JP2555651B2 - Alignment method and device - Google Patents

Description

The present invention relates to a wafer stepper for manufacturing a semiconductor, or a photomask for manufacturing a photomask used in a proximity type or contact type collective exposure apparatus. Concerning repeaters, etc., in particular, a circuit pattern of a mask or a reticle is sequentially superposed on a circuit pattern of a plurality of semiconductor elements formed on one photosensitive substrate (semiconductor wafer, hereinafter referred to as a wafer) by a step-and-repeat method. The present invention relates to an alignment method for exposing by exposure.

[Conventional technology]

Conventionally, a reduction projection type exposure apparatus, a so-called wafer stepper, is mainly used in the manufacturing process of a semiconductor element, and a circuit pattern formed on a mask or reticle (hereinafter referred to as a reticle) and a projection lens are used to form a wafer on the wafer. The circuit pattern already formed is accurately superimposed and exposed.
At this time, it is necessary to accurately align (align) the circuit pattern of the reticle and the circuit pattern of the wafer. At present, in this type of exposure apparatus, since one projection exposure field is about 5 mm to 20 mm square, the so-called step-and-repeat method is repeated in which the wafer is advanced by a fixed pitch and exposed. It is adopted. In case of this step-and-repeat exposure apparatus,
Since a plurality of shot areas to be exposed are formed in a matrix on the wafer, when the wafer is stepped, it is required that the shot areas are accurately aligned with the next shot area. However, when the exposure area for each shot increases with the high integration of semiconductor elements, and the line width of the circuit pattern to be transferred onto the wafer becomes about sub-micron, when the reticle is placed on the reticle stage. Due to the minute rotational deviation that occurs and the rotational deviation of the wafer stage that occurs due to the yawing of the wafer stage, the overlay accuracy of the reticle circuit pattern and the wafer circuit pattern decreases. Furthermore, when the circuit pattern of the first layer is transferred onto the wafer, the circuit pattern transferred onto the wafer is slightly rotated with respect to the array coordinates, and the superposition of the second and subsequent layers is insufficient. Becomes
For this reason, it becomes impossible to obtain a semiconductor element satisfying the desired characteristics with high productivity. Therefore, for example, an alignment mark (wafer mark) provided along with a shot area corresponding to one projection exposure field on the wafer and an alignment mark (reticle mark) formed around the circuit pattern of the reticle are provided. Simultaneously observing with an alignment optical system provided above the reticle, the position deviation is corrected by finely moving the wafer or reticle for each shot so that the position deviation amount of both marks becomes zero.
TTR (Through the Reticle) method step alignment is performed. As a result, it is possible to prevent a slight rotation deviation of the reticle, a rotation deviation of the wafer stage due to yawing of the wafer stage, and a decrease in overlay accuracy due to expansion and contraction (run-out) of the wafer.

[Problems to be solved by the invention]

However, in the exposure apparatus using this type of alignment method, the throughput is reduced because the reticle mark and the wafer mark are simultaneously detected for each shot to correct the positional deviation. Therefore, after correcting the relative rotation error between the reticle circuit pattern and the circuit pattern already formed on the wafer by rotating the θ table of the wafer stage by the rotation error, the circuit already formed on the wafer is corrected. There is a method in which the wafer stage is sequentially stepped by a predetermined amount along the arrangement coordinates of the pattern to perform overlay exposure. However, even in this method, if yawing occurs on the wafer stage, there is a problem that the overlay accuracy is reduced.

Also, when the first-layer circuit pattern is exposed on the wafer by the step-and-repeat method, the rotation deviation of the wafer stage due to yawing causes irregular rotation deviation for each shot with respect to the array coordinates. Occurs. For this reason,
There is also a problem that the overlay accuracy of the circuit pattern of the reticle on the second and subsequent layers and the circuit pattern of the wafer is lowered, and the alignment takes time and throughput is lowered.

The present invention has been made in consideration of the above points, and provides a method for preventing deterioration of overlay accuracy, throughput, and the like due to rotational displacement of a wafer stage, and enabling alignment with high accuracy and in a short time. is there.

[Means for solving problems]

In order to solve such a problem, according to the present invention, the reticle stage on which the reticle is placed and which moves two-dimensionally is provided with each circuit already formed on the wafer in the overlay exposure of the second and subsequent layers. In the relative rotation deviation between the pattern and the reticle, or in the exposure of the first layer, the movement coordinate of the wafer stage that moves the wafer stage two-dimensionally when the exposure is performed by the step-and-repeat method. It is configured to be driven to rotate in order to correct the relative rotational deviation of the reticle with respect to the coordinate system XY as a system. A second rotation amount detection device including a laser interferometer that detects the rotation amount of the wafer stage and a θ counter is reset to a predetermined value when the exposure apparatus is started, and a laser interferometer that detects the rotation amount of the reticle stage. The first rotation amount detection device, which is composed of a θ counter and a θ counter, is reset to a predetermined value after the relative rotation deviation is corrected. In this way, the first rotation amount detection device and the second rotation amount detection device are associated with each other, and the reset position of the first rotation amount detection device as the first reference position and the second position as the second reference position, respectively. The rotation amount from the reset position of the rotation amount detection device is detected. Further, when the reticle circuit pattern is exposed to each shot area on the wafer by the step-and-repeat method, the reticle stage is driven to rotate by a rotation amount substantially equal to the rotation amount detected by the first rotation amount detection device. By doing so, the second rotation amount detection device is configured to detect.

