JPH08162392A - Aligning method - Google Patents

Abstract

PURPOSE: To enhance a wafer aligning system in alignment accuracy without deteriorating it in throughput when an alignment operation of in-shot multipoint EGA type is carried out. CONSTITUTION: A first wafer of a lot is subjected to an aligning process of in-shot multipoint EGA type by the use of an LIA system and an FIA system, and an LIA parameter and an FIA parameter obtained through the above aligning process and a parameter difference between them are stored (step 103 to 105). When a second wafer and following ones are subjected to a light exposure process, they are aligned through a usual EGA method by the use of a LIA system (step 110), the scaling component of an EGA parameter obtained through the above method is corrected by the stored parameter difference, and the array coordinates of shot regions are calculated by the use of the stored in-shot parameters.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a mask pattern for sequentially exposing each shot area on a photosensitive substrate in an exposure apparatus used in a lithography process for manufacturing a semiconductor device, a liquid crystal display device or the like. The present invention relates to a suitable alignment method applied when positioning each shot area based on the array coordinates predicted using a statistical method.

[0002]

2. Description of the Related Art Generally, a semiconductor element is formed by superposing, for example, several to ten and several layers of circuit patterns on a wafer. When performing projection exposure, it is possible to perform highly accurate alignment between each shot area where a circuit pattern is already formed on the wafer and the pattern image of the reticle as a mask to be exposed, that is, the alignment between the wafer and the reticle. Need to do. The alignment apparatus for performing such alignment is roughly classified into an alignment sensor that detects the position of an alignment mark (wafer mark) attached to each shot area on the wafer and generates a photoelectric signal, and the alignment sensor that generates the photoelectric signal. The signal processing system includes a signal processing system for processing and obtaining the amount of deviation of the wafer mark from the original position, and a positioning mechanism for correcting the wafer or reticle position according to the obtained amount of deviation.

As the alignment sensor method, an alignment mark (reticle mark) on a reticle and a wafer mark are simultaneously observed (detected) via a projection optical system.
TTR (Through the Reticle) method and TTL (Through the Lens) method that detects only the wafer mark through the projection optical system without detecting the reticle mark, and the detection system that is separated from the projection optical system. There is an off-axis method in which only the wafer mark is detected via the.

Among these, in the TTR method or the TTL method, the wafer mark is detected through the projection optical system, and the projection optical system is designed to have the best chromatic aberration with respect to the exposure light. Therefore, the desirable light is a laser beam (monochromatic light) or a quasi-monochromatic light having a wavelength range similar to that of the exposure light (for example, a bright line spectrum of a mercury lamp such as g-line and i-line). Therefore, as a TTR or TTL alignment sensor, a dot row pattern wafer mark and a slit-shaped laser beam are relatively scanned to detect diffracted light generated in a predetermined direction. A laser step alignment (hereinafter referred to as “LSA”) method for detecting the position of a wafer mark, or a laser beam is irradiated from a plurality of directions on a diffraction grating wafer mark and the wafer mark is emitted in the same direction. A two-beam interference method for detecting the position of the wafer mark from the phase of the interference light of a plurality of diffracted light (hereinafter,
"LIA (Laser Interferometric Alignment) method"
The one using a laser beam as the detection light is mainly used. LSA method and LIA
The alignment sensor of the type is, for example, Japanese Patent Laid-Open No. 2-272.
This is disclosed in Japanese Patent No. 305.

On the other hand, in the off-axis method, since there is no limitation by the projection optical system, the illumination light of the wafer mark may be any light. Therefore, the above-mentioned LSA method or LIA method alignment sensor is also used. it can. Further, as an off-axis method, an image obtained by illuminating a wafer mark with illumination light (broadband light) having a predetermined bandwidth (for example, a width of about 200 nm) from a halogen lamp and capturing an image of the wafer mark An image processing method for obtaining the position of the wafer mark by image-processing the signal (hereinafter,
An alignment sensor of "FIA (Field Image Alignment) method" is also used.

Further, as a method of aligning each shot area of a wafer using these alignment devices, an enhanced global alignment (hereinafter, referred to as
An "EGA" type alignment method has been proposed (see, for example, Japanese Patent Laid-Open No. 61-44429).
In this EGA method, the position of a wafer mark attached to a shot area (sample shot) selected from a large number of shot areas on the wafer is detected, and the detection result is statistically processed to obtain each shot on the wafer. The coordinate position of the area is calculated, and the positioning of each shot area is performed based on the calculated coordinate position.

Further, not only the position of each shot area,
As a method for performing alignment while measuring the expansion and contraction of the chip pattern in each shot area and the rotation of the chip pattern and correcting the alignment, Japanese Patent Application Laid-Open No. 6-275496 discloses a method of performing wafer alignment at a plurality of measurement points within a sample shot. A method for measuring the one-dimensional or two-dimensional position of a mark (hereinafter referred to as "multi-point EGA method within shot")
Is also proposed. By thus taking into account the expansion (magnification) and rotation of the chip pattern within the shot area, the overlay accuracy within the shot area is also improved.

[0008]

Among the conventional alignment sensors as described above, the off-axis type and the FI type are used.
Since the A-type alignment sensor uses broadband illumination light, it has the advantage that it is less susceptible to thin film interference in the photoresist layer coated on the wafer and less susceptible to wafer mark asymmetry. . However, the off-axis method has a disadvantage that the throughput (the number of processed wafers per unit time) is poor because the measurement position and the exposure position are relatively far apart.

On the other hand, the TTL and LIA alignment sensor is of the TTL system and requires only a short moving distance from the measurement position to the exposure position (the moving distance may be almost 0). Is advantageous.
However, in the LIA type and LSA type alignment sensors, scaling (linear expansion / contraction of the entire wafer), or an asymmetric shape generated on the surface of a wafer mark formed on the wafer by vapor deposition of aluminum, for example, An error may occur in the measurement result of the magnification of the chip pattern in the shot area. Further, since the light used is a monochromatic laser beam, it may be affected by thin film interference of the photoresist. Since the degree of these influences may vary from lot to lot due to variations in process steps, it is difficult to correct with only the constants obtained in advance.

In particular, since the above-mentioned EGA type alignment method is a method of increasing both the positioning accuracy and the throughput (the number of wafers processed per unit time), the alignment sensor to be used has both the accuracy and the throughput. It is desirable to use the one that is superior. Further, it is necessary not only to simply improve the positioning accuracy of each shot area, but also to improve the overlay accuracy within each shot area.

In view of the above point, the present invention provides a positioning method which is excellent in terms of throughput as well as high positioning accuracy and superposition accuracy in each shot area when performing EGA type alignment. The purpose is to do.

[0012]

According to the alignment method of the present invention, for each substrate in N (N is an integer of 2 or more) substrates, two-dimensional alignment is performed on the substrate (W) according to the designed arrangement coordinates. When aligning each of the plurality of shot areas arranged in a line with a predetermined reference position in the stationary coordinate system (X, Y) that defines the movement position of the substrate, among the plurality of shot areas, , The coordinate positions of the plurality of preselected shot areas (S _{1 to} S ₈ ) in the static coordinate system (X, Y) are detected, and the plurality of measured coordinate positions are statistically calculated, thereby The coordinate position in each static coordinate system (X, Y) of the shot area is calculated, and the substrate (W) is calculated according to the coordinate position thus calculated.
The present invention relates to a method of aligning each of the plurality of shot areas with respect to its reference position by controlling the movement position of the above, that is, an EGA type alignment method.

Then, according to the present invention, according to the coordinate position calculated by the statistical calculation, each of a plurality of shot regions on the kth (k is an integer of 2 or more and N or less) substrate and thereafter is set with respect to its reference position. Prior to performing the alignment by using two alignment sensors (10, 36), at least one of the substrates up to the (k−1) th substrate is used at a plurality of points (SC1 to SC4) in the shot area. Of 1
-Dimensional or two-dimensional position measurement is performed, and the difference of the statistical calculation result of the coordinate position measured by each alignment sensor and the statistical calculation result of the coordinate position measured by each alignment sensor in the shot area are obtained. (Steps 104 to 106), and when aligning the k-th and subsequent sheets, only one of the two alignment sensors (10) is used for one-dimensional or two-dimensional measurement at one point in the shot area. Position measurement is performed (step 110), and the results obtained by statistically calculating these measurement results are used as the difference between the statistical calculation results of the coordinate positions measured by the two already stored alignment sensors, and 2
The coordinate position measured by each of the two alignment sensors is corrected using the statistical calculation result in the shot area (step 111), and the position is adjusted based on the correction result (step 112).

