system (substrate coordinate)
System), and high baseline measurement
It is an object of the present invention to provide an exposure method and apparatus that can be used
To do.

[0015]

In order to solve the above-mentioned object, the projection exposure method according to the present invention is held on a first stage.
Synchronized mask and substrate held on the second stage
The mask pattern by moving
In a projection exposure method for exposing the substrate had its dew
Prior to the light, a first mode in which a plurality of marks formed on the mask are detected by a detection system, and a number of marks formed on the mask that is less than the number of marks detected in the first mode are detected. Either one of the second mode to be detected by the detection system is selected, and in this selected mode
Using the detection system, the mark on the mask and the second scan
The mark of the tage is detected, and based on this detection result,
A first coordinate system and its basis for controlling the movement of the mask
The correspondence with the second coordinate system for controlling the movement of the plate, or
Project the image of the mask pattern onto the substrate.
Detects the projection reference point of the shadow system and the alignment information of the substrate
Off-axis alignment system detection reference point
It seeks the positional relationship with .

In this case, as an example, the second mode
Is a part of the marks detected in the first mode.
This is the detection mode.

Next, the exposure apparatus of the present invention comprises a first stage
The mask held on the substrate and the substrate held on the second stage
By moving in synchronization with an exposure apparatus that exposes the substrate using the pattern of the mask, and detectable mark detection system marks the mask held by the first stage, the mark detection system Execute a first mode to detect multiple marks on the mask by
A control system for selecting whether to execute the second mode for detecting the marks on the mask, the number of which is smaller than the number of marks detected by the mark detection system in the first mode, and prior to the exposure. The control system is
Select either the first mode or the second mode
And using its detection system in this selected mode
Detect the mark on the mask and the mark on the second stage
Control the movement of the mask based on this detection result.
To control the movement of the first coordinate system and its substrate for
Relationship with the second coordinate system of, or the pattern of the mask
And the projection reference point of the projection system that projects the image of
Off-axis method to detect board alignment information
The positional relationship with the detection reference point of the alignment system is obtained.

[0018]

[0019]

According to the present invention, the reticle coordinate system (mass
Correspondence between the wafer coordinate system) and the wafer coordinate system (substrate coordinate system)
And perform baseline measurement with high throughput.
The pattern image of the mask can be accurately projected and exposed on the substrate.

[0020]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS A first embodiment of the present invention will be described below with reference to the drawings. The present embodiment is an application of the present invention when a reticle pattern is exposed on a wafer by a slit scan exposure type projection exposure apparatus. FIG. 1 shows a projection exposure apparatus of this embodiment. In FIG. 1, a reticle 12 is formed by a rectangular illumination area (hereinafter, referred to as “slit-shaped illumination area”) by exposure light EL from an illumination optical system (not shown). The upper pattern is illuminated, and the image of the pattern is projected and exposed on the wafer 5 via the projection optical system 8. At this time, in synchronization with the scanning of the reticle 12 in the forward direction with respect to the paper surface of FIG. On the other hand, scanning is performed backward at a constant speed V / M (1 / M is a reduction magnification of the projection optical system 8).

A drive system for the reticle 12 and the wafer 5 will be described. A reticle Y drive stage 10 that can be driven in the Y-axis direction (direction perpendicular to the paper surface of FIG. 1) is placed on the reticle support base 9. A reticle micro-driving stage 11 is placed on the reticle Y driving stage 10, and a reticle 12 is held on the reticle micro-driving stage 11 by a vacuum chuck or the like. Reticle micro drive stage 1
Reference numeral 1 indicates the position control of the reticle 12 with a small amount and with high accuracy in the X direction, the Y direction, and the rotation direction (θ direction) parallel to the paper surface of FIG. 1 in the plane perpendicular to the optical axis of the projection optical system 8. To do. A movable mirror 21 is arranged on the reticle micro-driving stage 11, and an interferometer 14 arranged on the reticle support 9 is provided.
The position of the reticle micro-driving stage 11 in the X direction, Y direction, and θ direction is constantly monitored by. The position information S1 obtained by the interferometer 14 is the main control system 22A.
Is being supplied to.

On the other hand, a wafer Y-axis drive stage 2 that can be driven in the Y-axis direction is mounted on the wafer support base 1, and a wafer X-axis drive stage 3 that can be driven in the X-axis direction is mounted thereon. And a Zθ axis drive stage 4 is provided on it.
The wafer 5 is held on the Zθ axis drive stage 4 by vacuum suction. The movable mirror 7 is fixed also on the Zθ axis drive stage 4, and the interferometer 13 arranged outside
The positions of the Zθ axis drive stage 4 in the X direction, the Y direction, and the θ direction are monitored, and the position information obtained by the interferometer 13 is also supplied to the main control system 22A. Main control system 22A
Controls the positioning operation of the wafer Y-axis drive stage 2, the wafer X-axis drive stage 3, and the Zθ-axis drive stage 4 via the wafer drive device 22B and the like, and also controls the operation of the entire device.

As will be described later, the interferometer 13 on the wafer side is also provided.
In order to make a correspondence between the wafer coordinate system defined by the coordinates measured by the reticle and the reticle coordinate system defined by the coordinates measured by the reticle-side interferometer 14,
A reference mark plate 6 is fixed near the wafer 5 on the θ-axis drive stage 4. Various reference marks are formed on the reference mark plate 6 as described later. Among these reference marks, there are reference marks illuminated from the back side by the illumination light guided to the Zθ axis drive stage 4, that is, luminescent reference marks.

Above the reticle 12 of this example, reticle alignment microscopes 19 and 20 for simultaneously observing the reference mark on the reference mark plate 6 and the mark on the reticle 12 are provided. In this case, the deflection mirrors 15 and 16 for guiding the detection light from the reticle 12 to the reticle alignment microscopes 19 and 20, respectively, are movably arranged, and when the exposure sequence is started, a command from the main control system 22A is also issued. And the mirror drive devices 17 and 1
The deflecting mirrors 15 and 16 are retracted by 8, respectively. Further, the wafer 5 is attached to the side surface of the projection optical system 8 in the Y direction.
An off-axis type alignment device 34 for observing the upper alignment mark (wafer mark) is arranged.

A keyboard 22C for inputting commands from an operator is connected to the main control system 22A. The projection exposure apparatus of the present embodiment has a quick mode for simply measuring the baseline amount and the like as described later, in addition to the mode for highly accurate measurement, and the operator uses the keyboard 22C. Main control system 22A
Instruct whether the mode to be executed from now on is the high precision mode or the quick mode.

Next, the sequence from the loading of the wafer 5 and the reticle 12 to the completion of the alignment in the projection exposure apparatus of this embodiment will be described with reference to the flowchart of FIG. First, in step 101 of FIG. 2, pre-alignment of the reticle 12 is performed on the reticle loader (described later) on the basis of the outer shape. 3 shows a reticle loader system for transporting the reticle 12 onto the reticle micro-driving stage 11 shown in FIG. 1. The reticle loader shown in FIG. 3 has two reticle arms 23A and 23B and these reticle arms 23A, 23A. 23B, an arm rotating shaft 25 and a rotating mechanism 26 for rotating the arm rotating shaft 25. Reticle arm 23A
Grooves 24A and 24B for vacuum suction are formed on the reticle mounting surfaces of the reticle and 23B, respectively, and the reticle arms 23A and 23B are supported via an arm rotation shaft 25 so as to be independently rotatable.

When the reticle 12 is loaded, the reticle 12 is transferred onto the reticle arm 23A by another reticle transport mechanism (not shown) at the position A3. At this time, the other reticle arm 23B is used, for example, for carrying out the reticle used in the previous step. Next, the reticle 12 is placed on the reticle arm 23A by a reticle outline pre-alignment mechanism (not shown) installed near the position A3.
Are aligned with a certain accuracy on the basis of the outer shape, the reticle 12 is vacuum-sucked on the reticle arm 23A.
Next, in step 102 of FIG. 2, the rotation mechanism 26 rotates the reticle arm 23A via the arm rotation shaft 25, and the position B3 in the Y direction (standby position (transfer position) of the reticle drive stage 10 in FIG. 1) is reached. Up to reticle 12
To move.

At this time, since the vacuum suction groove 24A is in the direction orthogonal to the suction position on the reticle micro-driving stage 11 and outside the pattern area of the reticle 12, the reticle micro-driving stage 11 is in the scanning direction. Is y
Reticle arm 23A while moving to the forefront of the direction
The reticle 12 can be freely moved in and out of the reticle micro-driving stage 11. When the reticle 12 reaches the reticle micro-driving stage 11 (see FIG. 1), the arm rotation shaft 25 is lowered in the −Z direction, the reticle 12 is placed on the vacuum suction surface of the reticle micro-driving stage 11, and the reticle 12 moves. After the delivery is completed, the reticle arm 23A retracts. Then, the reticle micro-driving stage 11 carries the reticle 12 in the direction of the position C3. At this time, the reticle arms 23A and 23B are driven independently of each other, and, for example, the reticle loading and the reticle unloading are simultaneously performed, thereby improving the reticle exchange speed.

Next, the alignment of the reticle 12 will be performed in step 103 and subsequent steps of FIG. 2, and the mechanism and operation therefor will be described. FIG. 4A shows the arrangement of alignment marks (reticle marks) on the reticle 12, and FIG.
(B) is a slit-shaped illumination area bright 3 in the area 33R conjugate with the effective exposure field of the projection optical system on the reticle.
2 etc. are shown. The scanning direction is the y direction, and the direction perpendicular to the y direction is the x direction. In FIG. 4A, the reticle 1
A light-shielding portion 31 is formed around the central pattern area on the upper portion 2 of the reticle mark, and the reticle marks formed on the outer side of the light-shielding portion 31 are rough search alignment marks 27 and 28 and fine alignment marks 29A to 29D. 30A to 30D. The rough search alignment mark 27 on the right side is formed by a linear pattern that is long along the y direction, which is the scanning direction, and a cross pattern formed on both ends of this linear pattern, and the rough search alignment mark 27 on the left side is formed. Reference numeral 28 is symmetrical with the rough search alignment mark 27 on the right side.