[Action]

According to the present invention, the amount of rotation of the wafer stage due to yawing that occurs when the wafer stage is stepped is measured using a laser interferometer, and the amount of rotation that is approximately equal to the amount of rotation of the wafer stage is detected by the laser interferometer of the reticle stage. The reticle stage is rotationally driven as described above. Thus, the exposure is performed by correcting the relative rotation error between the reticle circuit pattern and the wafer circuit pattern due to yawing or the like for each shot.
In overlay exposure from the second layer onward, it is possible to prevent a decrease in overlay accuracy between the reticle circuit pattern and the wafer circuit pattern due to wafer stage rotation deviation caused by yawing and a decrease in throughput due to this decrease in accuracy. Also, in the exposure of the first layer, the circuit pattern of the first layer can be transferred to the coordinate system XY, which is the moving coordinate system of the wafer stage by the step-and-repeat method, without any rotational displacement, thus achieving high accuracy. The alignment is performed in a short time.

〔Example〕

FIG. 1 is a perspective view showing a schematic configuration of a suitable projection exposure apparatus to which the method according to the first embodiment of the present invention is applied. FIG. 2 is a block diagram of a control system of this projection type exposure apparatus.

In FIG. 1 and FIG. 2, exposure light from an illumination optical system (not shown) uniformly illuminates the reticle R mounted on the chicle stage 2 which moves two-dimensionally (X direction, Y direction and rotation direction). To do. The light image of the pattern region Pr of the reticle R is projected onto the wafer W via the projection lens 1 which is telecentric on both sides or one side. The wafer W is mounted on the wafer stage 6 which moves two-dimensionally (X direction, Y direction and rotation direction) by the step-and-repeat method. Further, the wafer stage 6 is provided with a reference mark FM having the same pattern shape as the alignment mark (wafer mark) formed on the wafer W. Here, the projection lens 1 used in this embodiment is an optical system in which the reticle R side is non-telecentric and the wafer W side is telecentric. Further, the driving unit 3 finely moves the reticle stage 2 in the X direction, and the driving units 4, 5 finely move the reticle stage 2 in the Y direction and the rotation direction orthogonal to the X direction. (Including also) positioning is possible. The two-dimensional position of the reticle stage 2, that is, the position in the X direction, the Y direction, and the rotation direction, is a light wave interferometer such as a laser interferometer as shown in FIG. 2 (hereinafter referred to as a laser interferometer). 2
Detected by 0, 21, 22. This laser interferometer 20, 2
The configurations 1 and 22 are disclosed in, for example, Japanese Patent Application Laid-Open No. 62-150721 filed by the applicant of the present application, so a detailed description thereof will be omitted, but a laser beam emitted from a laser light source is used for the reticle stage. Irradiation is performed on a right-angled mirror as a movable mirror fixed to 2 and a fixed mirror fixed to the barrel of the projection lens 1. The reflected light beam from the movable mirror and the reflected light beam from the fixed mirror are coaxially combined to generate an interference fringe on the light receiving surface of the receiver. This receiver photoelectrically detects the change in the interference fringes and produces an up / down pulse according to the change in the position of the movable mirror. These pulses are reversibly counted by the counters 23 and 24, and the position of the reticle stage 2 is detected. The X counter 23 reversibly counts up / down pulses output from the X-axis interferometer (receiver) 20 to detect the position of the reticle stage 2 in the X direction. Y-axis interferometer 21, θ-axis interferometer 22 and θ counter as a first rotation amount detecting device
24, the θ counter 24 is the Y-axis interferometer (receiver) 21 and θ
The rotation amount of the reticle stage 2 is detected by reversibly counting the difference between the up and down pulses output from the axis interferometer (receiver) 22. At this time, the position in the Y direction is simultaneously detected.