In this case, at the time of positioning the kth and subsequent sheets, one of the two alignment sensors (10) performs one-dimensional or two-dimensional position measurement at one point in the shot area. , A one-dimensional coordinate position in a predetermined direction at another point in the shot area is measured by one of the alignment sensors, and the measurement result at that one point in the shot area and the measurement at that other point The result obtained by statistically calculating the result is the difference between the statistical calculation results of the coordinate positions measured by the two already-stored alignment sensors and the shot area of the coordinate position measured by each of the two alignment sensors. It is also possible to make a correction using the statistical calculation result in 1. and perform the alignment based on the correction result.

[0015]

According to such an aspect of the present invention, it is composed of N substrates.
For the first substrate of the lot or a few (k) substrates from the first, two types of alignment sensors (for example, TT
L system LIA system 10 and off-axis system FIA
The system 36) is used to perform multi-point intra-shot EGA alignment for the same sample shot. Then, the difference between the respective sensor values of the in-shot multi-point EGA values obtained for both alignment sensors is obtained and stored.

When exposing subsequent wafers, for example, alignment is performed in the usual EGA system using only the alignment sensor of the TTL system LIA system 10 or the like having a high throughput, and a predetermined bias is generated in the LIA system 10 or the like. Parameters that may occur (for example, scaling) are corrected by the difference with the FIA system 36 stored previously, and parameters within the shot such as the chip pattern magnification (shot magnification) are the most important for the process of the substrate. Exposure is performed using a suitable alignment sensor (for example, shot magnification is FIA system, shot rotation is LIA system, etc.) and the parameters stored in advance are used. This makes it possible to improve the alignment accuracy and the overlay accuracy in the shot area while maintaining high throughput.

In addition, for the kth and subsequent substrates, for example, the alignment is performed in the normal EGA system by using only the alignment sensor such as the TTL LIA system 10 which has a high throughput, and within the shot area. For the parameters (shot magnification, shot rotation, etc.) that change easily from wafer to wafer in the parameters, refer to the results obtained by actually measuring at the few measurement points using the most suitable alignment sensor of the two alignment sensors. Determined based on As a result, the overlay accuracy for each shot area can be further increased without significantly reducing the throughput.

[0018]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of the alignment method according to the present invention will be described below with reference to the drawings. FIG. 2 shows a configuration of a main part of a projection exposure apparatus equipped with an alignment apparatus used in this embodiment. In FIG. 2, illumination light for exposure (g line, i line from a mercury lamp, or excimer laser) is shown. (UV pulsed light from the light source) IL is a condenser lens C
The pattern area PA of the reticle R is illuminated via L with a uniform illuminance distribution. Illumination light IL passing through the pattern area PA
Enters the projection optical system PL that is telecentric on both sides (may be one side) and reaches the wafer W. The projection optical system PL has the best aberration correction with respect to the wavelength of the illumination light IL, and the reticle R and the wafer W are conjugated with each other under the wavelength. In the following, the Z axis is taken parallel to the optical axis AX of the projection optical system PL, the X axis is taken parallel to the paper surface of FIG. 2 in the plane perpendicular to the Z axis, and the Y axis is taken perpendicular to the paper surface of FIG. Explain.

The reticle R is held on a reticle stage RS that can be finely moved two-dimensionally, and the reticle R has a reticle alignment mark formed on the periphery of the reticle R.
6, the objective lens 17 and the mark detection system 18 detect the reticle alignment system to position the optical system AX with respect to the optical axis AX. On the other hand, the wafer W is mounted on the wafer stage ST which is two-dimensionally moved by the drive system 13, and the coordinate value of the wafer stage ST is sequentially measured by the interferometer 12. The stage controller 14 controls the drive system 13 based on the coordinate measurement values from the interferometer 12 and controls the movement and positioning of the wafer stage ST. A reference mark FM used for baseline measurement and the like is provided on the wafer stage ST.
Baseline measurement is to measure the interval (baseline) between the detection center of the alignment sensor and the center of the exposure field of the projection optical system PL, and in this example, the baseline is obtained in advance.

Next, the wafer mark as an alignment mark provided in each shot area on the wafer W in this embodiment will be described. FIG. 3A shows 1 on the wafer W.
One shot area 37-n and this shot area 37-n
Wafer mark pairs 39 (n, 1) to 39 attached to n
3A shows the positional relationship with (n, 4), and in FIG. 3A, the four sides of the shot area 37-n are scribe lines S.
Surrounded by CL. Then, four measurement points SC1 to SC are provided at the positions of the substantially square vertices in the shot area 37-n.
4 is set, and wafer mark pairs 39 (n, 1) to 39 (n, 4) are formed near the measurement points SC1 to SC4, respectively. Wafer mark pair 3 near the measurement point SC1
9 (n, 1) is composed of a wafer mark MW1 arranged at a predetermined pitch in the X direction and a wafer mark MY1 arranged at a predetermined pitch in the Y direction. 39 (n, 4) is composed of wafer marks MX2 to MX4 for the X axis and wafer marks MY2 to MY4 for the Y axis, respectively. The wafer marks MX1 to MX4 have the same structure, and the wafer marks MY1
~ MY4 have the same configuration.

Reference point 38 at the center of shot area 37-n
-N is on the optical axis AX of the projection optical system PL at the time of exposure. Then, for example, the centers of the wafer marks MX1 and MY1 are located on the straight lines extending in the X and Y directions, respectively, passing through the measurement point SC1. Wafer mark MX1 is measurement point SC1
The wafer mark MY1 is used to detect the position of the measurement point SC1 in the Y direction, and is a multi-mark in which a plurality of linear patterns are arranged in parallel. Similarly, other wafer mark pairs 39 (n, 2) to 39
(N, 4) is also used to detect the positions of the measurement points SC2 to SC4.

FIG. 3B shows a typical wafer mark MX1.
5 is an enlarged view of 5 linear patterns P extending in the Y direction.
₁ , P ₂ , P ₃ , P ₄ , P ₅ are arranged in the X direction at a substantially constant pitch. FIG. 3C shows the wafer mark M.
The cross-sectional structure of X1 in the X direction is shown. Here, five linear patterns P _{1 to} P ₅ are formed in a convex shape protruding from the base of the wafer W, and the upper surface thereof is covered with a photoresist layer PR. As shown in FIG. 3B, the shot area 3
A straight line passing through the measurement point SC1 of 7-n and parallel to the Y-axis passes through the width center of the linear pattern P ₃ at the center of the wafer mark MX1. The same applies to the wafer mark MY1, which is composed of five linear patterns, and the center line of the central linear pattern coincides with the center line of the measurement point SC1 in the Y direction.

In this example, since four pairs of wafer marks are provided in each shot area 37-n, it is possible to perform EGA type alignment of four points in each shot. Returning to FIG. 2, in the present embodiment, an off-axis FIA (imaging method) alignment sensor (hereinafter referred to as “FIA system”) 36 as a first alignment sensor, and a second alignment sensor TT
An LIA type (two-beam interference type) alignment sensor (hereinafter referred to as “LIA system”) 10 is provided. In this embodiment, the wafer marks MX1 to MX4 for the X axis shown in FIG. 3 are replaced by the LIA system 10 for the X axis of the TTL method.
And the off-axis type FIA system 36 are commonly detected, and the Y-axis wafer marks MY1 to MY4 are set to TT.
The L system YIA LIA system (not shown) and the off-axis FIA system 36 are commonly detected. The LIA system 10 and the FIA system 36 may have different wafer marks to be detected. In this case, there is no problem if the positional deviation amounts of different types of wafer marks are obtained in advance.

Next, the configuration of the alignment sensor will be described in detail. First, in the latter LIA system 10, the laser beam LB from the laser light source 1 is red monochromatic light such as He—Ne laser light and is non-photosensitive to the photoresist layer on the wafer W. This laser beam LB is incident on the heterodyne beam generating optical system 2 including an acousto-optic modulator, and the two laser beams LB1 and LB2, which have a slightly different frequency from each other and are coherent with each other, cross a predetermined intersection. Ejected at the corner.

Two emitted laser beams LB1, L
B2 is once Fourier-transformed through the mirror 3a and the lens system 4, then is inverse-Fourier-transformed through the mirror 3b and the objective lens 6, and then is obliquely provided below the reticle R at a tilt angle of 45 °. Is reflected by and enters the periphery of the visual field of the projection optical system PL. Then, the laser beams LB1 and LB
After being Fourier-transformed again in the vicinity of the pupil EP of the projection optical system PL, the light beams 2 are incident on the wafer W as parallel light beams at a predetermined crossing angle in the XZ plane. In this case, the mirror 7 is fixed outside the periphery of the pattern area PA of the reticle R and within the visual field of the projection optical system PL. Therefore, the laser beams LB1 and LB2 intersecting on the wafer W are
Are located outside the projected image of the pattern area PA. To detect the position in the X direction of the diffraction grating-shaped alignment mark (wafer mark) attached to each shot area of the wafer W by the pair of intersecting laser beams, the pitch of the wafer mark and the two laser beams are used. The crossing angle is set to a predetermined relationship so that the diffraction grating is emitted from the wafer mark in the same direction.