Further, between the light-shielding portion 31 on the right side and one of the cross patterns of the rough search alignment mark 27, the fine alignment mark 29 is located close to the y direction.
A and 29B are formed between the light-shielding portion 31 on the right side and the other cross pattern of the rough search alignment mark 27, and the fine alignment mark 2 is formed in close proximity to the y direction.
9C and 29D are formed. Fine alignment marks 30A to 30D are formed on the left side symmetrically with these fine alignment marks 29A to 29D.
4A, only 30A to 30D are shown as cross-shaped marks, but actually, as shown in FIG. 4C, two sets of three linear patterns are arranged in the x direction at predetermined intervals. In addition to being arranged, two sets of three linear patterns are arranged in the y direction at predetermined intervals.

First, in step 103 of FIG.
Rough search alignment mark 28 on the left side of (a)
Are detected by the reticle alignment microscope (hereinafter referred to as “RA microscope”) 20 of FIG. FIG. 4B shows the observation areas 19R and 20R on the reticle 12 of the RA microscopes 19 and 20 in this case, and when the rough search is performed, the rough search alignment marks 27 and 28 are
Respectively outside the observation regions 19R and 20R,
Further, it is outside the region 33R which is conjugate with the effective exposure field. This is because it is necessary to make the rough search alignment marks 27 and 28 large for the rough search, but if the exposure field of the projection optical system is made large accordingly, the diameter of the projection lens needs to be made large, resulting in an increase in cost. Because of that. Therefore, the procedure for performing the rough search in this example will be described with reference to FIG.

FIG. 5A is an enlarged view in the vicinity of one cross pattern of the rough search alignment mark 28, and FIG.
FIG. 5B is a reduced view of FIG.
In (a) and (b), let W be the width of the square effective field of view 20R _ef of the RA microscope 20, and let ΔR be the design value of the sum of the pattern drawing error and the installation error with respect to the outer shape of the reticle 12. Therefore, as shown in FIG.
One cross pattern 28a of the rough search alignment mark 28 is always included in the R square area.
The detection target is the x-coordinate and the y-coordinate of the cross pattern 28a, but in this example, the width W of the alignment mark 28 intersects the two axes at 45 °, that is, diagonally to the x-axis and the y-axis. Scan the effective field of view 20R _ef of. Then, the x coordinate and the y coordinate of the cross pattern 28a are obtained as the x coordinate and the y coordinate when the alignment mark 28 is crossed during the oblique scanning.

For that purpose, the integer part of the positive real number a is set to IN.
As represented by T (a), the minimum number of search screens for scanning a square area of the width ΔR with the effective field of view 20R _ef of the width W is {INT (ΔR / W) +1}. The number of search screens is obtained in advance. Then, {INT (ΔR / W) +1} each having a width W is diagonally arranged in a square region having the width ΔR centered on the first effective visual field B5.
Number of effective visual field A5, B5, C5, and set the ‥‥, by driving the reticle fine driving stage 11 in Figure 1, the effective field of view 20R _ef sequential FIGS. 5 (a) of each active field by stepping
The image within each effective field of view is sampled while setting inside.

As shown in FIG. 5B, at least the width Δ
The cross pattern 28a of the alignment mark 28 to be searched exists in the search range of R × ΔR, and the alignment mark 28 is sufficiently large with respect to the search range. Therefore, it is understood that the coordinates of the cross pattern 28a of the alignment mark 28 can be detected with the minimum number of screens by stepwise feeding the effective field of view to the alignment mark 28. The image processing at that time may be one-dimensional image processing for an image signal obtained by adding the scanning lines of all the captured lines in the screen.

FIG. 6 shows various image signals obtained by adding all the scanning lines, and FIGS. 6 (a) and 6 (d) are obtained in the effective visual field A5 of FIG. 5 (b). Image signals along the x-direction and the y-direction, FIGS. 6 (b) and 6 (e) are shown in FIG.
Image signals along the x and y directions obtained in the effective visual field B5 of (b), and FIGS. 6 (c) and 6 (f) are images along the x and y directions obtained in the effective visual field C5 of FIG. 5 (b). It is a signal. The x coordinate of the cross pattern 28a is obtained from the image signal of FIG. 6 (b), and the y coordinate of the cross pattern 28a is obtained from the image signal of FIG. 6 (f).

In this way, the search reticle mark 28 is searched.
2 is detected, the alignment mark 27 for rough search is moved to the observation region of the RA microscope 19 this time, and the position of the alignment mark 27 is detected in the same manner. However, in this case, the pattern-free portion of the reference mark plate 6 in FIG. 1 is moved into the exposure field of the projection optical system 8 and the pattern-less portion is illuminated from the bottom. Thus, the rough search alignment marks 27 and 28 are illuminated from the back surface side by the illumination light emitted from the reference mark plate 6.

With the above sequence, it is possible to roughly associate the positions of the rough search alignment marks 27 and 28 and the reticle coordinate system with the observation regions 19R and 20R of the RA microscopes 19 and 20 of FIG. 4B. Further, the rough correspondence between the observation regions 19R and 20R of the RA microscope and the wafer coordinate system can be performed by measuring the reference marks on the reference mark plate 6 in FIG. 1 with the RA microscopes 19 and 20. Thereby, the fine alignment marks 29A to 29D and 30A to 30
The rough alignment (rough alignment) is completed to the extent that D and the reference mark (described later) on the reference mark plate 6 do not overlap.

However, in this example, in order to reduce the lens diameter of the projection optical system 8, the alignment marks on the reticle 12 are divided into rough search alignment marks and fine alignment marks. When the diameter can be increased, the rough search alignment mark and the fine alignment mark can be used as a common mark. Even in this case, as shown in FIG. 5, the method of searching the alignment mark by stepwise feeding in the oblique direction can be diverted and the RA microscopes 19 and 20 can simultaneously search the alignment mark.

Next, the fine alignment sequence will be described, but before that, the detailed configurations of the wafer stage and the reticle stage will be described. Figure 7 (a)
7A is a plan view of the wafer stage. In FIG. 7A, the wafer 5 and the reference mark plate 6 are arranged on the Zθ axis drive stage 4. Further, an X-axis moving mirror 7X and a Y-axis moving mirror 7Y are fixed on the Zθ-axis drive stage 4, and a slit-shaped corresponding to the slit-shaped illumination area 32 of FIG. The illumination area 32W is illuminated with the exposure light, and the observation areas 19W and 20W are respectively shown in FIG.
It is conjugated with the observation regions 19R and 20R of (b).

The movable mirror 7X is irradiated with the laser beams LWX and LW _{of the} intervals IL along the optical paths which are parallel to the X axis and pass through the optical axis of the projection optical system and the reference point of the alignment device 34, respectively. 7Y is irradiated with two laser beams LWY1 and LWY2 having an interval IL along an optical path parallel to the Y axis. At the time of exposure, the coordinate value measured by the interferometer using the laser beam LWX is used as the X coordinate of the Zθ axis drive stage 4, and the coordinate value Y measured by the interferometer using the laser beams LWY1 and LWY2 is used as the Y coordinate. Average of ₁ and Y ₂ (Y
₁ + Y ₂ ) / 2 is used. Further, for example, from the difference between the coordinate values Y ₁ and Y ₂ , the rotation direction of the Zθ-axis drive stage 4 (θ
Direction) is measured. The position and rotation angle of the XY plane of the Zθ axis drive stage 4 are controlled based on those coordinates.

In particular, in the Y direction, which is the scanning direction, the average value of the measurement results of the two interferometers is used to reduce the error due to air fluctuation during scanning due to the averaging effect. The position in the X-axis direction when the off-axis type alignment device 34 is used is controlled based on the measurement value of a dedicated interferometer using the laser beam LW _of so that so-called Abbe error does not occur. It is a composition.

FIG. 7B is a plan view of the reticle stage. In FIG. 7B, the reticle micro-driving stage 11 is placed on the reticle Y driving stage 10 and the reticle 12 is held thereon. Has been done. Further, the reticle micro-driving stage 11 has a moving mirror 21x for the x-axis.
And two moving mirrors 21y1 and 21y2 for the y-axis are fixed, and the laser beam LR is parallel to the x-axis on the moving mirror 21x.
The laser beams LRy1 and LRy2 are irradiated on the movable mirrors 21y1 and 21y2 in parallel with the y-axis.

Similar to the wafer stage, the y-direction coordinate of the reticle micro-driving stage 11 is the laser beam LR.
The average value (y ₁ + y ₂ ) / 2 of the coordinate values y ₁ and y ₂ measured by two interferometers using y1 and LRy2 is used. Moreover, the coordinate value measured by the interferometer using the laser beam LRx is used as the coordinate in the x direction. Further, for example, the rotation amount of the reticle micro-driving stage 11 in the rotation direction (θ direction) is measured from the difference between the coordinate values y ₁ and y ₂ .