On the other hand, the wafer stage 6 is moved in the X and Y directions by the drive units 8 and 9, respectively, and the wafer holder (θ table) 7 sucks the wafer W under vacuum, and the drive unit 10 makes a minute rotation on the wafer stage 6. To be configured. In order to detect the two-dimensional position of the wafer stage 6, an optical wave interferometer (hereinafter referred to as a laser interferometer) 11, 12, 13 such as a laser interferometer, and a laser as shown in FIG. X counter 25, Y counter 26, θ for reversibly counting up / down pulses output from interferometers 11, 12, 13 respectively.
A counter 27 is provided. Mirror as a moving mirror
Reference numeral 14 extends its reflection plane in the Y direction, and mirror 15 extends its reflection plane in the X direction.
It is provided above. Therefore, the laser interferometer 11 is a mirror
The laser beam Bx having an optical axis parallel to the coordinate axis X is emitted to the laser beam 14, and the fixed mirror provided inside the laser interferometer 11 (or under the lens barrel of the projection lens 1) is also irradiated with the laser beam. The laser interferometer 11 photoelectrically converts a change in an interference fringe generated by the interference between the reflected light beam from the mirror 14 and the reflected light beam from the fixed mirror, and outputs an up / down pulse. The X counter 25 reversibly counts the up / down pulses to detect the position of the wafer stage 6 in the X direction. Similarly for the laser interferometer 12 and the Y counter 26,
A laser light flux having an optical axis parallel to the coordinate axis Y on the mirror 15.
Is irradiated, and the position of the wafer stage 6 in the Y direction is detected by the interference between the reflected light beam from the fixed mirror and the reflected light beam from the mirror 15. The projection lens 1 and the laser interferometer 11,
Reference numeral 12 indicates that the laser beam Bx and the laser beam By are orthogonal to each other in the same plane, the optical axis AX of the projection lens 1 passes through the intersection, and a plane including these two measurement axes (for example, the center line of the laser beam) is projected. It is arranged so as to substantially coincide with the projection image plane of the lens 1.

On the other hand, the parallel laser beam from the laser interferometer 13 is split into two by a beam splitter (not shown). The laser beam reflected by the reflection mirror (not shown) passes through the laser beam Bishita ₁ and the beam splitter Bishita ₂ enters a vertical mirror 15. This laser light flux Bθ ₁ , Bθ
_The laser beam By is irradiated onto the mirror 15 so as to be symmetrical with respect to the laser beam By, that is, the Y axis of the coordinate system XY. Due to the interference of the two reflected light beams from the mirror 15, the laser interferometer 13 as a second rotation amount detection device and the θ counter 27 detect the rotation amount of the wafer stage 6 due to yawing or the like. Since the rotation amount of the wafer stage 6 is minute, the laser interferometer 13 can simultaneously receive the reflected light beams from the respective mirrors 15 of the laser light beams Bθ ₁ and Bθ ₂ and detect the rotation amount by the interference of the reflected light beams. It becomes possible.

Now, the TTR type die-by-die alignment system (hereinafter referred to as DDA) 16a shown in FIG.
Upper mark Rsa and mark on wafer W or reference mark F
The mark on M is superposed through the projection lens 1 and simultaneously observed, and both mark images are photoelectrically detected. Then, the positional deviation in the Y direction between the reticle R and the mark image of the wafer W or the reference mark FM is detected. Laser step alignment systems (hereinafter referred to as LSA systems) 17a and 17b are provided at intervals of 90 degrees with respect to the optical axis AX, as shown in the figure. These LSA systems 17a and 17b are TTL (through the
A lens is used to detect a predetermined mark (wafer mark) on the wafer W or a reference mark FM with a laser beam,
The positional deviation of the wafer W or the reference mark FM in the X and Y directions is detected.

The main control system 60 shown in FIG. 1 includes a processor (hereinafter, referred to as CPU) 30 such as a microcomputer and a minicomputer, an interface circuit (hereinafter, referred to as IFC) 31, etc., as shown in FIG. Through these, the operation of the entire apparatus including the above two alignment systems (DDA system, LSA system) is centrally controlled. Laser step alignment processing circuit (hereinafter referred to as LSAC) 32 is LSA system 17a, 17b
By using the LSA system 17 of the alignment mark on the wafer W.
The position with respect to the spot light (sheet beam) from a and 17b is detected. A die-by-die alignment system processing circuit (hereinafter referred to as DDAC) 33 cooperates with the DDA system 16a, and the mark Rsa of the reticle R and the alignment mark on the wafer W or the mark My of the reference mark FM in the Y direction. The amount of relative positional deviation regarding CPU30 is C
Depending on the calculated value of the PU 30 and the amount of positional deviation detected by various alignment systems, the drive unit 8 in the X direction of the wafer stage 6 (hereinafter, referred to as X-ACT8) and the drive unit 9 in the Y direction (via the IFC 31) ( Hereinafter, a predetermined drive command is output to each of the drive unit 10 (hereinafter, referred to as θ-ACT10) for rotating the wafer holder 7 and Y-ACT9. Similarly, in the reticle stage 2 as well, depending on the amount of rotation of the wafer stage 6 and the amount of positional deviation detected by the alignment system, the drive unit 3 (hereinafter X
-ACT3), a Y direction drive unit 4 (hereinafter referred to as Y-ACT4), and a rotation direction drive unit 5 (hereinafter referred to as? -ACT5), and outputs a predetermined drive command.