FIG. 5 shows a state where a predetermined wafer mark MX1 on the wafer W is symmetrically irradiated with two laser beams LBL1 and LB2. In FIG. 5, one laser beam LB1 from the wafer mark MX1 is shown. The + 1st-order diffracted light and the -1st-order diffracted light of the other laser beam LB2 are emitted vertically upward. Since the laser beam LB1 and the laser beam LB2 are coherent lights having slightly different frequencies, the diffracted light LB3 is a heterodyne beam in which the light intensity changes with the frequency difference as the beat frequency. Further, since the phase of the diffracted light LB3 changes in accordance with the position of the wafer mark MX1 in the X direction, the phase of the beat signal obtained by photoelectrically converting the diffracted light LB3, for example, a reference heterodyne beam is photoelectrically converted. By comparing with the phase of the beat signal obtained by conversion, the position of the wafer mark MX1 in the X direction can be obtained with extremely high resolution (for example, about several nm).

Returning to FIG. 2, the diffracted light generated substantially vertically upward from the wafer mark on the wafer W is projected by the projection optical system PL,
After passing through the mirror 7, the objective lens 6, and the mirror 3b, the light is reflected by the small mirror 5 arranged between the two incident side laser beams LB1 and LB2, and reaches the light receiving element 8.
The beat signal obtained by photoelectric conversion by the light receiving element 8 is a position measurement signal PDS of the wafer stage ST from the interferometer 12.
At the same time, it is supplied to the LIA arithmetic unit 9. A reference beat signal obtained by photoelectrically converting a standard heterodyne beam generated in the heterodyne beam generation optical system 2 is also supplied to the LIA calculation unit 9, and the LIA calculation unit 9 compares the phases of the two beat signals. To obtain information AP ₁ of the position of the wafer mark to be measured in the X direction,
This information is supplied to the main control system 50.

At this time, for example, when the phase of the reference beat signal and the phase of the beat signal corresponding to the wafer mark match, for example, the X coordinate of the wafer stage ST is
If the reference beat signal and the beat signal corresponding to the wafer mark are out of phase with each other as the X coordinate of the wafer mark as it is, the X coordinate of the wafer stage ST is converted into a value obtained by converting the amount of the phase deviation into displacement. The added coordinate becomes the X coordinate of the wafer mark. In addition, the phase of the beat signal is, for example, 36 at a 1/2 pitch cycle of the wafer mark.
Since it changes by 0 °, it is necessary to perform the positioning of the wafer W by search alignment (described later) or the like with an accuracy of, for example, ½ pitch of the wafer mark or less.

In the above, the laser light source 1, the heterodyne beam generation optical system 2, the mirrors 3a and 3b, the lens system 4, the mirror 5, the objective lens 6, the mirror 7, the light receiving element 8,
The LIA operation unit 9 and the projection optical system PL form an LIA system 10 for the wafer W. In addition, this LIA
The system 10 is an alignment sensor for detecting the position of the wafer mark for the X axis, and is also provided with a LIA system (not shown) for detecting the position of the wafer mark for the Y axis in the Y direction with the same configuration. There is.

Next, in the FIA system 36 as the first alignment sensor, the broadband light generated from the halogen lamp 20 is condensed on one end surface of the light guide 22 by the condenser lens 21. The light passing through the light guide 22 passes through the filter 23 that cuts the photosensitive wavelength (short wavelength) region and the infrared wavelength region of the photoresist layer, and reaches the half mirror 25 through the lens system 24. The illumination light reflected here is reflected substantially horizontally by the mirror 26, then enters the objective lens 27, and the field of the projection optical system PL is not blocked around the lower part of the lens barrel of the projection optical system PL. It is reflected by the fixed prism (mirror) 28 and irradiates the wafer W almost vertically. Although not shown here, in the optical path from the exit end of the light guide 22 to the objective lens 27, an appropriate illumination field diaphragm is provided at a position conjugate with the wafer W with respect to the objective lens 27. Further, the objective lens 27 is a telecentric system, and an image of the exit end of the light guide 22 is formed on the surface 27a of the aperture stop (same as the pupil), and Koehler illumination is performed. The optical axis of the objective lens 27 is set to be vertical on the wafer W so that the mark position does not shift due to the tilt of the optical axis when the mark is detected.

The reflected light from the wafer W passes through the objective lens 28 and the half mirror 25 and is imaged on the index plate 30 by the lens system 29. The index plate 30 is arranged conjugate with the wafer W by the objective lens 27 and the lens system 29, and has linear index marks extending in the X direction and the Y direction in a rectangular transparent window. Therefore, the image of the wafer mark on the wafer W is formed in the transparent window of the index plate 30, and the light flux from the wafer mark image and the index mark is
The light flux that has entered the half mirror 32 through the first relay system 31 and is split into two by the half mirror 32 is the second relay system 33.
Images are formed on the image pickup devices 34X and 34Y such as CCD cameras via X and SSY, respectively. Image sensor 34X, 3
The image pickup signal from 4Y is FIA (field image
The alignment) calculation unit 35 is supplied with the position measurement signal PDS from the interferometer 12. The FIA calculation unit 35 determines the deviation of the wafer mark image from the index mark on the index plate 30 based on the waveform of the image pickup signal. In this case, the position of the wafer mark in the X direction is detected by processing the image pickup signal from the X-axis image pickup device 34X,
The position of the wafer mark in the Y direction is detected by processing the image pickup signal from the Y-axis image pickup device 34Y.

FIG. 4 shows a state of the wafer mark MX1 detected by the X-axis image pickup device 34X. As shown in FIG. 4, the wafer mark MX1 to be detected is indicated by the index plate 30.
(See FIG. 2) Positioning is performed between the index marks 30a and 30b on the upper side, and a precise position XA in the X direction of the wafer stage ST at that time is obtained. Imaging device 34X is electrically scanned along five linear patterns P ₁ to P ₅ and index marks 30a of the wafer mark MX1, an image of the 30b to the scan line SL. At this time, for example, since only one scanning line is disadvantageous in terms of SN ratio, the levels of the image pickup signals obtained by a plurality of horizontal scanning lines in the video sampling area VSA indicated by the broken line are added for each pixel in the horizontal direction. Good to average. The image pickup signal has waveform portions of rising and falling with respect to each of the index marks 30a and 30b on both sides, and these positions (positions on pixels) XR ₁ and XR ₂ have been obtained in advance, and their midpoints. The position XR _{0 of is} also obtained.

On the other hand, the image pickup device 34X has a wafer mark MX.
Since 1 bright field image is photoelectrically detected, 5 linear patterns
Pawn P₁~ P_FiveAt the left and right step edges of each
The light returning to the objective lens 27 is extremely reduced due to the scattering.
It Therefore, the linear pattern P ₁~ P_FiveEach left of
The edge and the right edge are imaged like black lines. Follow
The waveform on the imaging signal is the left edge and right of each linear pattern.
It becomes the bottom at the position corresponding to the edge.

The FIA operation unit 35 uses the waveform as described above to form the wafer mark MX1 (patterns P ₁ to
The position Xm in the X direction of the center (straight line CX) of P ₅ ) is calculated. More specifically, the FIA operation unit 35 calculates the center position of each of the patterns P _{1 to} P ₅ based on the left and right edge positions, and then calculates the five linear patterns P.
_The positions of _{1 to} P ₅ are added and divided by 5, and the mark position in the X direction to be the center is detected.

The FIA operation unit 35 then calculates the difference ΔX between the previously determined position XR ₀ and the mark measurement position Xm.
F (= XR ₀ −Xm) is calculated, and a value obtained by adding the difference ΔXF to the position XA when the wafer stage ST is positioned is output to the main control system 50 as mark position information AP2. Further, in the FIA system 36, the filter 23
The illumination light of the wafer W passing through illuminates a local area (smaller than the shot area) including the wafer mark on the wafer W with a substantially uniform illuminance, and the wavelength range is set to a width of about 200 nm.

The members from the halogen lamp 20 to the FIA operation unit 35 in the order of codes form a FIA system 36. In addition, the objective lens 27 and the lens system 2
9. Since the light having a wavelength bandwidth of about 200 nm passes through the telecentric imaging optical system including the relay systems 31, 33X (or 33Y), it is naturally necessary to correct the chromatic aberration corresponding thereto. Further, the numerical aperture (NA) of the objective lens 27 on the wafer side is preferably set smaller than the numerical aperture of the projection optical system PL.