In this case, corner cube type reflecting elements are used as the movable mirrors 21y1 and 21y2 in the y direction which is the scanning direction, and the laser beams LRy1 and LRy2 reflected by the movable mirrors 21y1 and 21y2 are respectively reflected. It is reflected back at 39 and 38. That is,
The reticle interferometer is a double-pass interferometer, which is configured so that the laser beam position is not displaced by the rotation of the reticle micro-driving stage 11. Further, as in the case of the wafer stage, the slit-shaped illumination area 32 and the RA microscopes 19, 20 are formed on the reticle 12.
The observation areas 19R and 20R are arranged. And
The reticle 12 and the Zθ axis drive stage 4 of FIG. 7A can be observed only from the observation regions 19R and 20R. In this way, the reticle 12 and the Zθ axis drive stage 4 are
Is to improve the alignment accuracy at the time of exposure and the rotation accuracy of the reticle 12 and the wafer 5 by measuring the relationship with the above. A method thereof will be described with reference to FIGS. 8 and 9.

FIG. 8A shows the reticle 12 of FIG. 4A.
7A shows a reticle image 12W obtained by projecting it onto the reference mark plate 6 in FIG. 7A, and in FIG.
(A) Mark images 29AW to 29DW conjugated to the fine alignment marks 29A to 29D and mark images 30AW conjugated to the fine alignment marks 30A to 30D.
˜30 DW is shown. Each mark image 29AW-2
9DW and 30AW to 30DW are shown in FIG.
As shown in (3), three linear patterns are arranged on four sides.

FIG. 8C shows the arrangement of the reference marks on the reference mark plate 6. On the reference mark plate 6 of FIG. 8C, the mark images 29AW to 29DW of FIG. Three
The reference marks 35A to 35D and 36A to 36D are formed in substantially the same arrangement as 0AW to 30DW. These reference marks are illuminated from the back surface of the reference mark plate 6 with illumination light having the same wavelength as the exposure light. A reference mark 37A is formed on the reference mark plate 6 at a position separated from the midpoint of the reference marks 35A and 36A by a distance IL in the Y direction which is the scanning direction. The interval IL is shown in FIG.
It is equal to the baseline amount, which is the distance between the reference point of the projection optical system 8 and the reference point of the off-axis type alignment device 34. Similarly, fiducial marks 35B and 36B
The reference marks 37B, 37C and 37D are formed at positions separated from the center point, the center point of the reference marks 35C and 36C and the center point of the reference marks 35D and 36D by the distance IL in the Y direction.

Reference marks 35A to 35D, 36A to 36
As shown in FIG. 8D, each D is composed of a linear pattern of 7 rows × 7 columns, and these reference marks 35 are formed.
A to 35D and 36A to 36D are the mark images 2 in FIG.
It is a size that can be accommodated within 9 AW to 30 DW. Also,
As shown in FIG. 8E, the reference marks 37A to 37D use corresponding grid points in the grid pattern formed at a predetermined pitch in the X and Y directions.

In this case, first, in step 105 of FIG. 2, the relative positional relationship and the relative rotation angle between the reticle 12 and the RA microscopes 19 and 20 are calculated from the results obtained by the measurement in steps 103 and 104. And then Figure 4
(A) Fine alignment marks 29A and 30A
Are moved into the observation regions 19R and 20R of the RA microscopes 19 and 20, respectively. Then, in step 106, the fiducial mark 35 on the fiducial mark plate 6 of FIG.
A and 36A are the observation areas 19R and 20R, respectively.
It moves to the observation regions 19W and 20W (see FIG. 9) which are conjugate with. As a result, as shown in FIG. 9A, the mark image 29AW and the reference mark 35A can be simultaneously observed in the observation area 19W, and the mark image 30AW and the reference mark 36A can be simultaneously observed in the observation area 20W. After that, Figure 2
In step 107, the images observed by the RA microscopes 19 and 20 are converted into imaging signals and sampled, and at the same time, the off-axis alignment device 34 is used.
However, the detection signal of the corresponding reference mark image is sampled.

In FIG. 9A, a reticle image 12W, which is a projected image of the reticle 12, is projected on the reference mark plate 6. Further, as shown in FIG. 9C, the observation areas 19W and 20W are located at positions crossing the optical axis in the exposure field of the projection optical system 8 and are in the observation area of the off-axis alignment device 34. Fiducial mark 3
7A is settled. Then, similarly to the slit scan exposure, the reticle micro-driving stage 11 of FIG. 7B is moved downward in synchronization with the movement of the Zθ-axis driving stage 4 of FIG. 7A to the upper side (Y direction). When moved in the (-y direction), the fiducial mark plate 6 and the reticle image 12W move together in the Y direction as shown in FIGS. 9 (a) to 9 (b).
At this time, the observation areas 19W, 2 of the RA microscopes 19, 20
Since 0 W and the off-axis alignment device 34 are fixed, a group of marks (mark images 29AW, 30AW, reference marks 35A, 35A Three
6A, 37A), a group of marks (mark images 29DW, 30DW, reference marks 35D, 36D, 3) to which a symbol D is attached.
Up to 7D) moves.

First, at the first stationary position in FIG. 9A after the start of alignment, the mark image 2 is formed below the observation region 19W.
9AW and the reference mark 35A, there is a mark image 30AW and the reference mark 36A under the observation region 20W,
Below the off-axis type alignment device 34, there is a reference mark 37A, and all the marks marked with the symbol A are observed at the same time. When the measurement at the first stationary position is completed, the reticle image 12W and the reference mark plate 6 are moved in synchronization to the second stationary position by the stepping operation. The mark group existing below the observation regions 19W and 20W and the alignment device 34 at the first stationary position is a mark group denoted by the reference symbol A, and the mark groups of the observation regions 19W and 20W and the alignment device 34 at the second stationary position. The mark group existing below is a mark group denoted by reference numeral B (mark image 29BP in FIG. 8, reference marks 35B, 37B, etc.).

By repeating the above sequence with the third stationary position and the fourth stationary position (state of FIG. 9B), the mark image of the reticle image 12W and the reference mark on the reference mark plate 6 are formed. , The mark group with the reference symbol A, the mark group with the reference symbol B, the mark group with the reference symbol C, and the mark group with the reference symbol D in this order
9, 20 and off-axis alignment device 3
It will be measured by 4. This operation is the operation of steps 105 to 110 in FIG. The measurement results are shown in FIG. 10 in order to express the measurement results thus obtained in an easy-to-understand manner.

In FIG. 10, the vector of the alignment error from the reference mark 35A to the mark image 29AW, which is obtained by correcting the measurement result obtained by the RA microscope 19 as described later, is AL, and similarly the reference marks 35B to
Vectors of alignment errors from 35D to mark images 29BW to 29DW are BL to DL, respectively. Similarly, the vector of the alignment error from the reference mark 36A to the mark image 30AW is set as AR, and the reference mark 36
Vectors of alignment errors from B to 36D to the mark images 30BW to 30DW are BR to DR, respectively.
In addition, the alignment device 34 from the reference marks 37A to 37D, which is obtained by correcting the measurement result obtained by the off-axis type alignment device 34 as described below.
The error vectors up to the reference point of are respectively AO to DO.

Then, the error vectors AL, AR to DL,
When DR is obtained, the coordinate values in the x direction measured by the reticle-side interferometer 14 in FIG. 1, that is, the coordinate values obtained by using the laser beam LRx in FIG. 7B are respectively ReAx.
~ ReDx and error vectors AL, AR ~ DL, DR, the coordinate values in the y direction measured by the reticle-side interferometer 14 in Fig. 1, that is, the laser beam LR in Fig. 7B.
The coordinate values obtained using y1 and LRy2 are respectively Re
Ay1 to ReDy1 and ReAy2 to ReAy2.
Further, when the error vectors AL, AR to DL, and DR are obtained, the coordinate values in the X direction measured by the interferometer 13 on the wafer side in FIG. 1, that is, the laser beam LWX in FIG. When the obtained coordinate values are WaAX to WaDX and the error vectors AL, AR to DL, DR are obtained, the coordinate values in the Y direction measured by the interferometer 13 on the wafer side in FIG. 1, that is, FIG. 7A. Coordinate values obtained using the laser beams LWY1 and LWY2 of WaAY1 to WaDY, respectively.
1, WaAY2 to WaDY2.

Further, when the error vectors AO to DO are obtained, the coordinate values in the X direction obtained by the interferometer dedicated to the off-axis type alignment apparatus, that is, the laser beam LW _OF of FIG. 7A is used. The coordinate values obtained by
AOX to WaDOX. In this case, as shown in FIG. 7A, the laser beams LWY1 and LWY on the wafer side are formed.
2 is IL in the X direction, and RLy1 and LRy2 are the laser beams on the wafer side.

Next, the configuration of the RA microscope 19 shown in FIG. 1 will be described in detail in order to explain how to obtain the error vector AL and the like shown in FIG. FIG. 11 shows the RA microscope 19 and this illumination system. In FIG. 11, the illumination light EL having the same wavelength as the exposure light enters the Zθ-axis driving stage 4 from the outside of the Zθ-axis driving stage 4 via the optical fiber 44. Have been guided. The exposure light may be relayed by a lens system instead of the optical fiber 44. The illumination light guided in that way
Lens 45A, beam splitter 45B and lens 4
5C, the reference marks 35A to 35 on the reference mark plate 6
The illumination light that illuminates D and transmitted through the beam splitter 45B passes through the lens 45D, the lens 45E, the mirror 45F, and the lens 45G, and then the reference mark 36 on the reference mark plate 6.
Illuminating A to 36D.