Next, the operation of this embodiment will be described with reference to the schematic flow chart shown in FIG. FIG. 3 shows a second pattern which occurs when the circuit patterns of the second and subsequent layers are superposed and exposed on the wafer W on which the circuit patterns of the first layer are arranged.
A basic operation of correcting a relative rotation error due to yawing between a projected image Pr ′ of the circuit pattern of the reticle R on the layer after the layer and a circuit pattern (hereinafter referred to as a chip C) arranged on the wafer W is shown. ing. The operation from the alignment of the reticle R and the wafer W to the exposure, that is, step 201, will be briefly described because it is disclosed in, for example, Japanese Patent Application Laid-Open No. 60-186845 previously filed by the applicant of the present application. Hereinafter, a case where the second layer circuit pattern is overlaid on each chip C on the wafer W and exposed is described.

In step 200, the CPU 30 activates the exposure apparatus,
The count value of the θ counter 27 that reversibly counts the up / down pulse according to the rotation amount of the wafer stage 6 is reset to a constant value, for example, zero. By this reset operation, the rotation amount of the wafer stage 6 due to yawing or the like is always detected with the zero reset position as a reference, that is, the origin of the rotation direction of the wafer stage 6 (hereinafter referred to as the θ axis) is determined. Here, the position of the θ axis with respect to the coordinate system XY determined by the laser interferometers 11 and 12 is the X-axis and the reflection surface of the mirror 15 that reflects the laser beams Bθ ₁ and Bθ ₂ emitted from the laser interferometer 13. It is determined by the minute inclination that This is because it is possible that the mirror 15 is set by a slight rotation with respect to the coordinate system XY due to an attachment error or the like when the mirror 15 is manufactured. Therefore, in the present embodiment, it is assumed that the θ counter 27 is reset to zero when the θ axis (reflection surface of the mirror 15) is rotationally displaced by θm with respect to the coordinate system XY as shown in FIG. It should be noted that each rotation deviation is exaggerated in FIG.

Next, in step 201, the reticle alignment microscope (not shown) used for reticle alignment is adjusted (checked). Then, using this reticle alignment microscope, reticle alignment is performed so as to remove the rotational deviation of the reticle R with respect to the coordinate system XY, but the rotational deviation remains at a minute angle. Therefore, the residual rotation error of the reticle R, that is, the reticle rotation RR is measured. Therefore, the operation of reticle rotation measurement will be described with reference to FIGS. 4 and 5.
At the time of reticle rotation measurement, the wafer stage 6 is assumed to move in parallel with respect to the θ axis without rotating (yawing) as shown in FIG. In FIG. 5, the DDA system 16a irradiates the mark RSa of the reticle R with the illumination luminous flux Ba through the reflecting mirror 16ma, and the mark RSa is projected through the projection lens 1 onto the reference mark FM as shown in FIG. 4 (A). Image on. The CPU 30 moves the wafer stage 6 so that the linear mark My of the reference mark FM is accurately sandwiched in the center of the projected image RSa ′ of the mark RSa (rectangular transparent window) of the reticle R. This state is observed by the DDA system 16a as shown in FIG. At this time, the DDAC 33 detects the position of the wafer stage 6 in the Y direction by the laser interferometer 12, and the CPU 30 stores the Y coordinate value Y ₁ ′. Next, as shown in FIG. 4 (B), the CPU 30 moves the wafer stage 6 along the X axis of the coordinate system XY, and the illumination beam Bb of the DDA system 16b causes the mark RSb of the reticle R (rectangular transparent window). The projection image RSb 'and the mark My are positioned so that they can be detected at the same time. At this time, if the reticle R is placed on the reticle stage 2 with no rotational displacement with respect to the coordinate system XY, the mark My is accurately sandwiched in the center of the projected image RSb ′ and observed, as shown in FIG. The amount of reticle rotation is zero. However, in general, since there is a rotation error although it is a minute angle, the projected image RSb ′ and the mark My are relatively displaced and observed in the Y direction. Therefore, the DDAC 33 detects the positional deviation in the Y direction by the DDA system 16b and finely moves the wafer stage 6 in the Y direction so as to eliminate the positional deviation. Then, the CPU 30 displays the Y coordinate value Y ₂ ′ when the positional deviation is corrected as shown in FIG. 4 (C).
Is stored. Further, the moving distance of the wafer stage 6 in the X direction at this time can be obtained from a design value (a distance between the marks RSa and RSb of the reticle R), but here, the X of the wafer stage 6 detected by the laser interferometer 11 is detected. Travel distance
Use Lx.

