JPH08339957A - Exposure method - Google Patents

Abstract

PURPOSE: To use a sequence by a die-by-die method and choose a sample shot readily when carrying out EGA method alignment by an alignment sensor, of a TTR method, for example. CONSTITUTION: Misregistration between a reticle and a shot region of a wafer is detected by a TTR method alignment sensor, a reticle stage is moved slightly and positioned for a shot region wherein enough alignment signal is obtained, exposure is carried out and a position of a reticle stage is measured (steps 101 to 107). An EGA parameter is obtained by using a movement amount of a reticle stage regarding all the shot regions wherein enough alignment signal is obtained (step 110). An arrangement coordinate of an unexposed shot region is obtained from an obtained EGA parameter and exposure is performed for an unexposed shot region based thereon (step 112).

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an exposure method for exposing a pattern of a reticle to each shot area on a wafer based on array coordinates predicted by using, for example, a statistical method,
In particular, it is suitable for application when using a TTR (through the reticle) type alignment sensor.

[0002]

2. Description of the Related Art When manufacturing a semiconductor device, a liquid crystal display device or the like by a photolithography process, a reticle pattern as a mask is formed on a wafer (or a glass plate or the like) coated with a photoresist through a projection optical system. A projection exposure apparatus (stepper or the like) that transfers to each shot area is used. For example, a semiconductor element is formed by stacking multiple layers of circuit patterns on a wafer. Therefore, when projecting and exposing the circuit patterns of the second and subsequent layers, each circuit pattern already formed on the wafer is exposed. The shot area and the pattern of the reticle to be exposed from now on need to be aligned with high accuracy, that is, wafer alignment (wafer alignment). As a highly accurate method of aligning a wafer in a conventional projection exposure apparatus, for example, Japanese Patent Application Laid-Open No. 61-
As disclosed in Japanese Patent No. 44429, an alignment mark (wafer mark) attached to a predetermined number of shot areas (sample shots) selected on a wafer.
There is known an enhanced global alignment (hereinafter abbreviated as "EGA") type alignment method in which the coordinate position of each shot area is measured and the measurement result is statistically processed to calculate the array coordinates of each shot area on the wafer. There is.

On the other hand, before the exposure of each shot area on the wafer is performed, the position of the wafer mark (or the amount of positional deviation from the reticle mark) attached to each shot area is measured, and this measurement is performed. Site-by-site method that performs alignment based on the result, or die-by-site method
A die type alignment method is also known. The former site-by-site method is an alignment method that uses an alignment sensor such as a TTL (through-the-lens) method or an off-axis method in which the measurement position and the exposure position are different. is there. On the other hand, the latter die-by-die method is a TTR (through-the-reticle) method that directly measures the amount of misalignment between the wafer mark and the alignment mark (reticle mark) on the reticle via the projection optical system. As described above, this is an alignment method in the case of using an alignment sensor in which the measurement position and the exposure position are the same. When performing the die-by-die type alignment using the TTR type alignment sensor as described above, information on the amount of positional deviation measured by the alignment sensor is sent to the control system of the reticle stage, and the amount of positional deviation is measured. The position of the reticle stage has been finely adjusted so that is within a predetermined allowable value.

It is also possible to perform alignment by the EGA method using an alignment sensor of the TTL method or the off-axis method. At this time, since the measurement position and the exposure position are originally different, the measurement operation itself is the same as when the alignment is performed by the site-by-site method. Similarly, it is also possible to perform alignment by the EGA method using the TTR method alignment sensor. For this purpose, conventionally, the wafer stage is stepping driven to sequentially measure each wafer mark to be measured by the TTR method alignment sensor. The measurement position is set and the amount of positional deviation between each wafer mark and the corresponding reticle mark is measured. Then, as a control method after the measurement, conventionally, a method of sending the positional deviation amount obtained by the alignment sensor to the control system of the reticle stage via the central control system, or the positional deviation amount of the central control system and the reticle stage in parallel The method of sending to the control system was used.

[0005]

However, the method of sending the measured positional deviation amount to the control system of the reticle stage via the central control system as described above requires a long time for signal transmission and a long time for alignment. Therefore, there is a disadvantage that the throughput of the exposure process (the number of processed wafers per unit time) is reduced. In addition, the method of sending the measured positional deviation amount to the central control system as described above is a control method that is not used in the die-by-die method, and therefore the apparatus configuration becomes complicated to use that method. At the same time, it is necessary to incorporate another control sequence, which causes an inconvenience that the manufacturing cost increases.

Further, in the case of performing the EGA type alignment, a sample shot to be usually measured is designated in advance. However, if a shot area lacking a wafer mark is taken as a sample shot, the position of the sample shot cannot be measured. Therefore, for example, it is necessary to newly specify another sample shot and perform measurement, which causes a problem that the sequence becomes complicated.

In view of the above point, the present invention can utilize the control sequence in the case of performing the alignment by the die-by-die system when the alignment of the EGA system is performed by using the alignment sensor of the TTR system and the reticle. It is an object of the present invention to provide an exposure method capable of improving the overlay accuracy between the wafer and each shot area of the wafer and shortening the time required for alignment and exposure as compared with the conventional method. Another object of the present invention is to reliably and easily select sample shots.

[0008]

An exposure method according to the present invention exposes each of a plurality of shot areas on a substrate (6) with a pattern image on a mask (4), for example, as shown in FIGS. Of the plurality of shot areas on the substrate (6), a shot area that cannot be aligned with the pattern of the mask (4) using the alignment marks attached to each shot area is excluded. For each predetermined one of all the shot areas, the shot area is aligned with the pattern of the mask (4) by using the alignment mark provided in the shot area, and the pattern image is used. A first step (steps 101 to 109) of exposing a shot area and a mask obtained for each of the shot areas exposed in the first step. The second step of exposing each with the pattern image of the unexposed shot area on the substrate (6) on the basis of the positioning data of the turn (step 11
0 to 112).

In this case, an example of the alignment data is that the shot area to be measured and the pattern of the mask (4) are measured using the alignment marks attached to each shot area on the substrate (6). It is the amount of movement of at least one of the mask (4) and the substrate (6) with respect to the target position when aligning. Further, it is preferable that the second step has a step of calculating array data of unexposed shot areas on the substrate (6) based on the alignment data.

An example of a shot area which cannot be aligned with the pattern of the mask (4) in the first step is a shot area which is presumed to have a missing alignment mark. Another example of the shot area that cannot be aligned with the mask pattern in the first step is a shot area that is estimated in advance from the attached alignment mark.