For example, the light transmitted through the reference mark 35A is
An image of the reference mark 35A is formed on the fine alignment mark 29 on the reticle 12 via the projection optical system 8. The image of the reference mark 35A and the light from the alignment mark 29 are reflected by the deflection mirror 15, the lens 40A,
The light that has reached the half mirror 42 through the lens 40B and is split into two by the half mirror 42 is composed of a two-dimensional CCD.
It is incident on the image pickup surface of Y. These image pickup devices 43X and 43
Images of the fine alignment mark 29A and the reference mark 35 image 35AR as shown in FIG. 12A are projected on the Y imaging surface. In this case, the image pickup screen 43Xa of the X-axis image pickup element 43X is the X-axis on the wafer stage.
Although it is a region parallel to the direction and the direction of the horizontal scanning line is also the X direction, the image pickup screen 43Ya of the image pickup element 43Y for the Y axis is used.
Is a region parallel to the Y direction on the wafer stage, and the horizontal scanning line is also in the Y direction.

Therefore, the image pickup signal S4X of the image pickup device 43X
The positional deviation amount of the reference mark 35A and the alignment mark 29A in the X direction is obtained from the arithmetic mean of
The reference mark 35A from the arithmetic mean of the 3Y image pickup signal S4Y
The amount of positional deviation between the alignment mark 29A and the alignment mark 29A in the Y direction is obtained. The image pickup signals S4X and S4Y are supplied to the signal processing device 41.

More specifically, the case where the mark group with the symbol A is aligned will be described as an example.
The A microscope 19 simultaneously observes the alignment mark 29A and the reference mark image 35AR shown in FIG. 12A, for example. In FIG. 12A, the image signals S4X in the image pickup screens 43Xa and 43Ya surrounded by broken lines.
And S4Y are detected as digital signals by analog / digital conversion in the signal processing device 41. The image data on each scanning line is transferred to the X-axis in the signal processing device 41.
The image signal S4X 'for the X-axis and the image signal S4Y' for the Y-axis, which have been added and averaged independently on the axes and the Y-axis, are as shown in FIGS. 12B and 12C, respectively. Each of these image data is processed as a one-dimensional image processing signal.

When the signal thus obtained is arithmetically processed by the signal processing device 41, the X of the mark image 29AW of the reticle 12 and the reference mark 35A of the reference mark plate 6 of FIG.
Misalignment in the Y and Y directions AL ′ _X and AL ′
_Y is required. Then, the RA microscope 20 of FIG. 1 obtains the relative positional deviations AR ′ _X and AR ′ _Y between the mark image 30AW and the reference mark 36A in the X and Y directions. Similarly, the mark images 29BW to 29D of FIG.
Relative displacement between W and the reference marks 35B to 35D,
And mark images 30BW to 30DW and reference mark 36B to
The relative positional deviation from 36D is obtained.

However, for example, the image signal corresponding to the alignment mark 29A in FIG.
The position of the image signal corresponding to 5AR is controlled by the reticle side interferometer and the wafer side interferometer, respectively. Therefore, for example, a group of marks with the symbol A (see FIG.
0, 29AW, 35A, 30AW, 36A), measurement coordinates ReAx, Re of the reticle side interferometer when measuring
Ay1, ReAy2 and measurement coordinates W of the interferometer on the wafer side
For aAX, WaAY1, and WaAY2, a measurement error caused by a tracking error of each stage (= measured value-set value)
ΔReAx, ΔReAy1, ΔReAy2, and Δ
WaAX, ΔWaAY1, and ΔWaAY2 occur. This measurement error is included in the relative positional deviations AL ′ _X and AL ′ _Y obtained by the calculation.

[0061] Therefore, as in the following equation, the result obtained by subtracting those errors from relative positional deviation obtained by the measurement, the X component AL _X and Y components AL _Y vectors AL alignment error of 10 . However, in the following equation, (1 / M) is the reduction magnification of the projection optical system 8, and IL and RL are the intervals described in FIG. 7, respectively.

[0062]

## EQU1 ## AL _X = AL ' _X -ΔReAx / M-ΔWaAX

[0063]

## EQU2 ## AL _Y = AL ' _Y -ΔReAy1 / M-{(ΔW
aAY1 + ΔWaAY2) / 2- (ΔWaAY2-ΔW
aAY1) .RL / IL} Similarly, the alignment error vector AR of FIG.
The X component AR _X and the Y component AR _{Y of} are also obtained from the following equations.

[0064]

## EQU3 ## AR _X = AR ' _X -ΔReAx / M-ΔWaAX

[0065]

## EQU4 ## AR _Y = AR ' _Y -ΔReAy2 / M-{(ΔW
aAY1 + ΔWaAY2) / 2- (ΔWaAY2-ΔW
aAY1) × RL / IL} Next, the error vectors AO to DO of FIG. 10 obtained by correcting the results obtained by the off-axis type alignment device 34 will be described. For that purpose, the configuration of the alignment device 34 will be described. This will be described with reference to FIG.

FIG. 13 shows the configuration of the alignment device 34. In FIG. 13, the light from the reference mark on the reference mark plate 6 is deflected by the deflecting mirror section 46 and enters the half prism 47, and the half prism 47 is scanned. The light reflected by the prism 47 is directed to an image processing type alignment optical system (hereinafter referred to as “FIA optical system”) 48 using white light,
The light transmitted through the half mirror enters an alignment optical system (hereinafter referred to as “LIA optical system”) 52 for detecting the diffracted light from the grating mark by the heterodyne beam.

First, from the FIA optical system 48 side, the illumination light from the illumination light source 49 passes through the FIA optical system 48 and is then deflected by the half prism 47 and the deflecting mirror 46 to produce the reference mark on the reference mark plate 6. Illuminate. The return light follows the same optical path and returns to the FIA optical system 48, the light passing through the FIA optical system 48 enters the half prism 50A, and the light flux passing through the half prism 50A is imaged for the X-axis composed of a two-dimensional CCD. The reference mark image on the reference mark plate 6 is formed on the image pickup surface of the element 51X, and the light flux reflected by the half prism 50A is formed on the image pickup surface of the Y-axis image pickup element 51Y including a two-dimensional CCD. The reference mark image on 6 is formed.

An image as shown in FIG. 14A is formed on the image pickup surfaces of the image pickup devices 51X and 51Y. The reference marks on the reference mark plate 6 are the lattice points of the lattice pattern, and the image 37P of the lattice pattern is projected in FIG. When the lattice pitch of the image 37P of the lattice pattern on the reference mark plate 6 is P and the width of the dark line is L, the width L is set to be considerably smaller than the lattice pitch P. In addition, on the image pickup surface, reference mark (index mark) images 48X1, 48X2 and Y for the X direction, which are illuminated with illumination light different from the illumination light of the reference mark plate 6, are provided.
The index mark images 48Y1 and 48Y2 for directions are formed. The position of the reference mark on the reference mark plate 6 can be detected with reference to the positions of those index mark images.

Specifically, in FIG. 14A, the image pickup area 51Xa in the direction conjugate with the X direction and the image pickup area 51Ya in the direction conjugate with the Y direction are respectively the image pickup elements 51 in FIG.
Imaged at X and 51Y. Imaging devices 51X and 51Y
The horizontal scanning lines are in the directions conjugate with the X and Y directions, respectively, and the image pickup signals S5X and S5Y of the image pickup devices 51X and 51Y are the signal processing devices 56 of FIG.
Is supplied to. In the signal processing device 56, the image pickup signal S5X
And S5Y are respectively averaged to obtain the image signal S5X 'in FIG. 14B and the image signal S5Y' in FIG. 14C, and the position of the target reference mark on the reference mark plate 6 is obtained from these image signals. Find the gap. A more detailed structure is disclosed in Japanese Patent Application No. 4-16589.

When the fiducial mark to be detected is the fiducial mark 37A in FIG. 10, relative displacement of the fiducial mark 37A in the X and Y directions with respect to the reference mark obtained by the image processing of FIG. 14A. _Are AO ' _fX and AO' _fY , respectively. Since the position of the reference mark plate 6 at this time is managed by the wafer coordinate system, the value obtained by subtracting the tracking error and the rotation error of the Zθ axis drive stage 4 of FIG. 7A from the measurement result is shown in FIG. It becomes the X component AO _X and the Y component AO _{Y of the} error vector AO. However, the FIA of FIG.
The X component AO _x and the Y component AO _y corresponding to the optical system 48 are referred to as AO _fX and AO _fY , respectively. That is, the following equation is obtained.

[0071]

[Number 5] _{AO fX = AO 'fX - (} WaAOX-WaAX)

[0072]

AO _fY = AO ′ _fY − (WaAY1 + WaAY2) / 2 On the other hand, in the alignment system including the LIA optical system 52 of FIG. After that, the deflection mirror 4
The light is deflected by 5 and is incident on the reference mark in the form of a diffraction grating on the reference mark plate 6. The diffracted light from the reference mark follows the same optical path and returns to the LIA optical system 52, where the LIA optical system 5
The diffracted light that has passed through 2 is split into two by the half prism 50B and is incident on the light receiving element 55X for the X direction and the light receiving element 55Y for the Y direction.

At this time, the laser light from the laser light source 53 is divided into two in the LIA optical system 52, and a frequency difference of Δf is given to the frequencies of the two laser lights by the internal frequency shifter. The interference light of these two laser lights is received by the light receiving element 54, and the reference signal S6 of the frequency Δf is output from the light receiving element. Also, those 2
Laser beams having different frequencies (heterodyne beams) are irradiated onto a reference mark in the form of a diffraction grating on the reference mark plate 6 at an appropriate incident angle, and the ± 1st-order diffracted lights of the two laser beams by the reference mark are The reference mark plate 6 is returned in parallel and perpendicularly. The intensity of the ± 1st order interference light changes at the frequency Δf, but the phase changes depending on the X and Y coordinates of the reference mark. And
The light receiving element 55X outputs a beat signal S7X having a frequency Δf whose phase changes according to the X coordinate of the reference mark, and the light receiving element 55Y changes the phase according to the Y coordinate of the reference mark. Beat signal S7Y of frequency Δf
Is output, and the reference signal S6 and the beat signals S7X and S7 are output.
Y is supplied to the signal processing device 56.