When measuring reticle rotation here,
The wafer stage 6 was not rotated with respect to the θ axis as shown in FIG. However, in reality, since the wafer stage 6 is moved so as to simultaneously detect the mark My and the marks Rsa and Rsb of the reticle R, yawing may occur on the wafer stage 6. Therefore, the CPU 30 detects the Y coordinate systems Y ₁ ′ and Y ₂ ′ that include the rotational deviation of the wafer stage 6 due to yawing as an error. Therefore, in order to prevent the measurement accuracy from deteriorating due to the occurrence of yawing, the CPU 30 obtains the rotation amount of the wafer stage 6 when the mark My and the marks Rsa and Rsb are simultaneously detected from the laser interferometer 13 and the θ counter 27. . Next, the CPU 30 detects the Y coordinate value detected as described above based on the detected rotation amounts.
Correct Y ₁ ′ and Y ₂ ′ respectively, and use the corrected Y coordinate values respectively.
It is stored as Y ₁ ′ and Y ₂ .

Further, as shown in FIG. 4, the θ axis of the wafer stage 6 is a coordinate system.
It is determined by shifting θm from XY. Therefore, the moving distance Lx in the X direction and the moving distance ΔY = (Y _{2 in the} Y direction of the wafer stage 6 detected by the laser interferometers 11 and 12, respectively.
-Y ₁ ) is detected by deviating from the actual moving distance by X'and Y ', respectively. Therefore, the reticle rotation amount is calculated not by the coordinate system XY but by the CPU 30 as the apparent rotation amount θr of the reticle R with respect to the θ axis.

From this, the CPU 30 determines the rotation amount θ of the reticle R with respect to the θ axis.
r is calculated from the equation (1).

However, since the rotation amount θr is extremely small, the approximate calculation is performed by the equation (2).

This reticle rotation measurement is performed a plurality of times in order to improve accuracy, and the CPU 30 stores a value obtained by averaging the rotation amount θr obtained for each of them.

On the wafer W, the circuit pattern of the reticle R of the first layer is the coordinate system as the moving coordinate system of the wafer stage 6.
It is assumed that the image is transferred without rotation deviation with respect to XY. However, in reality, due to a decrease in alignment accuracy or the like, the rotational deviation of each chip C may remain.
Therefore, when the patterns of the second and subsequent layers are superposed and exposed, the residual rotation error of the chip C, that is, the chip
It is necessary to consider rotation.

FIG. 7 shows the arrangement state of the specific chips C transferred onto the X axis, and each rotation deviation is exaggerated. If the counterclockwise direction is the positive direction in FIG. 7, the center of each chip C is accurately located on the X axis of the coordinate system XY, but each chip C
Is rotated by θc in the positive direction with respect to the θ axis due to chip rotation. The projected image Pr ′ of the circuit pattern of the reticle R of the second layer is rotated by θr in the negative direction with respect to the θ axis due to the reticle rotation. Therefore, the projection image Pr ′ and the chip C are Δθ =
This means that they are relatively rotating by (θc−θr).
On both sides of each chip C on the wafer W, mark images Sy and Sy ′ of marks formed in association with the pattern region Pr of the reticle R of the first layer are formed. Both marks Sy and Sy 'are known diffraction grating patterns extending on the X axis.

Next, the CPU 30 measures the chip rotation.
The spot rotation by the LSA system 17a is used for the chip rotation measurement, and the rotation amount θc of the chip C is obtained by detecting the positional deviation of the marks Sy and Sy ′ in the Y direction. Here, in the same manner as the reticle rotation measurement, the tip rotation measurement also includes the rotation amount of the wafer stage 6 at the time of measurement detected by the laser interferometer 13 or the like in order to prevent a decrease in measurement accuracy due to yawing during measurement. The detected Y coordinate value is corrected based on the above.

Therefore, the CPU 30 first moves the wafer stage 6 so that the spot light scans the mark Sy of the arbitrary chip C on the wafer W in the Y direction. And spot light and mark
The Y coordinate value when Sy matches is read into the laser interferometer 12, and the value corrected based on the rotation amount (change amount from zero reset) of the wafer stage 6 at the time of measurement is stored as the Y coordinate value Y _3. To do. Next, the wafer stage 6 is moved in the X direction so that the spotlights are aligned in parallel with the marks Sy '. The amount of movement is the distance between the marks Sy and Sy ′, but the laser interferometer
It should be read exactly from 11, and this value is stored as Px. Similarly, the CPU 30 scans the mark Sy ′ with the spot light in the Y direction, reads the Y coordinate value when the spot light and the mark Sy ′ coincide with each other from the laser interferometer 12, and rotates the wafer stage 6 at the time of measurement ( The value corrected based on the change amount from the zero reset) is stored as the Y coordinate value Y ₄ . Further, these measured values include the rotation deviation θm with respect to the coordinate system XY of the θ axis as an error, as in the RR measurement. Therefore tip
The rotation is apparently calculated by the CPU 30 as the rotation amount θc of the chip C with respect to the θ axis.

From this, the CPU 30 calculates the rotation amount θc of the chip C by the following equation (3). However, since θc is extremely small, it is approximated.