[0011]

According to such an exposure method of the present invention, for example, TT
When aligning and exposing in the EGA method using an alignment system in which the measurement position and the exposure position are the same as in the R method, first, all M shot areas on the substrate (6) are measured in the first step. As a target shot region (sample shot), the mask-side stage and the substrate-side stage are aligned by a die-by-die method for these sample shots. In that case, after driving the stage on the substrate side to set each sample shot on the substrate (6) to the exposure position based on the designed array data, the sample shot and the mask (4) are set by the alignment sensor, for example. After measuring the position deviation amount (correctly, the position deviation amount of the two alignment marks), for example, by finely adjusting the position of the mask side stage so that the position deviation amount becomes a predetermined target value, The sample shot is exposed, and the amount of movement of the mask-side stage from the initial value at the time of this exposure is determined and stored. Then, at the time of the alignment, a sample shot that cannot be aligned due to a lack of alignment marks is left as an unexposed shot area.

The amount of movement of the stage on the mask side thus obtained is equivalent to the measured value when the mask (4) is fixed and the amount of positional deviation between each sample shot and the mask (4) is measured. Therefore, EGA alignment can be performed on the unexposed shot area using the movement amount as alignment data. Therefore, in the second step, the movement amount obtained in the first step and, for example, the substrate (6)
The designed array data of the upper unexposed shot areas is statistically processed to obtain the array coordinates of the unexposed shot areas on the substrate (6), and the alignment is performed based on the array coordinates.
Therefore, the alignment sequence of the EGA method is accumulated in a short time by using the alignment sequence of the die-by-die method as it is and measuring the moving amount of the mask (4) for each measurable sample shot. You can do it.

Further, since the position of the sample shot is measured and the sample shot is exposed at the same time by utilizing the characteristics of the die-by-die system, the precision of overlaying these sample shots is high. In addition, since the accurate position of the shot area is measured by using the exposure time, and the array coordinates of the unexposed shot area are calculated by the statistical processing based on the position data of each shot area, the exposure is performed. The yield of chips from 6) is increased.

Even if the alignment system is, for example, the TTL system or the off-axis system, the selection sequence of the sample shots can be simplified by, for example, using a shot area in which no alignment mark is missing as a sample shot. . If the shot area that cannot be aligned in the first step is a shot area that is estimated to have a missing alignment mark in advance, it is possible to shorten the alignment time by excluding it from the sample shots in advance. .

Further, when the shot area which cannot be aligned in the first step is a shot area which is preliminarily estimated from the attached alignment mark such that a predetermined measurement signal cannot be obtained by the alignment system. , It is possible to more accurately determine the lack or defect of the alignment mark.

[0016]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of the exposure method according to the present invention will be described below with reference to the drawings. In the present embodiment, the present invention is applied to a stepper type projection exposure apparatus in which two-beam heterodyne interference type TTR type alignment sensor is used to perform EGA type alignment exposure. The two-beam interference method is based on LIA (Laser In
It is also called the terferometric alignment method.

FIG. 2 shows an overall schematic configuration of the projection exposure apparatus of this example. In FIG. 1, the exposure illumination light of wavelength λ ₀ from the exposure illumination system 60 is reflected by the dichroic mirror 3 during exposure. Is irradiated onto the reticle 4, and under the illumination light, the pattern of the reticle 4 is reduced by, for example, 1/5 through the projection optical system 5 and projected onto each shot area on the wafer 6 coated with the photoresist. It As the illumination light for exposure, in this example, the i-line (wavelength: 36
5 nm), but other than that, excimer laser light (wavelength: 248 nm, 193 nm, etc.) can also be used.
Here, the Z axis is taken parallel to the optical axis AX of the projection optical system 5,
The plane perpendicular to the Z axis is parallel to the paper surface of FIG. 2 along the Y axis, and the X axis is perpendicular to the paper surface of FIG.

In this case, the reticle 4 is held on the reticle stage 9, and the reticle stage 9 moves in the X direction, the Y direction, and the rotation direction (θ direction) in a plane perpendicular to the optical axis AX of the projection optical system 5. Position 4 The X-coordinate, the Y-coordinate, and the rotation angle of the reticle stage 9 are constantly measured by the movable mirror 63m fixed on the reticle stage 9 and the laser interferometer 63 installed outside, and the measured value is the reticle stage control system 64. Further, it is supplied to an alignment signal processing system 68 which will be described later, and the measurement value thereof is also supplied to a central control system 61 which controls the operation of the entire apparatus through a reticle stage control system 64. Central control system 61
However, the reticle stage 9 is added to the reticle stage control system 64.
When the information on the target position and the target rotation angle is supplied, the reticle stage control system 64 sets the position and the rotation angle of the reticle stage 9 to the target values via the reticle stage drive unit 62 accordingly.

On the other hand, the wafer 6 is placed on a wafer stage composed of an X stage 8X, a Y stage 8Y and the like via a wafer holder 7. Actually, X stage 8X
A Z stage or the like for positioning the wafer 6 in the Z direction is also mounted on the top. The wafer stage positions the wafer 6 in the X direction, the Y direction, and the rotation direction (θ direction) within a plane perpendicular to the optical axis AX of the projection optical system 5. X stage 8
The X coordinate, the Y coordinate, and the rotation angle of the wafer 6 are constantly measured by the movable mirror 66m fixed on X and the laser interferometer 66 installed outside, and the measured values are supplied to the wafer stage control system 67. The measured value is also supplied to the central control system 61 via the wafer stage control system 67. When the central control system 61 supplies the information on the target position and the target rotation angle of the wafer 6 to the wafer stage control system 67, the wafer stage control system 67 responds to the information on the target stage and the target rotation angle of the wafer 6 via the wafer stage drive unit 65. Control the corners.

Further, the X stage 8X on the wafer stage side
The reference mark member 11 is fixed near the upper wafer 6,
A reference mark is formed on the reference mark member 11 as a reference for alignment of the reticle 4 with respect to the optical axis AX of the projection optical system 5. Correspondingly, two reticle alignment microscopes 39 and 40 (see FIG. 2) are provided above the peripheral portion of the reticle 4 for detecting the amount of positional deviation between the alignment mark on the reticle 4 and its reference mark. It is arranged. The detection results of these reticle alignment microscopes 39 and 40 are supplied to the central control system 61.

Next, the LIA type alignment sensor of this example will be described in detail. The alignment sensor includes an objective lens 2 above the dichroic mirror 3, an alignment optical system 1 above the objective lens 2, and an alignment signal processing system 68. When performing alignment, the laser beam emitted from the laser light source in the alignment optical system 1 undergoes predetermined frequency modulation and is emitted as alignment light. As alignment light,
Light in a wavelength range in which the photoresist coated on the wafer 6 is weakly sensitive (for example, a laser beam having a wavelength of 633 nm from a He—Ne laser light source) is used.
The alignment light is transmitted through the objective lens 2 and the dichroic mirror 3 and is a reticle mark 35A in the form of a diffraction grating on the reticle 4, and a light transmissive window portion (reticle window) 37.
The alignment light that has been irradiated to A and transmitted through the reticle window 37A is irradiated to the wafer mark 48A attached to the shot area of the positioning target on the wafer 6. Here, the measurement direction of the reticle mark 35A and the wafer mark 48A is the X direction.