In the present embodiment, the observation field area of the objective lens 27 is made to partly go under the lens barrel of the projection optical system PL by the prism 28 so as to be as close as possible to the field of view of the projection optical system PL. Generally, in this type of projection exposure apparatus, a focus sensor that accurately detects the distance (deviation) between the image plane of the projection optical system PL and the surface of the wafer W, the surface of the shot area on the wafer W, and the projection optical system. A leveling sensor for detecting the relative inclination of the PL with the image plane is provided. The focus sensor and the leveling sensor are used in the projection optical system PL.
The wafer W in which the projection field of view exists is obliquely irradiated with a light flux in the infrared region, and the shift of the light receiving position of the reflected light is obtained to perform focusing and leveling. At this time, if the numerical aperture of the objective lens 27 is small, the depth of focus of the objective lens 27 becomes deep, and if focusing is performed based on the detection result of the focus sensor, almost FIA system 36 also performs detection in the focused state. .

Further, in the configuration shown in FIG. 2, the detection center of the off-axis type FIA system 36 (the position of the conjugate image of the center of the index plate 30) is away from the center of the projection optical system PL.
The FIA is on a straight line connecting the measurement position of the interferometer 12 and the center of the projection optical system PL, that is, the measurement axis (measurement beam center line).
By providing the detection center of the system 36, the Abbe error (off-axis error due to the tilt of the stage) is minimized.

When performing the alignment, it is necessary to perform search alignment (global alignment) for rough alignment of the wafer and fine alignment for highly accurate alignment. Regarding this search alignment, as disclosed in, for example, Japanese Patent Application Laid-Open No. 60-130742, there is a method in which a TTL alignment system and an off-axis alignment system are mixed. The apparatus of the present embodiment employs a sequence in which a TTL LIA system 10 which normally has a high processing speed detects alignment marks at three or two positions on a wafer to perform search alignment. However,
Since the alignment may not be performed normally (especially when the mark detection does not work well) depending on the thickness or type of the wafer base or the photoresist layer, the FIA system 36 using the off-axis wide-bandwidth illumination wavelength is used. There is also provided means for switching the sequence so as to execute the search alignment. In this case, the mark detection time when performing search alignment in the TTL LIA system 10, the magnitude and distortion of the mark detection signal, etc. are determined and the sequence is switched.

Next, the TTL LIA system 10 is turned off.
The FIA system 36 of the axis system, and the main control system 50 that controls the stage controller 14 and the like will be described. It is assumed that the main control system 50 constantly inputs the position information PDS from the interferometer 12. Alignment (ALG)
The data storage unit 501 can input both the mark position information AP ₁ from the LIA calculation unit 9 and the mark position information AP ₂ from the FIA calculation unit 35.

The EGA (Enhanced Global Alignment) operation unit 502 uses the statistical operation method based on each mark position information stored in the ALG data storage unit 501 to calculate the actual shot array coordinate values on the wafer and the chips. The parameter magnification (shot magnification) or the like is calculated, and the calculation result is sent to the sequence controller 506.

Exposure (EXP) shot map data section 5
03 is 4 in all shot areas to be exposed on the wafer
The design value of the array coordinate value of each wafer mark pair is stored, and this design value is sent to the EGA arithmetic unit 502 and the sequence controller 506. Alignment (ALG)
The shot map data unit 504 stores design values of array coordinate values of four wafer mark pairs in all shot areas (sample shots) to be measured on the wafer, and these coordinate values are sequenced with the EGA operation unit 502. To the controller 506. The correction data storage unit 505 stores various data for alignment, data for correction of positioning for exposure shots, and the like, and these correction data are sent to the ALG data storage unit 501 and the sequence controller 506. The sequence controller 506 determines a series of procedures for controlling the movement of the wafer stage ST during alignment or during step-and-repeat exposure, based on the above-mentioned data.

The projection optical system PL of this embodiment is equipped with an image forming characteristic controller 19. The imaging characteristic control device 19 adjusts, for example, the distance between predetermined lens groups in the lens groups forming the projection optical system PL, or adjusts the pressure of gas in the lens chamber between the predetermined lens groups. Thus, the projection magnification and the distortion of the projection optical system PL are adjusted. Thereby, for example, the magnification of the projected image of the reticle is adjusted in accordance with the shot magnification measured by the intra-shot multipoint EGA system alignment. The operation of the imaging characteristic control device 19 is also controlled by the sequence controller 506.

Next, the basic alignment sequence of this embodiment will be described. Here, it is possible to obtain both high throughput and high alignment accuracy,
An in-shot multipoint EGA method (enhanced global alignment method) capable of correcting shot magnification and shot rotation will be described. FIG. 6A shows an example of the wafer W to be exposed in this embodiment.
In (a), the orthogonal coordinate system (α,
β) along the plurality of shot regions 37-n (n = 1,
, Are arranged in a matrix, and chip patterns are formed in each shot area 37-n by exposure and development in the previous step. In FIG. 6 (a),
Five shot areas 37-1 of the plurality of shot areas
Only ~ 37-5 is shown as a representative.

A reference position is defined for each shot area 37-n. For example, when the reference position is a reference point 38-n at the center of each shot area 37-n, the designed coordinate values of this reference point 38-n in the coordinate system (α, β) on the wafer W are ( C _Xn , C _Yn ) (see 6 (b)). Also, each shot area 3
7-n, four wafer mark pairs 39 (n, 1), 39 (n, 2), 39 (n,
3) and 39 (n, 4) are provided together. These wafer mark pairs 39 (n, 1) to 39 (n, 4)
Is actually composed of the wafer marks MX1 to MX4 for the X-axis and the wafer marks MY1 to MY4 for the Y-axis, respectively, as described with reference to FIG.
For convenience of description, in the following drawings, it is represented as a cross-shaped mark. In this case, the centers of the cross-shaped marks are shown in FIG.
It is located on the measurement points SC1 to SC4 in the shot area shown in (a).

In this case, the shot areas 37-n shown in FIG.
Coordinate system (x, y) on the shot area as shown in (b)
, The wafer mark pair 39 (n, 1), 39
(N, 2), 39 ( n, 3), 39 (n, 4) to the coordinate system of the corresponding measurement point (x, y) wherein the coordinates at the design on the _{(S 1Xn, S 1Yn),} (S _2Xn , S _2Yn ), (S _3Xn , S
_3Yn ) and (S _4Xn , S _4Yn ). For example, for the wafer mark pair 39 (n, 1), the X coordinate S _1Xn in the coordinates (S _1Xn , S _{1Y n} ) indicates the X coordinate of the wafer mark MX1 for the X axis, and the Y coordinate S _{1 Yn} is Wafer mark MY1 for Y axis
The Y coordinate of is shown.

Returning to FIG. 6A, the wafer W is placed on the wafer stage ST of FIG. 2 and the reticle is projected onto each of a plurality of shot areas in which chip patterns have already been formed by the step-and-repeat method. Exposure is performed by sequentially superimposing images. At this time, the correspondence between the stage coordinate system (X, Y) that defines the movement position of the wafer stage ST and the wafer coordinate system (α, β) is not necessarily the same as that in the previous process. Therefore, from the design coordinate values (C _Xn , C _Yn ) of the reference point 38-n of each shot area 37-n related to the coordinate system (α, β), the stage coordinate system (X,
Even if the coordinates on Y) are obtained and the wafer is moved based on these coordinates, the shot areas 37-n may not be precisely aligned. Therefore, in the present embodiment, first, it is assumed that the positioning error is caused by the following four factors.

Wafer rotation: This is represented by a rotation parameter Θ indicating a residual rotation error of the wafer coordinate system (α, β) with respect to the stage coordinate system (X, Y). Orthogonality of the stage coordinate system (X, Y): This occurs because the feed of the wafer stage ST in the X-axis direction and the Y-axis direction is not exactly orthogonal, and is represented by the orthogonality parameter W indicating an orthogonality error. To be done.

Linear expansion and contraction in the α and β directions in the wafer coordinate system (α, β): This means that the wafer W expands and contracts as a whole due to the processing process and the like. This expansion / contraction amount is scaled Rx in the α direction and β direction, respectively.
And Ry. However, Rx is represented by the ratio between the measured value and the design value of the distance between two points in the α direction on the wafer W, and Ry is represented by the ratio between the measured value and the designed value between the two points in the β direction. Stage coordinate system (X,
Offset for Y): This is caused by the wafer W being shifted (shifted) by a small amount as a whole with respect to the wafer stage ST, and the offset parameters O _X , O _Y indicating the shift amounts in the X and Y directions. It is represented by.

When the error factors (1) to (4) are added,
Coordinate values on the design of the reference point (C _Xn, C _Yn) for shot area, the stage coordinate system to be positioned when actually exposed (X, Y) on the coordinates _{(C 'Xn, C' Yn} )
Is represented as follows.