From the viewpoint of improving accuracy, this chip rotation measurement is similarly performed for a plurality of chips C located near the center of the wafer W or at the apex of a regular polygon around the wafer W, and θc obtained for each of the chips C is determined. What is averaged is used.

Next, the CPU 30 corrects the relative rotation error between the projected image Pr ′ of the reticle R and the chip C based on the measured chip rotation and reticle rotation.
Therefore, first, the relative rotation amount Δθ = (θc−θr) between the chip C and the projected image Pr ′ is obtained. Then, the reticle stage 2 is moved in the positive direction Δ with respect to the coordinate system XY via the drive unit 4.
Rotate by θ. The rotation of the reticle stage 2 is monitored by the laser interferometers 20, 21, 22 and the reticle stage 2 is rotationally driven without displacement in the X and Y directions. Therefore, the relative rotation error between each chip C on the wafer W and the projected image Pr ′ of the pattern of the second layer is corrected, and the superposition is performed accurately.

Next, in step 202, the CPU 30 projects the chip C and the projected image.
When the rotation error relative to Pr ′ is corrected, the count value of the θ counter 24 that detects the rotation amount of the reticle stage 2 is reset to zero. By this resetting operation, when the chip C and the projected image Pr ′ are accurately superimposed, θ
Both the count values from the counter 27 and the θ counter 24 become zero, and the θ counter 24 and the θ counter 27 are associated with each other, that is, the rotation reference of the reticle stage 2 and the rotation reference of the wafer stage 6. And CPU30
Wafer W for the second layer pattern using the and repeat method
In order to perform exposure by superimposing on each of the chips C above, the wafer stage 6 is next moved via the drive units 8 and 9, and the wafer W is set at a predetermined exposure start position.

Next, in step 203, the CPU 30 exposes the chips C on the wafer W and the projected image Pr ′ of the pattern of the second layer in an overlapping manner. Here, first, overlay exposure is performed on the first chip C on the wafer W. At this time, it is assumed that the first chip C and the projected image Pr 'do not have a relative rotation error (change in the count value of the θ counter 27) and are accurately superimposed and exposed.

Next, in step 204, the CPU 30 determines whether or not the second wafer circuit pattern has been overlaid and exposed on the entire wafer W. Since only the first chip C on the wafer W has been subjected to the overlay exposure here, the next step 205 is executed.

In step 205, the CPU 30 exposes the second chip C in a superimposed manner, so that the wafer stage 6 is moved by a predetermined amount along the X-axis or Y-axis of the coordinate system XY as a moving coordinate system by the step-and-repeat method. Only the driving unit 9 (or the driving unit 8) causes stepping.

When the wafer stage 6 is rotated due to this stepping and the wafer stage 6 rotates, the CPU 30
Is the rotation amount W of the wafer stage 6 in step 206.
Detect θ. The rotation amount Wθ is measured by the θ counter 27. Since the θ counter 27 is reset to zero in advance as described above, this count value corresponds to the rotation amount Wθ from the reference value of the wafer stage 6. With the above, the measurement of the rotation amount Wθ of the wafer stage 6 is completed, and the following step 207 is executed to correct this rotation deviation.

In step 207, the CPU 30 moves the reticle stage 2 by Wθ in the same direction as the rotation direction of the wafer stage 6 via the drive units 4 and 5 according to the rotation amount Wθ of the wafer stage 6 detected by the θ counter 27 in step 206. Rotate.
That is, the θ counter 24 of the reticle stage 2 is stepped.
Since it is reset to zero in 202 and is associated with the θ counter 27, the rotation amount Rθ of the reticle stage 2 detected by the θ counter 24 is detected to be equal to the θ counter 27, that is, the rotation amount Rθ is detected. Rθ = 0 to Rθ
The reticle stage 2 is rotated so that = Wθ.
The rotation of the reticle stage 2 is also caused by the laser interferometers 20, 21,
The reticle stage 2 is rotationally driven without being displaced in the X and Y directions by the monitor 22. As a result, the second chip C on the wafer W and the projected image Pr 'are accurately superposed, and then the CPU 30 executes step 203 again to perform superposition exposure.

As described above, in the overlay exposure of the second layer, the CPU 30 repeats the operations from step 203 to step 207, and corrects the rotation deviation of the wafer stage 6 by the rotation of the reticle stage 2 and the circuit pattern of the second layer. Is overlapped with each chip C on the wafer W and exposed by the step-and-repeat method. Therefore, each chip C on the wafer W and the projected image Pr 'are accurately superimposed, and the circuit pattern of the second layer is exposed due to yawing without lowering the overlay accuracy.

Also, for the exposure of the third and subsequent layers, the same operation as the exposure of the second layer, that is, the operations from step 200 to step 207 are performed. As a result, the rotational error between the overlay patterns due to the yawing of the wafer stage 6 is corrected, and the overlay exposure is performed accurately.