If the reference mark to be detected is the reference mark 37A shown in FIG. 10, the signal processing device 56 shown in FIG.
As shown in FIG. 4 (d), the reference signal S6 and the beat signal S7
From the phase difference Δφ _X with X, the positional deviation AO ′ _LX of the reference mark 37A in the X direction is obtained, and as shown in FIG.
From the phase difference Δφ _Y between the reference signal S6 and the beat signal S7X, the positional deviation AO ′ _LX of the reference mark 37A in the Y direction is obtained. Values obtained by subtracting the tracking error and the rotation error of the Zθ axis drive stage 4 of FIG. 7A from the measurement result become the X component AO _X and the Y component AO _Y of the error vector AO of FIG. 10. However, the X component AO _X and the Y component AO _Y corresponding to the LIA optical system 52 of FIG. 13 are AO _LX and AO _LY , respectively. That is, the following equation is obtained.

[0075]

[Expression 7] AO _LX = AO ' _LX- (WaAOX-WaAX)

[0076]

Equation 8] _{AO LY = AO 'LY - (} WaAY1 + WaAY2) / 2 or more in the manner, when the alignment at the position of symbol A is a group mark provided in Figure _{10, AL X, AL Y,} AR
_{X, AR Y, AO fX,} AO fY, AO LX, 8 pieces of data AO _LY is measured. In this sequence, 32 pieces (= 8 × 4) of data are obtained by measuring the mark group with the code A to the mark group with the code D. Of these 32 data, RA microscope 1
The data obtained by 9 and 20 are measured data Dxn, D
The data obtained by the off-axis alignment device 34 is stored as the actual measurement data Axn, Ayn. Thereafter, the operation shifts to step 111 in FIG.

In step 111 of FIG. 2, the measured data D _xn and D _yn corresponding to the RA microscopes 19 and 20 are coordinated so that the reticle coordinate system and the wafer coordinate system can be actually converted only by a linear error. X direction and y in the system
If the directional coordinates are F _xn and F _yn , then the relationship between them is as follows.

[0078]

[Equation 9]

In addition, the nonlinear error in the x direction and the y direction is expressed by ε
_{Given xn} and ε _yn , the following equation holds.

[0080]

[Equation 10]

Then, these nonlinear errors (ε _xn , ε _yn )
Using the least-squares approximation so that
The values of the two parameters Rx, Ry, θ, ω, Ox, Oy are calculated. Where the scaling parameter Rx in the x direction
Represents a magnification error between the reticle 12 and the reference mark plate 6 in the x direction, and the scaling parameter Ry represents a scaling error in the scanning direction (y direction) between the reticle coordinate system and the wafer coordinate system. Further, the angle parameter θ is the rotation error between the reticle 12 and the reference mark plate 6, the angle parameter ω is the parallelism in the scanning direction between the reticle coordinate system and the wafer coordinate system, and the offset parameters Ox and Oy are the x and y directions of both. The offset values are shown respectively.

Next, in steps 112 and 113 of FIG. 2, the baseline amount is obtained. In this case, the average values of the data A _xn and A _yn measured by the off-axis alignment device 34 are <Ax> and <Ay>, respectively.
The offset when measuring the baseline amount is (<A
x> -Ox, <Ay> -Oy). Therefore, at the time of alignment, an interferometer that uses the laser beam LWX of FIG. 7A (hereinafter, also referred to as “exposure interferometer LWX”)
To the interferometer using the laser beam LW _OF (hereinafter, also referred to as “off-axis interferometer LW _OF ”) and using the FIA optical system 48 of FIG. 13, measured data A _xn and the average value of a _yn respectively and <Afx> and <AFY>. Then, the offset (<Afx> -Ox, <Afy> -Oy) is set to the laser beam LWY1, LWY2, L of FIG. 7A.
The alignment process may be performed by adding the measurement value of the interferometer corresponding to W _OF . On the other hand, when the LIA optical system 52 of FIG. 13 is used, the average values of the measured data A _xn and A _yn are set to <ALx> and <ALy>, respectively. Then, the measured value of the interferometer (<ALx> -Ox, <ALy>
An offset of −Oy) may be given.

The above correction method means that the reference coordinate system of the stage coordinate system is set based on the reference mark on the reference mark plate 6. In this case, in other words, for example, the reference marks 37A to 37A on the reference mark plate 6
The axis passing through 37D serves as the reference axis, and the reading value of the off-axis interferometer LW _OF on this reference axis when the reading value of the exposure interferometer LWX is 0 on this reference axis (yaw value) Is required. At the time of exposure, the reading value of the exposure interferometer LWX and the off-axis interferometer LW
_The result obtained by correcting the yaw value to the actual _OF reading is referred to as an “interfering interferometer value”, and the wafer 5 is aligned based on the interfering interferometer value. .

On the other hand, for example, in FIG. 7A, a method in which the reference axis of the stage coordinate system is the movable mirror 7X for the X axis may be used. In this case, first, in FIG.
In the state of (a), the reading value of the exposure interferometer LWX and the reading value of the off-axis interferometer LW _OF are simultaneously reset (set to 0), and the delivery interferometer value is used in the subsequent exposure. Instead of using the measured value itself. On the other hand, at the time of alignment, for example, the inclination angle θ _XF of the reference axis passing through the reference marks 37A to 37D on the reference mark plate 6 with respect to the moving mirror 7X is obtained, and the interval I between the laser beams LWX and LW _OF is calculated.
Using L, the value obtained by correcting IL · θ _XF for the reading value of the off-axis interferometer LW _OF is used. As a result, the reading value of the exposure interferometer LWX and the reading value of the off-axis interferometer LW _OF can be used as they are during normal exposure.

Next, since the measurement data D _xn and D _yn represent only the relative error between the wafer coordinate system and the reticle coordinate system, they can be obtained when the least-squares approximation calculation is performed on the basis of the wafer coordinate system. Parameters Rx, Ry, θ, ω, Ox,
All Oy are represented by a linear error of the reticle coordinate system with the wafer coordinate system as a reference. Therefore, assuming that the x-coordinate and the y-coordinate of the reticle coordinate system are r _xn ′ and r _yn ′, respectively, the reticle is based on the new coordinates (r _xn , r _yn ) obtained from the following equation according to the movement of the wafer coordinate system. Should be driven.

[0086]

[Equation 11]

In this processing, the offsets Ox and Oy have already been set.
Is corrected on the reticle side, it is only necessary to correct the offset of (<Ax>, <Ay>) as the baseline amount. Further, when the reticle coordinate system is used as the reference, the results are all reversed, and it is possible to make corrections in the wafer coordinate system. Further, these corrections may be separately controlled such that the correction is performed in the wafer coordinate system during the rough alignment and the reticle coordinate system is performed during the fine alignment.

As described above, according to the present embodiment, the reticle alignment and the baseline amount are checked using a plurality of marks during one reticle alignment. It becomes possible to average the alignment error and improve the alignment accuracy. Furthermore, throughput is also improved because all these steps are performed simultaneously. Further, since the reference mark plate 6 capable of simultaneously measuring a plurality of reference marks in the non-scanning direction (X direction) is adopted, an error due to air fluctuation in the optical path of the interferometer does not occur.

However, since the reference mark plate 6 moves stepwise in the scanning direction, the influence of air fluctuations can be considered. Therefore, when checking the baseline amount, when performing the process using the LIA optical system 52 of FIG. 13, the position of the wafer stage (Zθ axis drive stage 4 etc.) is locked by using the output values of the light receiving elements 55X and 55Y. Then, if the reticle alignment and the baseline amount are checked, the influence of air fluctuation can be minimized. Further, the reticle marks of this example are arranged at the four corners of the reticle 12 at a total of eight locations. This is not only for offset, but for checking the correspondence between the reticle coordinate system and the wafer coordinate system.
This is because the parameters Rx, Ry, θ, and ω are necessary, and it is advantageous to arrange the marks at the four corners to determine the parameters Ry, θ, and ω. Furthermore, when the reference mark plate 6 having a light emitting property is used, the light emitting portion is limited, and it is difficult to cause the entire surface of the reference mark plate 6 to emit light.

If the number of reticle marks on the reticle 12 is n, the offset parameters Ox and Oy are 1 /
It is averaged to n ^1/2, and the error of other parameters is also small. Therefore, the error becomes smaller as the number of reticle marks n is increased. The following is a result of simulating the relationship between the reticle mark number n and the parameter error and the baseline amount error. In the following, the variation at the four corners in the new coordinate system of (Equation 11) is represented by 3 times the standard deviation σ and in the unit [nm].

[0091]

[Table 1] From the above, by setting the number of reticle marks n to eight,
It can be seen that even if the reticle drawing error is 50 nm and the stage stepping error is 10 nm, the accuracy of the reticle alignment and the baseline amount check can be 10 nm or less. That is, if the processing speed is increased within the limit of the luminescent reference mark plate 6 and the reticle mark number n is increased, the accuracy can be further improved.