Then, the heterodyne beam generated by the diffraction at the wafer mark 48A, and the reticle mark 35.
The heterodyne beam generated by the diffraction at A returns to the alignment optical system 1 via the dichroic mirror 3 and the objective lens 2, and two light receiving systems in the alignment optical system 1 generate two beat signals. These beat signals are supplied to the alignment signal processing system 68, where the phase difference between the two beat signals is detected, and the detected phase difference is supplied to the central control system 61.

Next, referring to FIGS. 2 to 6, the LI of this example
An optical path of alignment light from the A-type alignment sensor and a method of detecting a reticle mark and a wafer mark will be described. In FIG. 2, the alignment optical system 1
From, a pair of reticle alignment illumination lights RB ₁ and RB _{2 having} an average wavelength λ ₁ different from the exposure wavelength λ ₀ and a frequency difference Δf (50 kHz in this example) and wafer alignment illumination lights WB ₁ and WB ₂ are emitted. It

FIG. 3 is a side view of FIG. 2 viewed in the Y direction. As shown in FIG. 3, the reticle alignment illumination lights RB ₁ and RB ₂ are focused on the reticle 4 by the objective lens 2. The diffraction grating-shaped reticle mark 35A on the lower surface of the reticle 4 is irradiated with incident angles −θ _R1 and θ _R1 . FIG. 5 shows the reticle mark 35A of the reticle 4 of this example.
6 is an enlarged view of the periphery of FIG. 5, and in FIG. 5, the reticle mark 35A for the X axis is a mark made of a diffraction grating formed at a pitch P _R in the X direction.
A reticle window 37A for passing alignment light traveling toward the wafer is formed inside A. And
The reticle mark 35A is irradiated with the light flux 50 composed of the illumination lights RB ₁ and RB ₂ , and the reticle window 37A is illuminated with the illumination light W.
A light beam 51 composed of B ₁ and WB ₂ passes through.

Returning to FIG. 3, the incident angles −θ _R1 , θ _R1 and the grating pitch P _R of the reticle mark 35A have the following relationship, and the + _{1st order} diffracted light RB ₁ ^{+1 of the} illumination light RB ₁ is obtained. And illumination light RB
₂ -1st-order diffracted light RB ₂ ^-1 And occur directly above,
The alignment detection light (heterodyne beam) is returned to the alignment optical system 1 via the objective lens 2. sin θ _R1 = λ ₁ / P _R On the other hand, the wafer alignment illumination lights WB ₁ and WB ₂ pass through the reticle window 37A of the reticle 4 and reach the chromatic aberration control plate 10 in the projection optical system 5. Diffraction grating-like on-axis chromatic aberration control elements are formed in the portions of the chromatic aberration control plate 10 through which the illumination lights WB ₁ and WB ₂ pass, respectively.
_1, WB ₂ each angle - [theta] _G1, bent only theta _G1, respectively incident angle - [theta] _W1, to the diffraction grating-shaped wafer mark 48A is irradiated with theta _W1.

FIG. 6 is an enlarged view of the wafer mark 48A. In FIG. 6, the wafer mark 48A is composed of a concave and convex diffraction grating formed at a pitch P _W in the X direction. Then, the wafer mark 48A is irradiated with the light flux 51 composed of the illumination lights WB ₁ and WB ₂ . Returning to FIG. 3, the incident angles −θ _W1 , θ _W1 and the grating pitch P _W of the wafer mark 48A have the following relationship, and the + _{1st order} diffracted light WB of the illumination light WB ₁ is obtained.
₁ ⁺¹ And the −1st order diffracted light WB ₂ ^{−1 of the} illumination light WB ₂ Are generated directly above, and these two diffracted lights become alignment detection light (heterodyne beam).

Sin θ _W1 = λ ₁ / P _{W In} this case, as shown in FIG. 2, the wafer alignment illumination light is angled with respect to the wafer 6 in the non-measurement direction (Y direction) by the deflection action of the chromatic aberration control plate 10. Since the light is incident at an angle of θ _m , the position at which each of the alignment detection lights passes on the chromatic aberration control plate 10 is different from the position at which it is passed at the time of incidence. The alignment detection light from the wafer mark 48A passes through another axial chromatic aberration control element on the chromatic aberration control plate 10 so that the lateral chromatic aberration is corrected and goes to the reticle window 37A. After that, each detection light is transmitted through the reticle window 37.
It returns to the alignment optical system 1 again via A and the objective lens 2. Further, the wafer alignment illumination light illuminates a position shifted by Δβ in the Y direction on the surface of the wafer 6 as compared with the case where the chromatic aberration control plate 10 is not arranged.

The alignment optical system 1 will be described in detail with reference to FIG. FIG. 4A is a view of the alignment optical system 1 viewed from the same direction as FIG. 3, and FIG.
FIG. 4 is a bottom view of FIG. In FIG. 4A, H
Laser beam B emitted from the e-Ne laser light source 12
Is divided into two by the half prism 13, and each has a frequency F
Acousto-optic device driven by ₁ and F ₂ (hereinafter referred to as “AO
It is incident on 15, 16. Frequencies F ₁ and F
₂ is several tens of MHz, and the difference between both frequencies is 50 kHz. Of the diffracted lights emitted from the AOMs 15 and 16, + 1st order diffracted lights B ₂ and B, respectively.
Only ₁ is extracted by the slit plates 17 and 18. The light beam transmitted through the half prism 22 within one extracted the +1 order diffracted light B _2, and the light beam reflected by the other of the + half prism 22 among the diffracted light B _1, in parallel with some distance condenser Head to lens 23.

Then, the two light fluxes condensed by the condenser lens 23 are shaped by the field stop 24, and the pair of light fluxes passing through the field stop 24 is directed by the prism 25 in a direction perpendicular to the paper surface of FIG. 4 (a). And is divided into _two to generate a pair of reticle alignment illumination lights RB ₁ and RB ₂ and a pair of wafer alignment illumination lights WB ₁ and WB ₂ . These two pairs of alignment illumination light are reflected by the lens 26 and the half prism 2.
It goes toward the objective lens 2 of FIG.