Thus, according to the present embodiment, both Abbe errors due to yawing during RR (reticle rotation) and CR (chip rotation) measurements are corrected based on the rotation amount of the wafer stage 6 detected by the θ counter 27. ,
Calculate RR and CR. Since the reticle stage 2 is rotated by Δθ based on these measured values, the influence of yawing accompanying all movements of the wafer stage 6 is prevented, and the θ counters 24 and 27 are accurately associated with each other. Therefore, a relative rotation error between the projected image Pr ′ of the circuit pattern of the reticle R by the yawing of the wafer stage 6 and each chip C arranged on the wafer W is Wθ in the same direction as the rotation direction of the wafer stage 6. Only the reticle stage 2 can be rotated and corrected. Further, the θ counter of the reticle stage 2 is based on only the data of the yawing measurement of the wafer stage 6 by the laser interferometer 13.
Reticle stage 2 so that 24 measures the same amount of rotation
Since only the rotation is performed, the relative rotation error between the projected image Pr ′ of the pattern of the reticle R and each chip C arranged on the wafer W is corrected without performing the alignment (mark detection operation) for each shot. It

As described above, in the embodiment of the present invention, the case has been described in which the circuit patterns of the second and subsequent layers are overlapped with the respective chips C on the wafer W and exposed. However, even when the method of the present invention is applied when the circuit pattern of the first layer is exposed on the wafer W by the step-and-repeat method, the circuit pattern of the first layer is stepped by the yawing of the wafer stage 6. It is obvious that the wafer W is prevented from being exposed by being rotated with respect to the coordinate system XY as the moving coordinate system by the and repeat method. That is, the relative rotational deviation of the reticle due to the reticle rotation with respect to the coordinate system XY in the exposure of the first layer is corrected by rotating the reticle stage 2 in the positive direction by Δθ in the same manner as described above. At this time, the rotation amount Δθ of the reticle stage 2
Δθ = −θr is because the chip rotation is zero (no chips are formed on the wafer W). Then, the wafer stage 6 is stepped by a predetermined amount to perform sequential exposure. The rotation amount Wθ of the wafer stage 6 due to yawing is detected by using the laser interferometer 13 and the θ counter 27, and the reticle stage 2 is rotated by Wθ in the same direction as the rotation direction of the wafer stage 6 according to the rotation amount Wθ. By this, the rotation deviation of the circuit pattern is corrected.

Further, the θ counter 24 and the θ counter 27 are each reset to zero, and after being associated with each other, the rotation amount Wθ of the wafer stage 6 is measured, and the relative rotation error between the reticle R and the wafer W due to yawing is corrected. Was there. However, instead of resetting each of the θ counters 24 and 27 to zero, the count value θ ₁ of the θ counter 24 after rotating the reticle stage 2 by Δθ and the count value θ ₁ of the θ counter 27 when the exposure apparatus is started.
Each of ₂ and ₂ is stored in the CPU 30 as an initial value. The same effect can be obtained even if the rotation amounts of the reticle stage 2 and the wafer stage 6 are calculated with reference to the initial values θ ₁ and θ ₂ and the relative rotation error between the reticle R and the wafer W is corrected. That is clear. Also, θ counter 27
The zero reset or the storage of the count value θ ₂ is not limited to the time when the exposure apparatus is started, but is not limited to the time when the exposure apparatus is started.
It should be done before the R and CR measurements are taken.
Similarly, the zero reset of the θ counter 24 or the storage of the count value θ ₁ may be performed after the completion of the correction of RR and CR and before the execution of the overlay exposure of the first shot position.

In this embodiment, the LSA system 17a is used as in step 201, the CR measurement is performed by correcting the Abbe error, and the CRs of the plurality of chips C located near the center of the wafer W are averaged. The amount of rotation θc of the upper chip C with respect to the θ axis was obtained. However, the method of measuring the rotation amount θc used in this embodiment is not limited to the above method. For example,
The LSA system 17a detects CRs of a plurality of arbitrary chips C located near the center and near the periphery of the wafer W, corrects Abbe's error, and uses a statistical method, for example, the least squares method to determine the optimum rotation amount. It is obvious that the same effect can be obtained by obtaining θc.

〔The invention's effect〕

As described above, according to the present invention, even if the wafer stage rotation shift occurs due to the wafer stage stepping, the circuit pattern on the wafer and the reticle circuit in the overlay exposure of the second and subsequent layers due to this influence. Rotation error relative to the pattern, and the first arrayed on the wafer
Coordinate system XY as the moving coordinate system of the wafer stage 6 by the step-and-repeat method of the circuit pattern of the layer
There is an advantage in that the rotation deviation with respect to can be prevented, the exposure can be performed with high accuracy and in a short time according to the movement of the wafer stage, and the deterioration of the throughput can be prevented.