At this time, the patterning error on the fiducial mark plate 6 and the distortion error of the projection optical system 8 remain as errors in the new coordinate system, but since they hardly change, the exposure result is used as a reference data when adjusting the apparatus. There is no problem if the obtained error is removed as a system offset as compared with. In the embodiment described above, as shown in FIG. 8C, the reference marks 35A to 35A are formed on the reference mark plate 6.
A plurality of Ds are provided, and a plurality of reference marks 37A to 37D are also provided. However, even if only one reference mark 35A and one reference mark 37A are used, by scanning only the reticle 12 and averaging the measurement results, the influence of the drawing error of the pattern on the reticle 12 is reduced. It can be reduced.

Next, a second embodiment of the present invention will be described with reference to FIG.
16 and the flowchart of FIG. In this regard, the reticle alignment mode of the first embodiment described above uses four sets of fine alignment marks 29A to 29D and 30A to 30D on the reticle to perform fine reticle alignment. However, after the fine reticle alignment is performed once by the method of the first embodiment, when the scaling error in the scan direction or the parallelism between the reticle coordinate system and the wafer coordinate system in the scan direction is small, Reticle alignment and baseline measurement may be performed using a set of fine alignment marks. 1 like this
An alignment mode in which a set of fine alignment marks is used to perform measurement on three items of magnification (Rx) measurement in the non-scan direction, rotation (θ) measurement, and baseline measurement is called a “quick mode”. This quick mode can be applied even when it is known in advance that the drawing error of the fine alignment marks 29A to 30D on the reticle 12 is small.

In this quick mode, for example, a set of fine alignment marks 29A on the reticle 12,
30A and a set of reference marks 35 on the reference mark plate 6
A, 36A and one fiducial mark 3 on the fiducial mark plate 6
Magnification (Rx) measurement in the non-scan direction using 7A and
Rotation (θ) measurement and baseline measurement 3
Measure the items. However, in the case of this quick mode, one set of fine alignment marks 29A,
In order to correct the drawing error of 30A, it is necessary to store the drawing error of the marks 29A and 30A obtained by the fine alignment sequence.

The operation of the second embodiment will be described with reference to FIGS.
Will be described with reference to. The operation of FIGS. 15 and 16 is similar to that of FIG.
This is an operation in which the quick mode is added to the operation of, and it is possible to switch between the fine mode and the quick mode. In the steps of FIG. 15, steps corresponding to the steps of FIG. 2 are designated by the same reference numerals, and detailed description thereof will be omitted.

In FIG. 15, steps 101-104.
2, the reticle 12 is placed on the reticle holder, and the positions of the rough search alignment marks 27 and 28 are set to the RA microscopes 19 and 2, respectively.
Detects at 0. Next, at step 115, either the fine mode or the quick mode is selected. The selection result is instructed in advance by the operator via the keyboard 22C of FIG. However, a bar code not shown
The reader may read the pattern information of the reticle 12 and the like, and the main control system 22A may automatically select the alignment mode based on the result.

When the fine mode is selected, steps 105 to 113 of FIG. 15 are executed, and the measurement results of reticle alignment and fine alignment using a plurality of fine alignment marks and a plurality of reference marks are used as described above. Baseline measurement is performed. Then, in step 114, on the new coordinate system on the reticle 12, drawing errors of the actual positions of the fine alignment marks 29A and 30A with respect to the original positions (hereinafter, referred to as
“Mark error”) is obtained and the mark error is stored in the storage unit in the main control system 22A. When obtaining the mark error, the reticle coordinate system is obtained with the wafer coordinate system as a reference from the relationship (conversion parameter) obtained in step 113, and the fine alignment marks 29A to 29D and 30A to 30D are designed on this reticle coordinate system. Find the non-linear error of the measured coordinate values relative to the above coordinate values. This non-linear error becomes a mark error. In this way, at the time of fine alignment, the mark error on the new coordinate system on the reticle is stored from the results of steps 112 and 113. Further, when the reticle drawing error is measured in advance, the operator may directly input the drawing error. The effect is particularly great when the drawing error includes a linear component.

On the other hand, when the quick mode is selected in step 115, the operation shifts to step 116 in FIG. Then, in steps 116 to 118, the same operations as steps 105 to 107 in FIG. 15 are executed. That is,
The image of the pair of fine alignment marks 30A and 29A on the reticle 12 and the pair of reference marks 36A and 35A on the reference mark plate 6 is observed by the RA microscope in the quick mode, and the off-axis alignment device 3 is used.
4 detects one fiducial mark 37A. Also,
In the latter half of step 119, the positions of the marks observed by the RA microscope and the marks detected by the off-axis alignment device 34 are obtained. Then step 11
9, the fine alignment marks 30A and 29A on the reticle 12 are detected with respect to the detected positions shown in FIG.
The mark error obtained in step 114 is corrected. As a result, at least the number of marks measured in the quick mode can be corrected so that the pattern drawing error on the reticle 12 is almost the same as in the fine alignment mode of the first embodiment.

Next, in step 120, based on the position of each mark obtained by the correction in step 119, the six conversion parameters (Rx, Ry,
θ, ω, Ox, Oy), the magnification error Rx in the non-scan direction, the rotation θ, and the offsets Ox, Oy
Ask for. Specifically, as shown in FIGS. 8A and 8C, the measured mark intervals of the reference marks 35A and 36A in the X direction (non-scanning direction) and the mark images 29AW and 3A.
The magnification error Rx in the non-scanning direction is obtained from the difference from the 0 AW interval in the X direction. Further, the positional deviation of the reference marks 35A and 36A in the Y direction (scanning direction) and the mark image 29A
The rotation θ is obtained from the difference between the positional deviation of W and 30 AW in the Y direction and the mark interval. Also, the offset O
x and Oy are obtained from the average positional deviation amount between the reference mark and the mark image of the reticle.

In this quick mode, since there are two marks to be measured on the reticle side and the reference mark plate 6 side, four conversion parameters out of the six conversion parameters of (Equation 9) are used. Only parameters can be determined. Therefore, the values of the four conversion parameters are obtained as described above. Note that, for example, two fine alignment marks 29A and 29D arranged in the Y direction of FIG. 4 and FIG.
The magnification error Ry in the scanning direction can be obtained by selecting the two reference marks 35A and 35D as the measurement target.

Then, reticle alignment is performed based on the magnification error Rx in the non-scanning direction, the rotation θ, and the offsets Ox and Oy obtained in step 120. For the measurement of the magnification error, a magnification error corresponding to the deviation of the measured value of each mark with respect to the design value of each mark is prepared in advance as a table, and the deviation of the measured value of each mark with respect to the design value of each mark is May be applied to the table to obtain the magnification error.

Next, in step 121, the measured values of the central coordinates of the reference marks 35A and 36A and the reference mark 37.
Baseline measurement is performed using the measurement value of A and. As described above, according to this embodiment, the fine alignment mode is once executed to obtain the drawing error (mark error) of the pattern of the reticle 12, and when the alignment is executed in the quick mode, the mark error Since the correction is performed, the alignment of the slit scan type projection exposure apparatus can be performed with high throughput and high accuracy.

Next, a third embodiment of the present invention will be described with reference to the flowchart of FIG. In the third embodiment, reticle alignment and baseline measurement are performed in the above-mentioned quick mode every time a predetermined number of wafers are exchanged, that is, every time a predetermined number of wafers are exposed. In this embodiment, after exchanging the reticle with the projection exposure apparatus of FIG.
FIG. 17 shows an example of the operation when the two patterns are sequentially exposed.
Will be described with reference to.

First, in step 211 of FIG. 17,
The reticle used previously is replaced with the reticle 12 shown in FIG. 1, and the exposure operation is started. In this case, step 1 in FIG.
01-104 and 115, and step 11 of FIG.
The operations of reticle alignment and baseline check in the quick mode shown in 6 to 121 are executed. Thereafter, in step 212, the number of wafers to be exposed before the next reticle alignment and baseline check is set to the variable N as an initial value, and in step 213, the wafer is loaded on the wafer stage 4. However, if there is a wafer that has already been exposed in step 213, the exposed wafer is unloaded (unloaded) and then a new wafer is loaded.

Next, in step 214, it is checked whether or not the variable N is 0, that is, whether or not it is time to perform reticle alignment and baseline check.
If the variable N is greater than 0, then in step 215 the variable N
1 is subtracted from and the process proceeds to step 216. In this step 216, after the wafer is aligned by using the off-axis type alignment device 34 of FIG. 13 or the TTL type wafer alignment system, the pattern of the reticle 12 is exposed on each shot area of the wafer.
When exposure of all (designated number of) wafers is completed, the exposure process for the reticle 12 is completed, but if exposure of all wafers is not completed, step 21
Returning to step 3, the exposed wafer is unloaded and a new wafer is loaded. Thereafter, the operation is step 214.
Move to.

When N = 0 in step 214, that is, when it is the timing to perform reticle alignment and baseline check, in step 217 the rotation error and magnification error of the reticle 12 are measured. This is similar to step 120 in FIG. After that, the process proceeds to step 218, where the off-axis alignment device 34 (the alignment system including the FIA optical system 48 or the two-beam interference alignment wafer alignment system including the LIA optical system 52) is in the X and Y directions. A baseline check will be performed. Then step 21
In step 9, the number of wafers to be exposed before the baseline check is set as the variable N, and the operation returns to step 216.

As described above, according to this embodiment, the reticle alignment and the baseline measurement are performed every time the reticle is exchanged, and the reticle alignment and the baseline measurement are performed in the quick mode every time a predetermined number of wafers are exposed. Therefore, the overlay accuracy of each wafer and the pattern image of the reticle can be improved with high throughput.

Further, although the method of the above-described embodiment has been described with reference to the baseline measurement at the time of off-axis alignment, the TTL using the field of the projection optical system is used.
Even in the (through the lens) system, the same effect can be expected by applying the present invention. As described above, the present invention is not limited to the above-described embodiments, and various configurations can be adopted without departing from the gist of the present invention.