[0051]

[Equation 1]

Here, if the first-order approximation is performed assuming that the orthogonality W and the rotation Θ are minute amounts, (Equation 1) is as follows.

[0053]

[Equation 2]

Up to this point, the accurate alignment of the reference position on each shot area 37-n (in this example, the reference point at the center of each shot area) has been described. However, even if the reference points of the shot areas are accurately aligned, the entire chip pattern in each shot area and the projected image of the reticle do not necessarily exactly overlap each other.

Next, the overlay error in each shot area will be described. As described above, FIG. 6 (b)
In any shot area 37-n on the coordinate system (x, y) coordinate values of the design on the _{(S 1Xn, S 1Yn) ~} (S
_4Xn , S _4Yn ), and the wafer mark pair 39
(N, 1) to 39 (n, 4) are formed. In this example, the overlay error in each shot area is
It is assumed that it is caused by two factors.

Rotation of chip pattern (chip rotation or shot rotation): This is, for example, the wafer W.
What occurs when the reticle R is rotated with respect to the stage coordinate system (X, Y) when the projected image of the reticle R is exposed on the top, or when yawing is mixed in the movement of the wafer stage ST. And is represented by a shot rotation parameter θ indicating a rotation error with respect to the coordinate system (x, y) of the shot area. Orthogonality of chip: This is an error in the orthogonality of the chip pattern caused by distortion of the pattern itself on the reticle 2 or distortion of the projection optical system 7 when the projection image of the reticle 2 is exposed on the wafer W, for example. Yes, and is represented by a parameter w of the chip orthogonality corresponding to the angle error.

Linear expansion / contraction of chips (chip scaling, or shot magnification): This is, for example, an error in the projection magnification when the projection image of the reticle 2 is exposed on the wafer W, or the entire wafer W due to the processing process of the wafer W. It is caused by the partial or partial expansion and contraction. Here, the shot magnification rx in the x direction, which is the ratio between the measured value and the design value of the distance between two points in the x direction of the coordinate system (x, y) of the shot area, and the distance between the two points in the y direction, 2 in the y-direction shot magnification ry, which is the ratio of the measured value to the design value
It shall represent a linear stretch of direction.

For example, FIG. 7A shows a wafer W in which a rotation error and a magnification error occur in the chip pattern of each shot area 37-n formed in the previous step.
In (a), an example of a shot area in the case where there is no rotation error and magnification error is a shot area 40-6 surrounded by a broken line.
It is represented by 40-10. On the other hand, the shot areas 37-6 to 37-10 actually formed on the wafer W have different rotation angles and magnifications. These errors are
As shown in FIG. 7B, the shot rotation in which the shot area 37-n is inclined with respect to the original shot area 40-n and the magnification of the shot area 37-n as shown in FIG. The original shot area 40-n can be separated into shot magnifications different from those.

However, in the example of FIG. 7, the chip orthogonality w of the chip pattern is 0, and the shot magnification rx in the x direction.
And the shot magnification ry in the y direction are equal to each other. When the above error factors ( ₁ ) to ( ₄ ) are added, the designed coordinate values on the shot area 37-n are (S _NXn , S _NYn ) (N
= 1 to 4) for the wafer mark pair 39 (n, N),
The coordinate system of the shot area to be actually aligned (x,
The coordinate values (S ′ _NXn , S ′ _NYn ) on y) are expressed as follows.

[0060]

(Equation 3)

Here, if the linear approximation is performed assuming that the chip orthogonality w and the shot rotation θ are minute amounts, (Equation 3) is expressed by the following equation.

[0062]

[Equation 4]

Now, in FIG. 6B, the array coordinate value of the reference point 38-n of an arbitrary shot area 37-n on the stage coordinate system (X, Y) is (C _Xn , C _Yn ). Therefore, the design coordinate values (D _NXn , D _NYn ) of the arbitrary wafer mark pair 39 (n, N) on the arbitrary shot area on the stage coordinate system (X, Y) are expressed as follows. To be done.
However, as described above, the wafer mark pairs 39 (n, 1) to 39 (n, 4) are distinguished by the value of N (1 to 4).

[0064]

(Equation 5)

The above three errors (1) to (3) occur when the chip pattern is printed on the layer on which the wafer mark of each shot area on the wafer is printed. In actuality, the wafer mark pair 39 (n, N) is affected by the above-mentioned error caused by the wafer processing process. Measured coordinate values) (F _NXn , F _NYn ) (N = 1 to 4)
Then, this coordinate value (F _NXn , F _NYn ) is expressed as follows from ( _Equation 2) and ( _Equation 4).

[0066]

(Equation 6)

Next, in the present embodiment, in order to facilitate the application of the least squares method, the scaling Rx of the wafer in the α direction and the scaling Ry of the wafer in the β direction in (Equation 6) are performed.
Are expressed as the following (Equation 7) using new scalings Γx and Γy, respectively. Similarly, the shot magnification rx in the x direction and the shot magnification ry in the y direction in the (Equation 6) are respectively set as the new shot magnification parameter γx,
And γy are used to represent the following (Equation 7).

[0068]

(Equation 7)

By rewriting the equation (6) using the two scalings Γx and Γy and the two shot magnifications γx and γy, which show the new changes in linear expansion and contraction, the equation (6) is approximated. Is as follows.

[0070]

(Equation 8)

In this (Equation 8), if the two-dimensional vector is regarded as a matrix of 2 rows × 1 column, this (Equation 8) can be rewritten as a coordinate conversion equation using the following conversion matrix.

[0072]

[Equation 9]

However, each conversion matrix of (Equation 9) is defined as follows.

[0074]

[Equation 10]

The ten error parameters (θ, θ) included in the transformation matrices A, B, O in the coordinate transformation equation of (Equation 9).
W, .gamma.x (or Rx), Γy (or _{Ry), O X, O Y} ,
θ, w, γx (= rx−1), ry) are called intra-shot multipoint EGA parameters. Among these 10 error parameters, four error parameters representing the state within the shot, that is, shot magnifications γx (or rx), γy (or rx), shot rotation θ, and chip orthogonality w are intra-shot parameters. Call. The above 10 multi-point EGA parameters within a shot can be obtained by, for example, the method of least squares. In this example, based on the coordinate conversion formula of (Equation 9), the stage coordinate system (X,
Y) Calculated coordinates on the table and each error of the chip are obtained. Then, based on it, the error of shot rotation (chip rotation) and shot magnification (chip magnification)
After correcting the error and the like, the calculated array coordinates (F _xn , F _yn ) of each shot area 37-n of the wafer W are obtained,
After aligning each shot area 37-n with the reticle based on the array coordinates, the pattern image of the reticle is exposed.

Now, in association with the main control system 50 shown in FIG. 2, the designed array coordinate values of the four wafer mark pairs in all the shot areas on the wafer are EXP shot map data section 503. And the designed coordinate values of the four wafer mark pairs in the sample shot are stored in the ALG shot map data section 504, and the least square approximation for determining the above-mentioned ten multipoint EGA parameters in the shot. The arithmetic expression of is stored in the EGA arithmetic unit 502.

It is not always necessary to apply the least squares method to (Equation 9), and for example, 10 in-shot multipoint EGA parameters may be obtained at the stage of (Equation 6). Also,
In the above (Equation 8), there are four shot parameters (shot rotation θ, chip orthogonality w,
Shot magnification rx (= 1 + γx), shot magnification ry
Although all (= 1 + γy) are included, (Equation 8) (or (Equation 6)) may be used by focusing on only one of these four parameters. Specifically, when focusing on only the shot rotation θ, the chip orthogonality w is considered to be 0, and the shot magnifications rx and ry are considered to be 1 and Equation 8 is used. In this regard, when focusing on the shot magnifications rx and ry, (Equation 8) may be used as rx = ry, that is, the linear expansion and contraction is isotropic.

Of the four shot parameters, the parameter of interest may be selected according to, for example, the type (feature) of the wafer to be exposed. Here, it is assumed that the wafer mark is measured for sample shots selected in advance from all shot areas 37-n of the wafer W. However, the reference point 38-n of each sample shot
Since the distance between the shot mark and the wafer mark in the shot area is relatively short, four shot parameters (θ, w, r) among the above ten shot multipoint EGA parameters
x, ry) has a large error due to measurement reproducibility and the like. Therefore, in order to reduce the error, it is necessary to relatively increase the number of sample shots and the number of wafer marks to be measured, but this lowers the throughput of the exposure process. Therefore, in this example, a decrease in throughput is prevented and the alignment accuracy is maintained high in the sequence described below. The following method utilizes that the linear error amount and the non-linear error amount that generally occur on the wafer and the deformation of the chip pattern in each shot area have the same tendency in the same lot.