[Brief description of drawings]

FIG. 1 is a perspective view showing a schematic configuration of a system suitable for carrying out an alignment method according to one embodiment of the present invention, and FIG. 2 is a control of an apparatus to which the alignment method according to one embodiment of the present invention is applied. FIG. 3 is a block diagram of the system, FIG. 3 is a flow chart for explaining the operation of this embodiment, FIG.
5 and 6 are diagrams for explaining the measurement operation of the reticle rotation, and FIG. 7 is for explaining the correction of the relative rotation error between the tip C and the reticle R due to the tip rotation and the reticle rotation. It is a figure to offer. 1 ... Projection lens, 2 ... Reticle stage, 3, 4,
5 ... Reticle stage drive unit, 6 ... Wafer stage, 10, 11, 12, 20, 21, 22 ... Laser interferometer, 16a, 1
6b …… die-by-die alignment system, 17a, 17b ……
Laser step alignment system, 24, 27 …… θ counter, 30 …… Processor (CPU), R …… Reticle,
W ... wafer.

Claims (9)

(57) [Claims] 1. An alignment method used in an apparatus for exposing each of a plurality of shot areas on a photosensitive substrate with an image of a pattern on the mask, wherein a relative rotational displacement of the mask with respect to the photosensitive substrate is predetermined. A first step of setting the mask so as to be within an allowable value of; a first rotation amount detection means for detecting a rotation amount of the mask from a predetermined first reference position; and a predetermined first rotation amount of the photosensitive substrate. Two
The second detection result obtained by correlating both detection results with the second rotation amount detection means for detecting the rotation amount from the reference position after the first step.
And a step of detecting the detection result of the second rotation amount detection means associated with the first rotation amount detection means in the second step when exposing each of the plurality of shot areas with the image of the pattern. And a third step of rotating the mask so as to detect an amount of rotation substantially equal to. 2. The first rotation amount detecting means and the second rotation amount detecting means.
2. The method according to claim 1, wherein at least one of the rotation amount detecting means in the above step uses an interferometer. 3. The first rotation amount detecting means includes a first counting circuit for measuring the rotation amount of the mask from the first reference position by an increment method, and the second rotation amount detecting means, A second counting circuit for measuring an amount of rotation of the photosensitive substrate from the second reference position by an increment method;
The second counting circuit is reset to a second predetermined value before the first step, the first counting circuit is reset to a first predetermined value after the first step, and in the third step, 2. The method according to claim 1, wherein the mask is rotated so that the change in the count value by the second counting circuit and the change in the count value by the first counting circuit are substantially equal to each other. 4. In the first step, after detecting a relative amount of rotation deviation between a shot area on the photosensitive substrate and the mask, the mask is rotated until the amount of rotation deviation becomes substantially zero. 4. The method according to claim 3, wherein in the second step, the first counting circuit is reset to the first predetermined value. 5. An alignment apparatus used in an apparatus for exposing each of a plurality of shot areas on a photosensitive substrate with an image of a pattern on a mask, the alignment apparatus having a mark detection system for detecting the position of a mark on the photosensitive substrate. First rotation amount detecting means for detecting an amount of rotation of the mask from a predetermined first reference position; second rotation amount detecting means for detecting a rotation amount of the photosensitive substrate from a predetermined second reference position And; correcting the position of the mark based on the output of the second rotation amount detecting means when the mark is detected by the mark detecting system, and shot on the photosensitive substrate based on the corrected position. Rotation deviation amount detecting means for detecting a relative rotation deviation amount between an area and the mask; stepping the photosensitive substrate to expose each of the plurality of shot areas with the image of the pattern. In order to illuminate, the mask is removed using the first rotation amount detecting means based on the detected rotation deviation amount and the output of the second rotation amount detecting means when the photosensitive substrate is stepped. An alignment apparatus comprising: a drive unit for rotating the alignment apparatus. 6. The first rotation amount detecting means and the second rotation amount detecting means.
6. The method according to claim 5, wherein at least one of the rotation amount detecting means in step 1 uses an interferometer. 7. The driving means rotates the mask so that the rotation deviation amount becomes substantially zero before the exposure, and at the time of the exposure, the driving means is operated based on an output of the second rotation amount detecting means. Device according to claim 5, characterized in that the mask is rotated. 8. The detection results of both the first rotation amount detecting means and the second rotation amount detecting means are associated with each other after the mask is rotated so that the rotation deviation amount becomes substantially zero. The driving unit rotates the mask so that the first rotation amount detecting unit detects a rotation amount substantially equal to a detection result of the associated second rotation amount detecting unit. The device according to claim 7. 9. The first rotation amount detecting means and the second rotation amount detecting means each have a counting circuit for counting the rotation amount, and the first rotation amount detecting means is the second rotation amount detecting means. 9. The device according to claim 8, wherein the counting circuit is reset to a predetermined value so as to be associated with the rotation amount detecting means.