In the projection exposure method described above, the illumination area of the predetermined shape is illuminated with illumination light, and the pattern image on the mask (12) in the illumination area of the predetermined shape is passed through the projection optical system (8). The mask (1) is exposed by exposing the substrate (5) on the stage (4) and scanning the mask (12) and the substrate (5) in synchronization with each other with respect to an illumination region having a predetermined shape.
2) In the method of exposing a pattern image having a larger area than the illumination area of the predetermined shape on the substrate (5), a plurality of measurement marks (29A) are provided on the mask (12) in the relative scanning direction. .About.29D) are formed, and a plurality of reference marks (35A to 35D) are formed at positions substantially conjugate with the plurality of measurement marks.
35D) is formed on the stage (4), and the reference mark member (6) is placed on the stage (4), and the mask (12) and the substrate (5) are moved in synchronism with the relative scanning direction. ) One of the plurality of measurement marks on the measurement mark (29A, 29B, ...) and the corresponding reference mark (35A, 35B, ...) on the stage (4) are sequentially displaced. The measurement is performed, and the correspondence between the coordinate system on the mask (12) and the coordinate system on the stage (4) is obtained from the positional deviation amounts of the plurality of measurement marks and the plurality of reference marks. .

Further, in the projection exposure method described above, the off-axis for detecting the position of the positioning mark on the substrate (5) in the vicinity of the projection optical system (8) is the same premise as in the invention described above. System alignment system (34) is arranged, a plurality of measurement marks (29A to 29D) are formed on the mask (12) in the relative scanning direction, and the measurement marks (29A to 29D) are formed in the exposure field of the projection optical system (8). A reference mark member (6) having a first reference mark (35A) and a second reference mark (37A) formed at an interval corresponding to the interval between the reference point and the reference point of the off-axis alignment system (34). ) Is placed on the stage (4).

Then, with the off-axis alignment system (34) observing the second reference mark (37A) on the reference member (6), the mask (12) is moved in the relative scanning direction. By moving, one measurement mark (29A, among the plurality of measurement marks on the mask (12),
29B, ...) and the first reference mark (35A) on the stage (4) are sequentially measured, and the plurality of measurement marks and the first reference marks are respectively displaced. Of the reference point in the exposure field of the projection optical system (8) and the off-axis alignment system based on the average value of the The distance from the reference point of (34) is obtained.

Further, in the above-mentioned projection exposure method, a plurality of first reference marks are further provided on the reference mark member (6) so as to correspond to the plurality of measurement marks (29A to 29D) on the mask (12). Individually (35A to 35D) are formed, and the reference points in the exposure field of the projection optical system (8) and the off-axis alignment system (34) are formed from the plurality of first reference marks (35A to 35D). A plurality of second reference marks (37A to 37D) are formed at an interval corresponding to the interval with the reference point, and the mask (12) and the stage (4) are moved in synchronization with the relative scanning direction. Of the measurement marks on the mask (12)
One measurement mark (29A, 29B, ...) And the corresponding first reference mark (35A, 35) on the stage (4).
B, ...) is sequentially measured, and a plurality of second axes are provided by the off-axis alignment system (34).
Corresponding reference mark (37A, 37
B, ...) are observed, and the average value of the positional deviation amounts of the plurality of measurement marks and the plurality of first reference marks and the off-axis alignment system (3
Based on the average value of the positional deviation amounts of the plurality of second reference marks observed in 4), the reference point in the exposure field of the projection optical system 8 and the off-axis alignment system 3
The distance from the reference point in 4) is obtained.

Further, in the projection exposure method described above, a plurality of measurement marks (29) on the mask (12) are used in the same premise as in the invention described above, similarly to the first projection exposure method described above.
A, 29B, ...) and corresponding reference marks (35A, 35B)
B, ...) and a first step of obtaining respective positional deviation amounts; and a predetermined one of a plurality of measurement marks on the mask.
The amount of positional deviation between one measurement mark (29A) and the corresponding reference mark (35A) on the stage is measured only once, and the amount of positional deviation between the measurement mark (29A) and the reference mark (35A) is simplified. Second step to be obtained in a desired manner; either one of the first step and the second step is selected, and the amount of positional deviation between the measurement mark and the reference mark obtained in the selected step Based on the mask (1
2) The third step of obtaining the correspondence between the coordinate system on the stage and the coordinate system on the stage (4).

Further, in the projection exposure method described above, the mask ((A), (37A, 37B, ...) On the fiducial mark member (6) is observed by an off-axis alignment system while observing the mask ( 12) is moved in the relative scanning direction so that one of the plurality of measurement marks (29A, 29B, ...) On the mask (12) and the first reference mark (35A, 35B, ...) Sequentially measuring the amount of positional deviation with the mask; while observing the second fiducial mark (37A) on the fiducial mark member (6) with the off-axis alignment system, the mask ( 12) A second step of simply measuring the amount of positional deviation between a predetermined one measurement mark (29A) among the plurality of measurement marks above and the first reference mark (35A); Either the process or the second process And a third step of selecting; and a measurement result in the step selected in the third step, the amount of positional deviation between the measurement mark and the reference mark, and the off-axis alignment system. Correspondence between the coordinate system on the mask and the coordinate system on the stage, the reference point in the exposure field of the projection optical system, and the off-axis alignment system based on the displacement amount of the second reference mark. Distance from the reference point (baseline amount)
And a fourth step for obtaining.

In the projection exposure method described above, the off-axis alignment system (34) for detecting the position of the positioning mark on the substrate (5) is arranged near the projection optical system (8). On the mask (12), a plurality of measurement marks (29A, 29B,
...) is formed, and a reference mark member (6) having a plurality of reference marks formed at positions substantially conjugate to the plurality of measurement marks is arranged on the stage (4), and the plurality of reference marks are projected. The first (35A, 35B, ...) And the second (3) formed at an interval corresponding to the interval between the reference point of the optical system and the reference point of the off-axis alignment system.
7A, 37B, ...) and the substrate (5)
A plurality of measurement marks on the mask (12) with the second fiducial mark (37A) on the fiducial mark member (6) being observed by the off-axis type alignment system every time a predetermined number of sheets are exchanged. Of the predetermined one of the measuring marks (29A) and the corresponding first reference mark (35A) is measured, and the amount of positional deviation measured in this way and the off-axis method thereof are measured. From the position shift amount of the second reference mark (37A) observed with the alignment system,
Correspondence between the coordinate system on the mask and the coordinate system on the stage, and the distance (baseline amount) between the reference point in the exposure field of the projection optical system and the reference point of the off-axis alignment system. Is to seek.

In such a projection exposure method, a plurality of measurement marks are arranged on the mask (12), and a reference mark almost conjugate to the measurement marks is arranged on the reference mark member (6). , Mask (12) and stage (4)
The position deviation of each mark is measured while feeding with the stepping method. Then, finally, in accordance with the positional shift obtained at each position, for example, by parameters such as least square approximation, the parameters (magnification, scaling in the scanning direction, rotation, parallelism in the scanning direction) that associate the mask coordinate system with the substrate coordinate system are associated. By obtaining the angle, the offset in the X direction, and the offset in the Y direction, the influence of the drawing error of the measurement mark on the mask (12) can be suppressed to a small level. Further, since the mark measurement in the relative scanning direction is sequentially performed separately, the measurement is non-simultaneous, but since the measurement is performed at a plurality of points, there is an averaging effect and high-precision measurement is possible.

Further, according to the above-mentioned projection exposure method, the influence of the drawing error of the measurement mark of the mask (12) is reduced by averaging the measurement results of the plurality of measurement marks on the mask (12) side. Then, the baseline amount, which is the distance between the reference point of the projection optical system (8) and the reference point of the alignment system (34), can be accurately measured. Further, according to the above-mentioned projection exposure method, a plurality of the first reference marks ((A) to (D) corresponding to the plurality of measurement marks (29A to 29D) on the mask (12) are provided on the reference mark member (6). 35A ~ 3
5D), the reference point in the exposure field of the projection optical system (8) and the reference point of the off-axis alignment system (34) are formed from the plurality of first reference marks (35A to 35D). Since a plurality of second reference marks (37A to 37D) are formed at intervals corresponding to the intervals, averaging is also performed on the reference mark side.
The baseline amount is measured more accurately.

Further, according to the projection exposure method described above, when high throughput is required, the second step is selected and the amount of positional deviation between the measurement mark (29A) and the reference mark (35A) is set once. By performing only the measurement and performing the first step when high accuracy is required, the requirement for swiftness can also be satisfied. In this case, in the first step, the amount of positional deviation of the measurement mark (29A) from the original position (referred to as “mark error”) is further obtained and stored in advance, and the second step is executed. By correcting the mark error, it is possible to meet the requirements of both high throughput and high accuracy.

According to the projection exposure method described above, when high throughput is required, the second step is selected and the second fiducial mark (37) is set by the alignment system (34).
One measurement mark (29A) while observing A)
The requirement for quickness can also be satisfied by measuring the positional deviation amount between the first reference mark (35A) and the first reference mark (35A) only once and executing the first step when high accuracy is required. In this case, in the first step, the mark error is further calculated and stored in advance, and when the second step is executed, the mark error is corrected to meet the demands for both high throughput and high accuracy. I can respond.

According to the projection exposure method described above, the off-axis type alignment system (34) is replaced each time a predetermined number of substrates (5) are exchanged, that is, each time a predetermined number of substrates (5) are exposed. While observing the second reference mark (37A), the amount of misalignment between one measurement mark (29A) and the first reference mark (35A) was measured only once, and the measurement result on the mask The base line amount and the correspondence relationship between the coordinate system of 1) and the coordinate system on the stage are obtained. Therefore,
Measurement is performed with high throughput.