Further, in the present embodiment, since two alignment sensors are provided, the 10 multi-point EGA parameters within a shot obtained by the alignment sensor which detects the coordinate value of each wafer mark of the sample shot is 2. It becomes a street. That is, when the EGA method alignment is executed based on the measurement result of the LIA system 10, (Equation 1
0) 10 shots Uchida point EGA parameters in, i.e. scaling .gamma.x, Ganmawai, rotation theta, orthogonality W, offset O _x, O _y, shot magnification .gamma.x, Ganmawai,
The shot rotation θ and the chip orthogonality w are obtained. Similarly, when the EGA method alignment is executed based on the measurement result of the FIA system 36, ten in-shot multipoint EGA parameters in (Equation 10) are obtained. However, for example, the parameters obtained based on the measurement results of the LIA system 10 may have a certain bias. Therefore, for such a biased parameter, FI
It is used after being corrected by the conversion parameter obtained based on the measurement result of the A system 36.

Next, with reference to the flowchart of FIG. 1, an example of the operation in the case of performing the exposure by performing the alignment (alignment) by the multipoint EGA method within a shot for one lot of wafers in the present embodiment will be described. To do. The following operation shows the operation when exposing a predetermined circuit pattern layer (process layer) on the wafer. Then, in the process layer, scaling Rx, Ry and shot magnification rx in the multipoint EGA parameters in the shot, which are obtained based on the result of the experiment and the evaluation at the time of trial manufacture, based on the measurement result by the LIA type alignment sensor, in advance.
ry (or γx, γy) has an error with a nearly constant tendency, that is, a certain deviation from the true value is mixed, but the error of other parameters is small, while the measurement result by the FIA alignment sensor Scaling Rx, Ry in the multipoint EGA parameters within the shot obtained based on
And it is known that the error between the shot magnifications rx and ry is small.

First, with respect to the main control system 50, the LIA system 1
A first shot multipoint EGA alignment sequence using 0 and a second intrashot multipoint EGA alignment sequence using the FIA system 36 are stored, and a sample shot (correct The wafer marks are arranged in the same manner in the first and second alignment sequences. Then, basically, the in-shot multipoint EGA parameters obtained in the first alignment sequence are used, but only the scaling Rx, Ry and the shot magnifications rx, ry among them are shots obtained in the second alignment sequence. The correction is performed based on the inner multipoint EGA parameters as described later.

Specifically, in step 101 of FIG. 1, the leading wafer W of one lot is loaded on the wafer stage ST of FIG. At this time, the alignment of the reticle R is completed, and the amount of deviation of the reticle R in the X direction, the Y direction, and the rotation direction with respect to the orthogonal coordinate system defined by an interferometer (not shown) is substantially zero. FIG. 8 shows a wafer W to be exposed. In FIG. 8, a large number of shot areas are arranged on the wafer W along a coordinate system (sample coordinate system) (α, β) on the wafer W. However, only the sample shots 37-1 to 37-8 of the measurement target among them are shown. In the shot area 37-n (n = 1 to 8), the wafer mark pair 39 (n, n of FIG.
4 wafer mark pairs 39 (n, 1) to 3 identical to 1)
9 (n, 8) are formed. Then step 10
In 2, search alignment (global alignment) is performed. That is, on the wafer W, several alignment marks for rough alignment are formed along the sample coordinate system (α, β) in addition to the wafer marks attached to each shot area. Therefore, for example, the LIA system 10 of FIG. 2 (including the LIA system for the Y axis, the same applies hereinafter), or FI
The coordinate values of those alignment marks in the stage coordinate system (stationary coordinate system) (X, Y) are measured by the A system 36,
Based on this measurement result, approximate conversion parameters (scaling, rotation, offset, etc.) from the sample coordinate system (α, β) to the stage coordinate system (X, Y) are obtained, and the main control system 50 of FIG. EGA operation unit 50 in
It is stored in the memory in 2.

After that, when performing the alignment in the multi-point EGA method within the shot, the array coordinates of the sample coordinate system (α, β) at the reference point of the shot area to be measured, and the reference point as the origin. Based on the array coordinates of the wafer mark pair and its approximate conversion parameters, the EGA arithmetic unit 50
2 the stage coordinate system (X,
The coordinate value in Y) is calculated, and this coordinate value is supplied to the stage controller 14 via the sequence controller 506. Then, based on the supplied coordinate values, the X-axis and Y-axis wafer marks of the wafer mark pair to be measured are sequentially moved to the observation field of the FIA system 36 or the irradiation area of the laser beam from the LIA system 10. To be done.

Then, in step 103 of FIG. 1, the in-shot multipoint EGA alignment is executed using the TTL LIA system 10 of FIG. That is, all the sample shots 37-1 to 37-on the wafer W shown in FIG.
8 wafer mark pairs (39 (1,1), ..., 39
Coordinate values (FM _NXn , FM) of (1, 4) to 39 (8, 1), ..., 39 (8, 4)) on the stage coordinate system (X, Y)
_NYn ) is actually measured for LIA system 10 (including LIA system for Y axis). This means that all wafer mark pairs in all sample shots on the wafer 8 are set as measurement targets.
However, it is not always necessary to target all the wafer mark pairs in all the sample shots. Since one wafer mark pair has two one-dimensional wafer marks,
In order to determine the values of 10 or more parameters, it is necessary to actually measure the coordinate values of at least 5 or more wafer mark pairs. However, with respect to a certain wafer mark pair, it is also possible to select such that only the X-axis wafer mark is measured and other wafer mark pairs are measured only with the Y-axis wafer mark.

In this case, the reference points 38-n (FIG. 6) of the sample shots 37-n (n = 1 to 8) on the wafer W are shown.
(B)), the designed array coordinate values (C _Xn , C _Yn ) on the coordinate system (α, β) on the wafer W, and the measured wafer mark pair 39 (n, N) (N). = 1 to 4), the design coordinate values (relative coordinate values) (S _NXn , S _NYn ) in the coordinate system (x, y) on each sample shot 37-n are known in advance. Therefore, in step 103, on the right side of (Equation 9), the designed array coordinate values (C _Xn , C _Yn ) of the reference point of the shot area to which the measured alignment mark belongs, and the design regarding the reference point of the alignment mark By substituting the relative coordinate values (S _NXn , S _NYn ) above, the wafer mark pair 39 (n, N) is moved to the stage coordinate system (X,
Y) _{Find the} calculated coordinate values (F _NXn , F _NYn ) which should be on.

Then, by the calculation of the least squares method in the EGA calculation unit 502, ten intra-shot multipoint EGA parameters (hereinafter referred to as “LIA parameters”) satisfying (Equation 9) are calculated. This LIA parameter is
Scaling Γx (or Rx), Γy (or Ry), rotation Θ, orthogonality W, offsets O _x and O _y , shot magnifications γx (or rx), γy (or ry), shot rotation θ, chip orthogonality w, It consists of

Specifically, the actually measured coordinate values (F
The difference (E _NXn , E _NYn ) between M _NXn , FM _NYn ) and its calculated coordinate value (F _NXn , F _NYn ) is considered as an alignment error. Therefore, E _NXn = FM _NXn -F _NXn , E _NYn = FM _NYn -F
_NYn has been established. Then, the sum of squares of the alignment errors (E _NXn , E _NYn ) relating to the measured alignment marks is sequentially partial differentiated with these 10 parameters, and an equation is set so that the value becomes 0, respectively.
If 10 simultaneous equations are solved, 10 parameters can be obtained.

After that, in step 104, each wafer mark pair 39 in the sample shots 37-1 to 37-8 on the wafer W in FIG. 8 is used by using the FIA system 36 in FIG.
The coordinate value in the stage coordinate system (X, Y) of (n, N) is measured, and alignment in the multi-point within shot EGA method is executed. The EGA arithmetic unit 502 performs arithmetic processing on the measurement values obtained in this manner to obtain ten multipoint EGA parameters within a shot (hereinafter referred to as “FIA parameters”).
Is called) is calculated. This FIA parameter is also scaled Γx (or Rx), Γy (or Ry), rotation Θ, orthogonality W, offsets O _x and O _y , shot magnifications γx (or rx), γy (or ry), shot rotation θ, The chip orthogonality w.

In the next step 105, scaling Γx (or Rx), Γy (or Ry) and rotation Θ which are normal EGA parameters (parameters related to the wafer) among the 10 multi-point EGA parameters in the shot obtained. , Orthogonality W, and offsets O _x , O _y
The difference ΔΓ x _FL (or ΔR) obtained by subtracting the LIA parameter from the FIA parameter is
x _FL ), ΔΓ y _FL (or ΔRy _FL ), ΔΘ _FL , ΔW _FL ,
Delta O.D. _XFL, and calculates a delta O.D. _YFL, stores these differences in the correction data storage unit 505 in the main control system 50. Furthermore,
Regarding the shot magnifications γx (or rx), γy (or ry), shot rotation θ, and chip orthogonality w, which are intra-shot parameters in the multi-point EGA parameter within shot, the LIA parameter and the FIA parameter themselves are corrected. The data is stored in the data storage unit 505.