On the other hand, the reticle mark 35A shown in FIG.
The alignment detection light from the emission mark 48A is shown in FIG.
After returning to the alignment optical system 1 in (a),
Is reflected by the prism 27, passes through the lens 28, and the reticle,
And the detection light separation prism 29 at a position conjugate with the wafer
Is separated into reticle detection light and wafer detection light by
It As shown in FIG. 4B, the detection light separation prism 29
Is a part that reflects the wafer detection light and transmits the reticle detection light
A reticle detection light RB₁ ⁺¹ , RB₂
^-1 Is transmitted through the detection light separation prism 29, and is a photoelectric detection element.
Received by 30. Then, the wafer detection light WB₁
⁺¹ , WB₂ ^-1 Is reflected by the detection light separation prism 29
Is received by the photoelectric detection element 31 of FIG.
It Corresponds to the position of the reticle mark from the photoelectric detection element 30
Reticle beat signal S_RIs output, the photoelectric detection element
Wafer beat signal corresponding to the position of the wafer mark from 31
No. S_WIs output.

FIG. 11 shows an example of the reticle beat signal S _R and the wafer beat signal S _W. The reticle beat signal S _R and the wafer beat signal S _W each have a frequency Δ.
This is a sinusoidal beat signal of f (= 50 kHz), and the phase difference Δφ _T [rad] between them changes depending on the relative movement amount of the reticle 4 and the wafer 6 in the X direction, and the relative movement amount Δx is It is as shown in the formula.

[0032] [Delta] x (on the _{reticle) = P R · Δφ T /} (4π), Δx ( on _{wafer) = P W · Δφ T /} (4π) Note that the pitch P _R of the reticle marks, the projection optical system of FIG. 3 The product of the projection magnification (reduction magnification) of 5 is the wafer mark pitch P _W. Reticle beat signal S _R
And the wafer beat signal S _W are respectively supplied to the alignment signal processing system 68 of FIG. 2, where both signals S _R and S _W are supplied.
Is detected. The detected phase difference is supplied to the central control system 61. For example, when performing die-by-die type alignment, the central control system 61 determines that the positional deviation between the reticle mark and the wafer mark is a predetermined target drive-in value based on the phase difference between both beat signals S _R and S _W. The position of the reticle stage 9 is adjusted via the reticle stage control system 64 so that In addition, in addition, a 3-axis LIA type alignment optical system is provided in addition to the above, and the phase difference between the two beat signals from these 3-axis alignment optical systems can be controlled so as to reach predetermined target drive-in values. Done. After that, the pattern image of the reticle 4 is projected and exposed on the shot area of the wafer 6.

Next, an example of alignment and exposure operations in the projection exposure apparatus of this example will be described with reference to the flow chart of FIG. 1 and FIGS. In this example, all shot areas in the design of the wafer 6 are used as sample shots to perform positioning by using an LIA alignment sensor, and position data necessary for EGA alignment for shot areas that can be positioned by the alignment sensor. In addition to obtaining the shot areas, the shot areas are exposed, and the unexposed shot areas for which an alignment signal has not been obtained due to lack of a wafer mark or the like are positioned and exposed based on the array coordinates calculated by the EGA method. . Hereinafter, a specific description will be given.

First, FIG. 7 shows a reticle 4 used in this example.
The pattern arrangement of the reticle 4 is shown in FIG.
A frame-shaped light-shielding band 33 on the lower surface of the area becomes a pattern area 32, and a circuit pattern for transfer is drawn in the pattern area 32. Then, in the rectangular pattern region 32, a pair of reticle marks 35A and 35B in the form of diffraction gratings arranged in the X direction at a predetermined pitch in the vicinity of each side, and a pair of diffraction gratings arranged in the Y direction at a predetermined pitch. A pair of reticle marks 36A, 36B are formed, and reticle windows 37A, 37B, 38A, 38B are formed inside these reticle marks 35A, 35B, 36A, 36B, respectively. Further, cross-shaped alignment marks 34A and 34B are formed so as to sandwich the light shielding band 33 in the X direction.

Correspondingly, FIG. 8 shows an enlarged view of a part of the shot area on the wafer 6, and in FIG.
Inside the central rectangular shot area 47, X near each side
A pair of wafer marks 48A, 48B in the form of diffraction gratings arranged at a predetermined pitch in the Y direction, and a pair of wafer marks 49A, 49 in the form of diffraction gratings arranged at the predetermined pitch in the Y direction.
B is formed. Similarly, also in the other shot areas, a pair of X-axis wafer marks and a pair of Y-axis wafer marks, respectively.
A shaft wafer mark is formed. Note that these wafer marks may be formed on the street line areas between the shot areas.

In this case, when the shot area 47 is a sample shot to be measured, the amount of positional deviation between the wafer mark 48A and the corresponding reticle mark 35A in FIG. 7 is shown in FIG.
Of the other three alignment optics (not shown) whose positional deviations between the other three wafer marks 48B, 49A, 49B and the corresponding reticle marks 35B, 36A, 36B of FIG. The detection result is supplied to the central control system 61 shown in FIG. That is, the degree of freedom of the measured position shift amount is four. On the other hand, since the degree of freedom of relative movement between the wafer 6 and the reticle 4 is only 3 (X direction, Y direction, θ direction), for example, X
The average value of the positional deviation amounts of the two sets of X-axis marks is used as the positional deviation amount in the direction, and the average value of the positional deviation amounts of the two sets of Y-axis marks is used as the positional deviation amount in the Y direction. . For the deviation amount in the θ direction, the average value of the positional deviation amounts of the four sets of marks on the X axis and the Y axis is used. This provides an averaging effect.

FIG. 9 shows the pattern arrangement of the reference mark member 11 on the wafer stage of this example.
On the reference mark member 11 made of a glass substrate, frame-shaped reference marks 41A and 41B are arranged at predetermined intervals in the X direction.
Are formed. The distance between the reference marks 41A and 41B is set to the alignment mark 34 on the reticle 4 shown in FIG.
The distance between A and 34B is set to be the same as the distance obtained by multiplying the projection magnification of the projection optical system 5, and the reference mark 41
A and 41B are illuminated from the bottom surface side to the projection optical system 5 side with illumination light in the same wavelength range as the illumination light for exposure. Also,
Between the reference marks 41A and 41B, the X-axis reference diffraction grating mark 42A, which serves as a reference for positioning the wafer mark,
42B and the reference diffraction grating marks 43A and 43B for the Y axis
Are formed.

Then, in step 101 of FIG.
Positioning (reticle alignment) of the reticle 4 of FIG. 2 with respect to the projection optical system 5 is performed. Specifically, the central control system 6
In No. 1, the wafer stage is driven based on the data obtained in advance, and the midpoint of the reference marks 41A and 41B of the reference mark member 11 in FIG. After that, the central control system 61 takes in image pickup signals from the two reticle alignment microscopes 39 and 40 in FIG.