Further, according to the above-mentioned projection exposure method, finally, for example, in accordance with the positional deviations obtained at the respective positions of the plurality of measurement marks on the mask, the mask coordinate system is calculated by least-squares approximation or the like. By obtaining the parameters (magnification, scaling in the scanning direction, rotation, parallelism in the scanning direction, offsets in the X and Y directions) that are associated with the substrate coordinate system, the influence of drawing errors of the measurement marks on the mask can be minimized. be able to.

Further, according to the above-described projection exposure apparatus, the drawing error of the measurement mark on the mask is reduced by averaging the measurement results of the plurality of measurement marks on the mask side, and the reference of the projection optical system is reduced. It is possible to accurately measure the baseline amount, which is the distance between the point and the reference point of the alignment system. Further, according to the projection exposure apparatus described above, a plurality of first reference marks corresponding to the plurality of measurement marks on the mask are formed on the reference mark member, and the plurality of first reference marks are formed. Since a plurality of second reference marks are formed at an interval corresponding to the interval between the reference point in the exposure field of the projection optical system and the reference point of the off-axis alignment system, Since the averaging is performed, the baseline amount is measured more accurately.

Further, according to the above-mentioned projection exposure method, by selecting a simple measurement process in the quick mode, the correspondence between the coordinate system on the mask and the coordinate system on the stage can be established with high throughput if necessary. You can ask. Further, according to the projection exposure method described above, by selecting the simple measurement process in the quick mode, the correspondence relationship between the coordinate system on the mask and the coordinate system on the stage and the base can be set with high throughput as necessary. The line amount can be obtained.

Further, according to the above-described projection exposure method, since a simple measurement process in the quick mode is executed every time a predetermined number of substrates are exposed, a large number of substrates can be continuously scanned. When performing exposure, the correspondence between the coordinate system on the mask and the coordinate system on the stage and the baseline amount can be obtained with high throughput.

[0125]

According to the present invention, the reticle coordinate system (mass
Correspondence between the wafer coordinate system) and the wafer coordinate system (substrate coordinate system)
And perform baseline measurement with high throughput.
The pattern image of the mask can be accurately projected and exposed on the substrate.

[Brief description of drawings]

FIG. 1 is a configuration diagram showing a projection exposure apparatus according to an embodiment of the present invention.

FIG. 2 is a flowchart showing an alignment method and a baseline amount checking method of the first embodiment.

FIG. 3 is a perspective view showing a reticle loader system.

4A is a layout view of alignment marks on a reticle, FIG. 4B is a layout view showing alignment marks and the like in a region conjugate with an effective visual field of the projection optical system, and FIG. It is an enlarged view which shows 30D.

FIG. 5A is an explanatory diagram for performing a rough alignment of a reticle, and FIG. 5B is a reduced diagram of FIG. 5A.

FIG. 6 is a waveform diagram showing various image pickup signals obtained from an image pickup device when performing rough alignment of a reticle.

FIG. 7A is a plan view of a wafer-side stage,
(B) is a plan view of the stage on the reticle side.

8A is a projection view showing mark arrangement on the reticle, FIG. 8B is an enlarged projection view showing an example of marks on the reticle, and FIG. 8C shows arrangement of reference marks on the reference mark plate 6. A plan view, (d) is an enlarged view showing an example of the reference mark 35A and the like, and (e) is a plan view showing an example of the reference mark 37A and the like.

FIG. 9 is a plan view showing a relationship among a reference mark plate, a reticle, a projection optical system, and an alignment device during reticle alignment and measurement of a baseline amount.

FIG. 10 is a diagram showing an error vector obtained by reticle alignment and baseline amount measurement.

FIG. 11 is a partially cutaway configuration diagram showing configurations of a reticle alignment microscope 19 and an illumination system.

12A is a diagram showing an image observed by the image sensor of FIG. 11, and FIGS. 12B and 12C are X corresponding to the image.
It is a wave form diagram which shows the image signal of the Y direction.

FIG. 13: Off-axis type alignment device 3
4 is a configuration diagram showing FIG.

14A is a diagram showing an image observed by the image pickup device of FIG. 13, and FIGS. 14B and 14C are X corresponding to the image.
13A and 13B are waveform diagrams showing image signals in the Y direction and the Y direction, and FIGS. 14D and 14E are waveform diagrams showing detection signals obtained through the LIA optical system in FIG.

FIG. 15 is a flowchart showing a part of the operations of the alignment method and the baseline amount checking method of the second embodiment.

FIG. 16 is a flowchart showing the remaining operations of the alignment method and the baseline amount checking method of the second embodiment.

FIG. 17 is a flowchart showing the operation of the exposure method of the third embodiment.

FIG. 18 is a partially cutaway view showing an alignment system of a conventional stepper.

[Explanation of symbols]

4 ... Zθ axis drive stage, 5 ... Wafer, 6 ... Reference mark plate, 7 ... Wafer side moving mirror, 8 ... Projection optical system, 11 ... Reticle micro-driving stage, 12 ... Reticle, 19, 20
... Reticle alignment microscope (RA microscope), 21 ...
Reticle-side movable mirror, 27, 28 ... Rough search alignment mark, 29A-29D, 30A-30D ... Fine alignment mark, 34 ... Off-axis type alignment device, 35A-35D, 36A-36D ...
Reference mark, 37A to 37D ... Reference mark

─────────────────────────────────────────────────── ─── Continuation of the front page (58) Fields surveyed (Int.Cl. ⁷ , DB name) H01L 21/027 G01B 21/00 G03F 7/22 G03F 9/00

Claims (10)

(57) [Claims] 1. A mask held on a first stage and a second stage.
To move in synchronization with the substrate held on the stage
Therefore, the substrate is exposed using the pattern of the mask.
In the projection exposure method described above, prior to the exposure, a first mode in which a plurality of marks formed on the mask are detected by a detection system, and a number smaller than the number of marks detected in the first mode, Second detection of the mark formed on the mask by the detection system
Selects one of the modes, on the mask using the detecting system in the selected mode
Mark and the mark on the second stage are detected, and the movement of the mask is controlled based on the detection result.
A first coordinate system for controlling the movement and a second coordinate system for controlling the movement of the substrate.
A projection exposure method characterized by finding a correspondence with a coordinate system . 2. A mask held on the first stage and a second stage.
To move in synchronization with the substrate held on the stage
Therefore, the substrate is exposed using the pattern of the mask.
In the projection exposure method described above, prior to the exposure, a first mode in which a plurality of marks formed on the mask are detected by a detection system, and a number smaller than the number of marks detected in the first mode, Second detection of the mark formed on the mask by the detection system
Selects one of the modes, on the mask using the detecting system in the selected mode
Mark and the mark of the second stage are detected, and based on the detection result, an image of the pattern of the mask is detected.
The projection reference point of the projection system for projecting onto the substrate and the substrate reference point.
Off-axis alignment that detects alignment information
The projection exposure method is characterized in that the positional relationship with the detection reference point of the ment system is obtained . 3. The method according to claim 1, wherein the second mode detects a part of the plurality of marks detected in the first mode. 4. The parallelism between a scanning direction defined by the first coordinate system and a scanning direction defined by the second coordinate system by detecting a plurality of marks on the mask in the first mode. The method according to claim 1, wherein 5. In the first mode, by detecting a plurality of marks on the mask, a scaling in a scanning direction defined by the first coordinate system and a scaling in a scanning direction defined by the second coordinate system are performed. The method according to claim 1, wherein a correspondence relationship with is obtained. 6. The second mode can be selected after the first mode is executed.
5. The method according to any one of 5 to 5. 7. The method according to claim 6, wherein in the second mode, only some of the parameters obtained in the first mode are obtained. 8. The method according to claim 7, wherein a drawing error of the mask is obtained in the first mode, and the drawing error is applied to the second mode. 9. A mask held on the first stage and a second stage.
To move in synchronization with the substrate held on the stage
Therefore, an exposure apparatus for exposing a substrate by using the pattern of the mask, a plurality of marks of the mark detecting system capable of detecting the marks of the mask held by the first stage, said mask by said mark detection system Is selected, or a second mode for detecting a number of marks in the mask that is smaller than the number of marks detected in the first mode by the mark detection system is selected. And a control system, the control system prior to the exposure.
Mode or the second mode, and the detection system is used in the selected mode on the mask.
Mark and the mark on the second stage are detected, and the movement of the mask is controlled based on the detection result.
A first coordinate system for controlling the movement and a second coordinate system for controlling the movement of the substrate.
An exposure apparatus characterized by finding a correspondence with a coordinate system . 10. A mask held on the first stage and a first stage.
To move the substrate held on two stages in synchronization
Accordingly, an exposure apparatus for exposing a substrate by using the pattern of the mask, a plurality of marks of the mark detecting system capable of detecting the marks of the mask held by the first stage, said mask by said mark detection system Is selected, or a second mode for detecting a number of marks in the mask that is smaller than the number of marks detected in the first mode by the mark detection system is selected. And a control system, the control system prior to the exposure.
Mode or the second mode, and the detection system is used in the selected mode on the mask.
Mark and the mark of the second stage are detected, and based on the detection result, an image of the pattern of the mask is detected.
The projection reference point of the projection system for projecting onto the substrate and the substrate reference point.
Off-axis alignment that detects alignment information
The exposure apparatus is characterized in that the positional relationship with the detection reference point of the ment system is obtained .