Then, in step 106, the offsets O _x and O are applied to the wafer W measured here.
_{For y} , rotation Θ, orthogonality W, chip orthogonality w, and shot rotation θ, the values of LIA parameters are used, and scaling Γx (or Rx), Γy (or R) is used.
y), shot magnification γx (or rx), γy (or r
For y), use the value of the FIA parameter. Then, the reticle R is passed through the reticle stage RS of FIG. 2 so as to correct the wafer rotation Θ in the conversion matrix A of (Equation 9) and the shot rotation θ in the conversion matrix B.
Is rotated or the wafer W is rotated to correct the rotation of the chip pattern with respect to the stage coordinate system (X, Y). This means that the reticle R or the wafer W is rotated in accordance with the sum (θ + θ) of the rotation Θ forming the elements of the conversion matrix A of (Equation 10) and the shot rotation θ forming the elements of the conversion matrix B. means.

However, when the wafer W is rotated, the offset (O _X , O _Y ) of the wafer W may change.
After measuring the coordinate values of the wafer mark again, the calculation for obtaining the parameters by the above-mentioned least squares method (multipoint E in the shot)
It is necessary to recalculate the error parameter by performing GA calculation). That is, for example, when the wafer W is rotated by the angle (Θ + θ), the above steps 103 to 106 need to be repeated. This also means to confirm whether the new rotation Θ after the rotation of the wafer W is a value corresponding to the angle rotated in step 106.

The chip orthogonality w cannot be corrected in a strict sense, but by rotating the reticle R appropriately,
The error can be kept small. Therefore, it is possible to optimize the rotation amount of the reticle R or the wafer W so that the sum of the absolute values of the rotation Θ, the shot rotation θ, and the chip orthogonality w is minimized. Next, projection of the projection optical system PL is performed via the imaging characteristic control device 19 of FIG. 2 so as to correct the shot magnifications γx (or rx) and γy (or ry) in the conversion matrix B of (Equation 9). Adjust the magnification. This is the shot magnifications γx (= rx−1) and γy forming the elements of the conversion matrix B shown in (Equation 10).
It means adjusting the projection magnification of the projection optical system PL in accordance with (= ry−1).

After that, the transformation matrix A including the element consisting of the multipoint EGA parameters in the shot obtained in step 105.
And O in the EGA arithmetic unit 505,
In the following formula, the reference point 3 of each shot area 37-n on the wafer W is
By substituting the designed array coordinate values (C _Xn , C _Yn ) of 8-n, the calculated array coordinate values (G _Xn ) on the stage coordinate system (X, Y) of the reference point 38-n. , G _Yn ).

[0094]

[Equation 11]

Then, in step 107, the step
Array coordinates (G _Xn, G_Yn) And in advance
Required baseline amount (alignment sensor detection
Arrangement calculated based on the distance between the center and the exposure center)
The stage through the sequence controller 506.
By sequentially supplying to the controller 14, the wafer W
The reference points 38-n of the upper shot areas 37-n are sequentially illustrated.
Alignment with the center of the exposure field of the second projection optical system PL
Reticle for the shot area 37-n.
The pattern image of R is projected and exposed. And on the wafer W
After exposing all shot areas,
Move to 108.

In step 108, the exposed wafer W is unloaded, and the i-th (i = 2, 3, ...) Wafer to be exposed next in this lot is set to the wafer stage S in FIG.
Load on T. In this embodiment, it is assumed that there is no great difference in process state within the same lot.
Bias (process offset) mixed in the scalings Γx (or Rx) and Γy (or Ry) in the LIA parameter obtained based on the measurement result by the IA system 10.
Is considered to be almost constant within the lot. Furthermore, the shot magnifications γx (= rx−1), γy (= ry−1), and the shot rotation are considered to be substantially constant values within the lot.

Therefore, for the second and subsequent wafers,
Normal EGA alignment is performed using only the LIA system 10, and the scaling in the obtained LIA parameters is corrected using the difference obtained for the first wafer, and the shot rotation and shot magnification are set to 1 The value obtained and stored for the first wafer is used. Therefore,
As a sequence, step 109 to step 102
After performing search alignment in the same manner as in step 1,
10 and using the LIA system 10, sample shot 37-1 on the i-th wafer having the same arrangement as that in FIG.
-37-8, for example, the coordinate values of only the first wafer mark pair 39 (1,1) to 39 (8,1) are measured, and the measurement results are processed to obtain a normal EGA parameter (LIA parameter). Calculate the value of. For the wafer W measured here, in the subsequent step 111, the offsets O _x , O _y , the rotation Θ, and the orthogonality W are calculated by using the value of the LIA parameter obtained immediately before and scaling Γ x (or Rx), Γy (or R
For y), the scaling difference ΔΓx stored in step 105 is added to the value of the LIA parameter obtained immediately before.
_The value obtained by adding _FL (or ΔRx _FL ) and ΔΓy _FL (or ΔRy _FL ) is used, and the shot rotation θ and the chip orthogonality w
For, the value of the LIA parameter stored in step 105 is used, and the shot magnification γx (or rx), γy
For (or ry), F stored in step 105
Use the value of the IA parameter.

Thus, the values of all parameters are determined
However, one shot rotation and shot magnification
Since the second wafer has been corrected, the second and subsequent wafers
There is no particular need to make any corrections to the exhaust. However, low
Rotation of wafer or reticle R for rotation Θ
You may correct it by doing. Then multiple points within those shots
Transformation matrices A and O containing elements consisting of EGA parameters
In the EGA calculation unit 505, (Equation 1
1) The reference point 38 of each shot area 37-n on the wafer
-N designed array coordinate value (C_Xn, C _Yn)
And the stage coordinate system (X,
Y) Calculated array coordinate value (G_Xn, G_Yn)
It

Then, in step 112, the step
Array coordinates (G _Xn, G_Yn) And in advance
Sequence loci determined based on the required baseline amount
The stage through the sequence controller 506.
By sequentially supplying to the controller 14, the wafer W
The reference points 38-n of the upper shot areas 37-n are sequentially illustrated.
Alignment with the center of the exposure field of the second projection optical system PL
Reticle for the shot area 37-n.
The pattern image of R is projected and exposed.

Next, in step 113, it is judged whether or not there are wafers to be exposed in this lot. If there are wafers to be exposed, step 108 is performed.
Repeat 112 to 112 to perform alignment and exposure. Then, when the wafers to be exposed run out in step 113, this process is ended. As described above, in this embodiment, since the LIA system 10 and the FIA system 36 are used to perform alignment by the multi-point EGA method within a shot, not only the alignment accuracy between shots but also the shot rotation, shot magnification, and the like are There is an advantage that the alignment accuracy within the shot is also improved. However, in view of the fact that the parameters within a shot are almost the same within a lot, only the first wafer actually measures the multipoint EGA method within a shot, and the subsequent wafers are processed by a normal EGA method. Method, and the parameter is either a value obtained by measurement, a value stored in advance, or a value corrected with the value stored in advance, and thus the throughput is the There is also an advantage that there is almost no decrease in alignment with the EGA method.

In the embodiment of FIG. 1, in order to obtain the difference between the FIA parameter and the LIA parameter and the in-shot parameters (shot magnification, shot rotation, chip orthogonality), the FIA is performed only for the first wafer in the lot.
The system 36 and the LIA system 10 respectively measure sample shots. However, for several wafers from the beginning of the lot, for example, the in-shot multipoint EGA system alignment using the LIA system 10 and the in-shot multipoint EGA parameter system alignment using the FIA system 36 are performed, respectively. The average value of the LIA parameters, the average value of the FIA parameters, or the difference between the two values may be obtained for a single wafer.

For subsequent wafers, the LIA system 10 having a high throughput was used for multipoint EGA within a shot.
The alignment may be performed by a method. In this case, LIA
Parameters that are likely to be biased in system 10 (scaling,
The shot magnification) is corrected by the difference between the average value of the FIA parameters and the average value of the LIA parameters that have already been obtained, so that high throughput and high alignment accuracy can be obtained.