FIG. 10 shows the observation fields 39a, 40a of the reticle alignment microscopes 39, 40. In FIG. 10, in one observation field 39a, the image 41AR of the reference mark 41A on the reticle and the alignment on the reticle side are shown. The mark 34A is observed, and the image 41BR of the reference mark 41B and the alignment mark 34B are observed in the other observation visual field 40a. Therefore, the central control system 61 of FIG. 2 finely adjusts the position of the reticle stage 9 via the reticle stage control system 64, so that the alignment marks 34A and 34 in FIG.
Centers of B are reference mark images 41AR and 41B, respectively.
The reticle 4 is positioned in the X direction, Y direction, and θ direction so as to match the center of R. Note that, for example, when the distance between the alignment marks 34A and 34B deviates from the design value, the centers of the alignment marks 34A and 34B may be symmetrically displaced from the centers of the reference mark images 41AR and 41BR, respectively. .

With the reticle alignment completed in this way, the alignment optical system 1 of the LIA type alignment sensor shown in FIG. 2 and the other three alignment optical systems are further operated to set four sets of reticle beat signals and The wafer beat signal is taken into the alignment signal processing system 68, and the phase difference Δφ _{1 to} Δφ of the four beat signals at this time
Alignment signal processing system 68 with ₄ as the target driving-in value
It is stored in the internal storage unit. In this case, the phase difference Δφ ₁ between the two beat signals output from the alignment optical system ₁ is
Reference diffraction grating mark 42 on the reference mark member 11 of FIG.
This corresponds to the amount of positional deviation in the X direction between A and the reticle mark 35A on the reticle 4 in FIG. Similarly, other phase differences Δ
φ _{2 to} Δφ ₄ are the reference diffraction grating marks 42B, 43A,
43B and reticle marks 35B and 36 corresponding to 43B, respectively.
This corresponds to the amount of positional deviation from A and 36B.

Next, in step 102, the reticle is
Initialize the position of reticle stage 9 after alignment is completed
The position is recorded in the storage unit in the alignment signal processing system 68.
Remember The initial position is measured by the laser interferometer 63.
X coordinate x₀, Y coordinate y ₀, And rotation angle θ₀More
Therefore, its initial position is (x₀, Y₀, Θ₀).

Next, based on the exposure data file for the wafer 6, all shot areas recorded in the exposure data file for the wafer 6 are selected as sample shots. Here, the number of all shot areas is M. At this time, the center of all shot areas on the wafer 6 and the designed array coordinates of each wafer mark in the coordinate system (sample coordinate system) on the wafer 6 are stored in the storage unit in the central control system 61. Has been done. In the following, nth (n = 1 to M)
The designed array coordinate of the center of the sample shot of is set to (X _n , Y _n ). Further, in FIG. 2, rough alignment (pre-alignment) of the wafer 6 with respect to the wafer stage has already been executed, and the stage coordinate system defined by the measurement value of the laser interferometer 66 on the wafer stage side is calculated from the coordinates on the sample coordinate system. It is assumed that the rough value of the conversion coefficient for obtaining the upper coordinates has been obtained.

Then, the sample shots are aligned by a die-by-die method, and exposure light is irradiated to expose the pattern on the reticle 4 to the shot area of the wafer 6. That is, the central control system 61 initializes an integer n indicating the order of sample shots to 1 in step 103, and then shifts to step 104, where n
Design array coordinates (X
_n , Y _n ) to array coordinates in the stage coordinate system.
Then, based on the array coordinates after this conversion, the X stage 8X and the Y stage 8 are passed through the wafer stage control system 67.
The stepping drive of Y causes the center of the nth sample shot to be the center of the exposure field (exposure position).
Set to. However, in reality, the center of the sample shot is deviated from the exposure position due to expansion and contraction of the wafer 6, rotation error, and the like. This amount of positional deviation is an alignment error, and in this example, this alignment error is corrected by moving the reticle stage 9 side. Here, for convenience of explanation, the n-th sample shot is taken as the shot area 4 in FIG.
Assume that it is 7.

Therefore, in step 105, the central control system 61 activates the four alignment sensors of the LIA system by the TTR system, and the four reticle marks 35A, 35B, 36A, 36B of FIG. 8 n
Four wafer marks 48 of the fourth sample shot 47
The phase differences Δφ _A , Δφ _B , Δφ _C , and Δφ _D of the beat signals corresponding to the positional shift amounts from A, 48B, 49A, and 49B are measured. In the alignment signal processing system 68, the phase error obtained by subtracting the target drive-in values Δφ _{1 to} Δφ ₄ obtained in step 101 from the measured phase differences Δφ _{A to} Δφ _D is supplied to the central control system 61. In response to this, in the central control system 61, for example, the absolute value of the average value of the two X-direction phase errors among those phase errors becomes equal to or smaller than a predetermined first threshold value, and the average of the two Y-direction phase errors is calculated. The reticle stage 9 is moved through the reticle stage control system 64 in the X direction so that the absolute value of the value is less than or equal to the first threshold value and the sum of the absolute values of the four phase errors is less than or equal to the predetermined second threshold value. , Y direction, and θ direction. A state in which the positioning is performed in this way is called an alignment completed state.

In this step 105, a wafer mark chipping or the like occurs due to the process processing in the previous process,
There are cases where a predetermined alignment signal is not output from the alignment sensor or the signal strength is extremely weak even if there is an alignment signal. For example, in FIG. 8, since the wafer mark corresponding to the wafer mark 48B is missing in the shot area 71 on the right side, the wafer mark 48
The alignment signal (beat signal) of the wafer mark corresponding to B is only a noise signal. Therefore, if the signal intensity of the wafer beat signal of the alignment sensor is equal to or higher than a predetermined threshold value in the next step 106, step 1
If the signal intensity is not sufficient, the process proceeds to step 108 to measure the next shot area without performing positioning and exposure. In general, the shot areas around the wafer often have missing wafer marks (hereinafter, the shot areas where these wafer marks are missing are referred to as “missing shots”). Therefore, the peripheral shot areas having many missing shots may be excluded from the sample shots in advance. The states of the wafer marks on the wafer 6 are all recorded in the exposure data file.

Then, the wafer bee of the alignment sensor
If the strength of the signal is sufficient, go to step 107.
Command of the central control system 61 when the alignment is completed.
The exposure illumination light is emitted from the exposure illumination system 60 of FIG.
The circuit pattern on the reticle 4 to the shot area 47.
Project and expose. Then, the alignment signal processing system 6
8 shows whether the laser interferometer 63 on the reticle side during this exposure
These measurements are averaged to obtain the X coordinate of the reticle stage 9
Mean x_n, Y-coordinate average value y_n, Average rotation angle θ
_nAnd calculate these values as the average position of the reticle stage 9.
(X_n, Y_n, Θ _n) Is stored as.