Further, as in steps 103 and 104 of FIG. 1, the number of sample shots for obtaining the correction value of the multipoint EGA parameter in the shot is determined by the step 1 of FIG.
The number may be larger than the number of normal sample shots to be measured in 10. Next, in the embodiment of FIG. 1, for the second and subsequent wafers, the LIA system 10 measures the two-dimensional coordinates of one point in each sample shot (X coordinate, Y coordinate of one wafer mark pair). However, for the second and subsequent wafers (or several and subsequent wafers), the normal EGA system measurement is performed using the LIA system 10 having a high throughput. The Y coordinate of one Y-axis wafer mark at a different position may be measured by the LIA system 10 or the FIA system 36. For example, referring to FIG. 8, each of the sample shots 37-1 to 37-on the wafer W using the LIA system 10.
One wafer mark pair 39 (1, 1) to 39 in 8
Measure the X and Y coordinates of (8,1) and
Each sample shot 37-1 to 37-3 using the IA system 10
Wafer mark pair 39 (1, 2) or 39 within 7-8
Measure the Y coordinate of (1,3).

In this case, since multipoint measurement is performed with respect to the Y axis within each sample shot, multipoint EG within the shot is measured.
The shot rotation θ, which is one of the in-shot parameters in the A parameter, can be obtained. Then, in the process corresponding to step 111 in FIG. 1, the shot rotation value obtained immediately before is used for the shot rotation θ, and
Rotation Θ that constitutes the elements of the transformation matrix A of 0)
Then, the reticle R or the wafer W is rotated in accordance with the sum (θ + θ) of the shot rotation θ obtained immediately before the elements of the conversion matrix B are formed. After that, by exposing,
Exposure is performed in a state where the rotation error between the reticle R and each shot area on the wafer is small.

Furthermore, within each sample shot, LIA system 1
Aside from measuring one wafer mark pair at 0,
The A system 36 may be used to measure the X coordinate of a pair of wafer marks having substantially the same Y coordinate. As a result, the shot magnifications γx (or rx) and γy (or ry) which are in-shot parameters are obtained. However, regarding the shot magnification, the FI stored by measuring and storing up to several wafers.
There is no particular inconvenience to using the value of the A parameter. This is because the shot magnification is usually stable in the lot, but the shot rotation may be different for each wafer.

In the above embodiment, as shown in FIG. 3B, the wafer mark pair 39 consisting of one-dimensional wafer marks arranged in the X and Y directions on the stage coordinate system.
(N, 1) to 39 (n, 4) are provided at four corners on the diagonal of the shot area 37-n. However, if four wafer mark pairs are not arranged on one straight line, for example, the arrangement is not necessarily required. Further, the number of wafer mark pairs may be any number of two or more.
Further, for example, the position of the X-axis wafer mark and the position of the Y-axis wafer mark may be different.

Furthermore, depending on the combination of alignment sensors used, a two-dimensional mark such as a cross pattern may be used as the wafer mark. Also,
It is not always necessary to form a wafer mark inside each shot area 37-n, and for example, in FIG. A wafer mark may be formed on the wafer.

In the above embodiment, the FIA system 3 shown in FIG. 2 is used.
Although 6 is an off-axis system and LIA system 10 is a TTL system, FIA system 36 may be used in, for example, TTL system or TTR system, and conversely, LIA system 10 may be used in, for example, TTR system. Good. Further, in the above-described embodiment, the LIA system 10 and the FI are used as alignment sensors.
Although the A system 36 is used, in addition to this, for example, a laser step alignment (LSA) type alignment sensor that relatively scans a wafer mark and a laser beam focused in a slit shape may be used. .

Furthermore, in the above embodiment, multiple points E in the shot are recorded.
As the GA method, a method of treating the wafer marks in each sample shot equally is applied, but each wafer mark in the sample shot on the wafer is given a weight determined according to, for example, the distance from the center of the wafer. A weighting method may be applied that determines the value of the conversion parameter so that the residual error component obtained as a result is minimized.

As described above, the present invention is not limited to the above-mentioned embodiments, and various configurations can be adopted without departing from the gist of the present invention.

[0111]

According to the first alignment method of the present invention, two types of alignment sensors (for example, LIA system and FIA system) can be used for k substrates from the beginning of N substrates (wafers of one lot). System) and the like, the same sample shot is used to perform the measurement in the multi-point shot EGA method, and the values of the obtained parameters and the differences between them are stored. Then, when handling the subsequent substrates, an alignment sensor with a high throughput (for example, a TTL LI
Normal EGA measurement is performed using only the A system). At the time of alignment, the EGA parameters in a normal substrate are biased by the alignment sensor having a high throughput, so that a component (for example, LI
Scaling in A system) is corrected by the difference from the stored value in another alignment sensor, and parameters in the shot (shot magnification, shot rotation, etc.)
For example, parameters stored as measured values by an alignment sensor considered to be suitable for the process (for example, shot magnification is FIA system, shot rotation is L).
High accuracy of alignment (such as accuracy of alignment between the center of each shot area and the projected image, and accuracy of overlay between each shot area and the projected image, for example) without significantly reducing throughput. Is obtained.

For the kth and subsequent substrates, when one-dimensional coordinate measurement is performed at another measurement point in each sample shot by the other alignment sensor, the shot magnification or shot rotation in the shot parameters is further performed. Since it is possible to correct even the above, there is an advantage that the overlay accuracy in the shot area is further improved.

[Brief description of drawings]

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

FIG. 2 is a configuration diagram showing a main part of a projection exposure apparatus used in an embodiment.

3A is an enlarged plan view showing an example of a shot area and a wafer mark on a wafer, FIG. 3B is an enlarged plan view showing a wafer mark, and FIG. 3C is a sectional view of FIG. 3B. .

FIG. 4 is a diagram showing an example of an observation image by an FIA type alignment sensor.

FIG. 5 is an explanatory diagram of a detection principle of an LIA type alignment sensor.

6A is a plan view showing a relationship between a stage coordinate system and a coordinate system on a wafer when alignment is performed by an in-shot multipoint EGA method in the embodiment, and FIG. 6B is a plan view of FIG. 6A. It is an enlarged view which shows the coordinate system in the shot area 37-n.

FIG. 7 is an explanatory diagram of shot rotation and shot magnification.

FIG. 8 is a plan view showing an example of an array of sample shots on a wafer W to be exposed in the embodiment.

[Explanation of symbols]

 R Reticle PL Projection optical system W Wafer ST Wafer stage 8 Light receiving element 10 LIA system 34X, 34Y Image sensor 36 FIA system 50 Main control system 502 EGA arithmetic unit 505 Correction data storage 37-n Shot area 38-n Reference point 37- 1-38 Sample shots 39 (n, 1) to 39 (n, 4) Wafer mark pair WX1 X-axis wafer mark WY1 Y-axis wafer mark

Claims (2)

[Claims] 1. For each of N (N is an integer of 2 or more) substrates, each of a plurality of shot areas arranged two-dimensionally on the substrate according to design arrangement coordinates is defined as follows. When aligning with respect to a predetermined reference position in a static coordinate system that defines the moving position of the substrate, of the plurality of shot areas, measure the coordinate position in the static coordinate system of a preselected shot area, By statistically calculating the plurality of measured coordinate positions, the coordinate position of each of the plurality of shot areas on the stationary coordinate system is calculated, and the movement position of the substrate is controlled according to the calculated coordinate position. Accordingly, in the method of aligning each of the plurality of shot areas with respect to the reference position, k is calculated according to the coordinate position calculated by the statistical calculation.
(K is an integer of 2 or more and N or less) Prior to aligning each of the plurality of shot areas on the second and subsequent substrates with respect to the reference position, For at least one sheet, two-dimensional alignment sensors are used to perform one-dimensional or two-dimensional position measurement at a plurality of points in the shot area, and the difference between the statistical calculation results of the coordinate positions measured by the respective alignment sensors, And the statistical calculation result in the shot area of the coordinate position measured by each alignment sensor is obtained and stored, and when aligning the kth and subsequent sheets, only one of the two alignment sensors is used. One-dimensional or two-dimensional position measurement at one point in the shot area is performed, and the result obtained by statistically calculating the measurement result is
The difference between the statistical calculation results of the coordinate positions measured by the two alignment sensors already stored and the statistical calculation result of the coordinate positions measured by each of the two alignment sensors in the shot area are used for correction, A positioning method characterized by performing positioning based on the correction result. 2. When performing alignment on the k-th and subsequent sheets, one of the two alignment sensors performs one-dimensional or two-dimensional position measurement at one point in the shot area and The alignment sensor measures a one-dimensional coordinate position in a predetermined direction at another point in the shot area, and statistically calculates the measurement result at the one point and the measurement result at the other point in the shot area. The obtained result is calculated as the difference between the statistical calculation results of the coordinate positions measured by the two alignment sensors which have already been stored, and the statistics of the coordinate positions measured by the two alignment sensors in the shot area. The alignment method according to claim 1, wherein the alignment is performed based on the result of the correction and the alignment is performed based on the result of the correction.