In the following step 108, the value of the integer n is incremented by 1, and then the process proceeds to step 109 to determine whether or not the measurement has been completed for M sample shots. Therefore, steps 104 to 108 are repeated M times, and the average position (x) of the reticle stage 9 in the alignment completed state for each of the K sample shots for which a sufficient alignment signal strength of the M sample shots is obtained. ₁ , y ₁ , θ ₁ ) to (x _K , y _K , θ _K ) are calculated and stored. In this way, when the average position of the reticle stage 9 is obtained for the K sample shots for which the measurable alignment signal is obtained among all (M) sample shots on the wafer 6, the operation starts from step 109. Move to 110.

Then, the alignment signal processing system 68 is
With respect to the central control system 61, for each of the K sample shots, the movement amount (x _n −x ₀ , y _n −y from the initial position of the reticle stage 9 in the alignment completed state).
₀ , θ _n −θ ₀ ) (n = 1 to K) is supplied as alignment data. The central control system 61 calculates coordinate conversion parameters (hereinafter referred to as “EGA parameters”) when performing alignment by the EGA method using the supplied alignment data. In addition, the E
The GA parameter may be calculated.

Here, an example of the EGA parameter and its
An example of the calculation method of will be described. First, in FIG.
Let the orthogonal coordinate system (sample coordinate system) on the wafer 6 be (x, y).
Then, the shot area 47 is set as the nth sample shot.
I do. The seat of the reference point 47a at the center of this sample shot
Design coordinate values in the standard (x, y) are (C_Xn,
C _Yn). This coordinate value (C_Xn,
C_Yn) Is the nth sample used in step 104
Leshot design array coordinates (X_n, Y_nSame as
is there. Here, the reference point 4 is parallel to the sample coordinate system (x, y).
The Cartesian coordinate system with the origin at 7a is (α, β), and
Coordinate system of Hammark 48A, 48B, 49A, 49B
Design coordinates on (α, β) are respectively
(S_1Xn, S_1Yn), (S_2Xn, S_2Yn), (S_3Xn,
S_3Yn), (S_4Xn, S_4Yn).
The wafer marks 48A, 48B, 49A, 4 shown in FIG.
9B is a one-dimensional mark, so the non-measurement direction of each mark
The coordinates at are used when finding the EGA parameters.
No need. At this time, Nth (here, N = 1 to 4)
Of the wafer mark in the sample coordinate system (x, y)
Column coordinates (D_NXn, D_NYn), The following relation holds.
are doing.

[0050]

[Equation 1]

Next, the following six EGA parameters are used as coordinate conversion parameters from design coordinates on the sample coordinate system (x, y) to coordinates on the stage coordinate system (X, Y). Wafer residual rotation error Θ: This is represented by the rotation of the sample coordinate system (x, y) with respect to the stage coordinate system (X, Y). Orthogonality W of stage coordinate system (X, Y): This is mainly X
This occurs because the wafer stage feeds in the axial and Y-axis directions are not exactly orthogonal. Scaling (linear expansion / contraction) Rx in the x direction and scaling R in the y direction in the wafer coordinate system (x, y)
y: This occurs because the wafer 6 is entirely expanded and contracted due to the processing process and the like. For example scaling R
x is the ratio between the measured value and the design value of the distance between two points on the wafer 6 in the x direction. Stage coordinate system (X,
Offsets Ox and Oy with respect to Y): This occurs because the wafer 6 is entirely displaced from the wafer stage by a minute amount.

Further, there is a factor of overlay error in each shot area. Considering these error factors excluding the offset component, the following is used to obtain the coordinates of the coordinate system (x, y) on the wafer from the design coordinates on the coordinate system (α, β) in FIG. 4 coordinate conversion parameters (also referred to as “EGA parameters”). Chip rotation θ: This is a rotation error of each shot area with respect to the coordinate system (x, y) on the wafer. Chip orthogonality w: This is an error in the orthogonality of the chip pattern caused by distortion of the pattern itself on the reticle, distortion of the projection optical system, or the like. Chip scaling rx, ry in the x and y directions:
This is a linear expansion / contraction of the chip pattern in each shot area caused by, for example, an error in projection magnification.

The above-mentioned 10 EGA parameters (1) to (4), the design coordinate values (C _Xn , C _Yn ) of the reference point 47a of the nth sample shot, and the wafer marks 48A, 4A.
8B, 49A, 49B design coordinates (S _NXn ,
S _NYn ) (N = 1 to 4), the wafer mark 48
A, 48B, 49A, 49B stage coordinate system (X,
Y) The coordinates of the position that should actually be on (F _NXn , F _NYn ).
(N = 1 to 4) is expressed as follows. However, A and B
Is a matrix of 2 rows × 2 columns.

[0054]

[Equation 2]

In this case, when the residual rotation error Θ of the wafer, the orthogonality W, the chip rotation θ, and the chip orthogonality w are assumed to be minute amounts, the approximation calculation is performed to obtain the matrix A.
And B are expressed as follows.

[0056]

(Equation 3)

Further, in order to facilitate the application of the least squares method described later, those matrices A and B may be approximated by the following matrices A'and B '. In this case, the wafer scalings Rx and Ry are calculated using new parameters Γx and Γy, respectively, and Rx = 1 + Γx, and Ry =
It is represented by 1 + Γy. Similarly, chip scaling rx and ry are represented by rx = 1 + γx and ry = 1 + γy using new parameters γx and γy, respectively. Then, assuming that the parameters Γx, Γy, γx, γy are minute amounts, the matrices A and B of (Equation 3) can be approximated by the following matrices A ′ and B ′.

[0058]

[Equation 4]

Next, assume that matrices A'and B'are used.
Then, the 10 EGA patterns included in the coordinate conversion formula of (Equation 2) are
Parameters (Θ, W, Γx (= Rx-1), Γy, Ox,
Oy, θ, w, γx (= rx−1), γy) is the least square
Find by law. Specifically, the nth sample shot
Coordinate values (C_Xn, C_Yn),as well as
Design of wafer marks 48A, 48B, 49A, 49B
Coordinates of (S_NXn, S_NYn) (N = 1 to 4) to (Equation 2)
Coordinates calculated by substituting (F _NXn, F_NYn)age
Is measured on the stage coordinate system (X, Y).
Ehamark 48A, 48B, 49A, 49B coordinate values
(FM_NXn, FM_NYn). And K sumps
For each wafer mark in
Coordinate values (FM_NXn, FM_NYn) And its calculated coordinate values
(F_NXn, F_NYn) Amount obtained by calculating the sum of squares of the difference between
Is the residual error component ΔE as follows.

[0060]

(Equation 5)

The residual error component ΔE is
Partially differentiated by EGA parameter and the value is 0
Make an equation such that
By solving the equation, it is possible to obtain 10 EGA parameters.
Wear. This is the EGA calculation. However, the steps of this example
At 110, the alignment of the nth sample shot
Movement from the initial position of the reticle stage 9 as data
Quantity (x_n-X₀, y_n-Y ₀, θ_n−θ₀) (N = 1 to K)
The measurement data for each wafer mark is supplied.
Not in. In this case, first, K times from the initial position.
Rolling error (θ_n−θ ₀) The average value of
, The other chip scaling rx, ry,
And the value of the chip orthogonality w is unknown. Furthermore, (Equation 2)
Then, the matrix B (more precisely, the matrix B ') is set to 0, and the n-th
Designed coordinate value (C of the reference point 47a of the sample shot)
_Xn, C_Yn) Is the calculated coordinate value
(F_{NX n}, F_NYn). Also, the projection of the projection optical system 5
The conversion factor obtained by pre-alignment, with the magnification being ζ
Using the coordinate values (C_Xn, C_Yn) Is obtained by converting
Roughly calculated coordinate values, that is, the position in step 104
The nth sun on the stage coordinate system, which is the target of placement
Set the pull shot coordinate value to (X '_n, Y '_n).

In this case, the coordinate values (FM _NXn , FM _NYn ) actually measured on the stage coordinate system (X, Y) of the n-th sample shot are (X ' _n -ζ (x _n -x ₀ ) 、
Y ′ _n −ζ y _n −y ₀ )). Therefore, the residual error component ΔE of ( _Equation 5) represented by these actually measured coordinate values (FM _NXn , FM _NYn ) and the calculated coordinate values (F _NXn , F _NYn ) should be minimized. , The four EGA parameters of the matrix A ′ (residual rotation error Θ of wafer, orthogonality W, scaling Γx (= Rx−1), Γy) and the values of the remaining two offsets Ox and Oy are determined. Good.

Next, in step 111, the central control system 61 uses the matrix A'containing the EGA parameters obtained in step 110 and the offsets Ox, Oy to calculate all (M By sequentially substituting the designed array coordinate values (C _Xn , C _Yn ) of the reference points of the shot areas in (4), the calculated coordinates of the reference points of the shot areas on the stage coordinate system (X, Y) are calculated. Array coordinate values (G _Xn , G _Yn )
Ask for.

[0064]

(Equation 6)

Further, since the chip rotation θ is also obtained in this example, the reticle stage 9 is rotated according to the sum (θ + θ) of the residual rotation error Θ of the wafer in the matrix A ′ and the chip rotation θ. Although the chip orthogonality w and the chip scaling rx and ry are not obtained in this example, if the chip orthogonality w is obtained, the stepper type (batch exposure type) projection exposure is performed as in the present example. The device cannot correct in a strict sense. However, when using a scanning exposure type projection exposure apparatus that performs exposure by scanning the reticle and the wafer relative to the projection optical system, for example, the reticle scanning direction and the wafer scanning direction are shifted during exposure. Thus, the chip orthogonality w can be corrected.

Further, it is assumed that chip scaling rx, ry
, The chip scaling rx,
It is desirable to correct the projection magnification of the projection optical system 5 in accordance with ry. Also in this case, when the scanning exposure type projection exposure apparatus is used, the chip scaling in the scanning direction can be corrected by adjusting the scanning speed ratio between the reticle and the wafer during exposure.

Then, in step 112, based on the array coordinates (G _Xn , G _Yn ) obtained by the calculation of ( _Equation 6), the remaining (M−K) unexposed shot areas on the wafer 6 are formed. The reference point is sequentially aligned with the center of the exposure field of the projection optical system 5 in FIG. 2, and the pattern image of the reticle 4 is projected and exposed on the shot area. Then, after the exposure of all shot areas on the wafer 6 is completed, processing such as development of the wafer 6 is performed.

As described above, according to the present example, when the EGA type alignment is performed using the TTR type alignment sensor, the sequence for performing the alignment and exposure by the die-by-die type can be used as it is. Furthermore, when measuring the amount of displacement of each sample shot, the amount of movement of the reticle stage 9 is only measured with the reticle mark and the wafer mark being aligned, so the measurement time is shortened. As a result, the throughput of the exposure process is improved. Further, since a measurable alignment signal is obtained and all the shot areas that can be exposed are exposed, the overlay accuracy is improved. Further, the position of the shot area is observed during the exposure time, and the position of the shot area for which an alignment signal cannot be obtained due to a missing shot or other causes is estimated based on the position and the exposure is performed. And the time of the exposure process is shortened. In addition, on the contrary, accurate alignment is performed even for a shot area where the wafer mark is missing.
Therefore, the yield of chips from the wafer is improved.

[0069] In the above embodiment, may not necessarily be determined rotation angle theta _n when it is seeking the rotation angle theta _n of the reticle stage 9 in step 107, for example, it is not necessary to correct the chip rotation. Further, in the EGA type alignment used in the above-mentioned embodiment, the contribution from each sample shot (or each wafer mark) is equal to the residual error component ΔE of (Equation 5). However, for example, when the wafer 6 is distorted around a predetermined reference point, it is desirable to give different weights to the respective terms of the residual error component ΔE of (Equation 5) depending on the position of the sample shot. . The method of giving a weight to each sample shot (or wafer mark) in this way is called a weighted EGA method, but the present invention can also be applied to the case of performing alignment by the weighted EGA method.

Next, a modification of the above embodiment will be described. This modified example is particularly effective when the positioning accuracy of the wafer stage in the projection exposure apparatus of FIG. 2 is insufficient. When the positioning accuracy of the wafer stage is insufficient as described above, the position of the wafer stage may deviate from the target position in the alignment completed state in step 107 of FIG. Just measuring is not enough. Therefore,
In the process corresponding to step 107 of FIG. 1 of this modification, in addition to the average position (x _n , y _n , θ _n ) of the reticle stage 9 in the alignment completed state, the measurement value by the laser interferometer 66 on the wafer side is used. Based on the wafer stage X
The average value (ΔX _n , ΔY _n , Δφ _n ) of the deviation amount from the target position in the direction, the Y direction, and the rotation direction is measured. Since the projection magnification of the projection optical system 5 from the reticle to the wafer is ζ, the average position of the reticle stage 9 is corrected by the following equation.

[0071]

(7) (x _n ', y _n ', θ _n ') = (x _n , y _n , θ _n )
− (1 / ζ) (ΔX _n , ΔY _n , Δφ _n ) Then, EGA alignment is performed using the corrected average position (x _n ′, y _n ′, θ _n ′). According to this modification, since the movement amount of the wafer stage is also measured, the alignment is performed with higher accuracy.

The alignment signal processing system 68 of the above embodiment outputs the phase difference between the two beat signals corresponding to the relative positional deviation amount between the reticle mark and the wafer mark. Even if the relative displacement amount with the mark cannot be quantitatively measured, only the sign of the relative displacement amount and the information indicating whether the absolute value of the relative displacement amount is smaller than a predetermined allowable value are output. It is also possible to use an alignment sensor. Also in this case, in step 107 of FIG.
The alignment is performed in the same manner as in the above-described embodiment by obtaining the average position of the reticle stage 9 while the information indicating that the absolute value of the relative positional deviation amount is smaller than the predetermined allowable value is output from the alignment sensor. be able to.

Further, although the LIA type alignment sensor is used as the alignment sensor in the above-described embodiment, for example, the images of the reticle mark and the wafer mark are taken under a predetermined illumination light, and the picked-up image data is processed. Even when an alignment sensor such as an image processing method (FIA method) that measures the amount of positional deviation is used, by applying the present invention, the same effect as that of the above-described embodiment can be obtained.

Further, even when using a TTL or off-axis type alignment sensor, for example, by using the sample shot excluding the shot region where the wafer mark is missing, the EGA-direction alignment is performed, Selection of sample shot becomes easy. As described above, the present invention is not limited to the above-described embodiments, and various configurations can be taken without departing from the gist of the present invention.

[0075]

According to the exposure method of the present invention, in the first step, for example, a predetermined shot area (sample shot) in the shot area that can be aligned with the mask pattern using a TTR type alignment sensor is used. In the state where the alignment is performed and the exposure is performed, for example, the movement amount of at least one of the mask and the substrate with respect to the target position on the data is measured. Then, by considering this movement amount as alignment data, alignment of the unexposed shot area on the substrate is performed in the second step. Therefore, for example, when the alignment data of the shot area that can be aligned is used to perform the EGA alignment, the control sequence when performing the alignment by the die-by-die system can be used, and the mask and Since it is only necessary to measure and store the movement amount of at least one of the substrates, there is an advantage that the time required for the measurement can be shortened as compared with the conventional case. Further, since the shot area that cannot be aligned is excluded from the sample shot, the sample shot can be easily selected.

Further, in the first step, since the positional deviation between the sample shots of the mask and the substrate is detected and the sample shots are exposed, the precision of overlaying the sample shots with the mask becomes extremely high. In addition, when the accurate position of the shot area is measured using the exposure time and the array coordinates of the unexposed shot area are calculated by statistical processing based on the position data of each shot area, for example, a wafer mark It is also possible to accurately align a shot area lacking in a chip, which has an advantage of increasing the yield of chips from the substrate.

If the shot area in which the alignment cannot be performed in the first step is the shot area in which the alignment mark is estimated to be missing in advance, the shot area may be excluded from the sample shot area in advance. Time is shortened. Further, when the shot area that cannot be aligned is, for example, the shot area determined by the signal intensity of the detection signal, that is, the shot area estimated from the attached alignment mark, the actual measurement result is used, and thus the accurate position is determined. The shot area that cannot be aligned can be specified.

[Brief description of drawings]

FIG. 1 is a flowchart showing an embodiment of an exposure method according to the present invention.

FIG. 2 is a schematic configuration diagram showing an entire projection exposure apparatus used in an embodiment of the present invention.

3 is a side view of a stage system and an alignment optical system of the projection exposure apparatus of FIG. 2 as viewed in a Y direction.

4A is a configuration diagram showing the alignment optical system 1 in FIG. 3, and FIG. 4B is a bottom view of FIG. 4A.

FIG. 5 is an enlarged plan view showing a reticle mark 35A and a reticle window 37A formed on the reticle.

FIG. 6 is an enlarged plan view showing a wafer mark 48A attached to a shot area on the wafer.

FIG. 7 is a plan view showing a pattern arrangement of a reticle used in an example.

FIG. 8 is an enlarged plan view showing a part of a shot area on a wafer which is exposed in the embodiment.

9 is a fiducial mark member 11 on the wafer stage of FIG.
It is an enlarged plan view showing the upper pattern arrangement.

FIG. 10 is a diagram showing an observation visual field of a reticle alignment microscope.

FIG. 11 is a view of the LIA type alignment optical system of FIG.
Output reticle beat signal S _RAnd wafer beat signal
No. S_WIt is a waveform diagram showing an example.

[Explanation of symbols]

 1 Alignment Optical System 2 Objective Lens 3 Dichroic Mirror 4 Reticle 5 Projection Optical System 6 Wafer 8X X Stage 8Y Y Stage 9 Reticle Stage 30, 31 Photoelectric Detection Element 35A, 35B, 36A, 36B Reticle Mark 39, 40 Reticle Alignment Microscope 48A, 48B, 49A, 49B Wafer mark 61 Central control system 68 Alignment signal processing system

Claims (5)

[Claims] 1. An exposure method for exposing each of a plurality of shot areas on a substrate with a pattern image on a mask, wherein an alignment mark attached to each shot area of the plurality of shot areas on the substrate is formed. For each predetermined one of all shot areas except the shot area that cannot be aligned with the pattern of the mask by using the alignment marks attached to the shot area, the pattern of the shot area and the mask A first step of exposing the shot area with the pattern image by performing alignment with the pattern image, and the substrate based on the alignment data with the mask pattern obtained for each of the shot areas exposed in the first step. A second step of exposing each of the unexposed shot areas above with the pattern image; And an exposure method. 2. The exposure method according to claim 1, wherein the alignment data includes a shot area to be measured by using alignment marks attached to each shot area on the substrate and the mask. An exposure method, which is an amount of movement of at least one of the mask and the substrate with respect to a target position when aligning with a pattern. 3. The exposure method according to claim 1, wherein the second step includes a step of calculating array data of unexposed shot areas on the substrate based on the alignment data. An exposure method characterized by the above. 4. The exposure method according to claim 1, 2, or 3, wherein the shot area, which cannot be aligned with the pattern of the mask in the first step, has alignment marks in advance. An exposure method, which is an estimated shot area. 5. The exposure method according to claim 1, 2, or 3, wherein a shot region that cannot be aligned with the pattern of the mask in the first step is estimated in advance from an attached alignment mark. Exposure method, which is a shot area to be exposed.