JP3204248B2 - Exposure method, circuit pattern body manufacturing method using the exposure method, or exposure apparatus - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an exposure method or an exposure apparatus used for manufacturing semiconductor devices, liquid crystal display devices, printed circuit boards, and the like.

[0002]

2. Description of the Related Art In this type of exposure apparatus, an object to be exposed, such as a semiconductor wafer, a glass plate, or an original wiring pattern cell, is mounted on a large movable stage, and an image or shadow of a pattern to be exposed is formed. It is positioned and exposed. In a step-and-repeat type (or step-and-scan type) exposure apparatus that sequentially exposes a pattern at a plurality of positions on a wafer or a glass plate, a so-called stepper, the positioning accuracy of a substrate to be exposed is less than submicron. Orders are required, and high speed is required to increase the productivity of device manufacturing.

In general, a stepper is provided with a projection optical system for projecting a mask pattern called a reticle onto a photosensitive substrate (wafer) at a fixed magnification.
This projection optical system has an extremely high resolution and is composed only of a refraction system (lens), one composed only of a reflection system (mirror), or one composed of a combination of a refraction system and a reflection system. , There are various types. Also, the projection magnification may have various values (× 1, × 1/2, × 1 /
4, × 1/5, × 1/10, etc.).

Further, there are various field sizes which can be projected at one time. Generally, the field size in a projection optical system for a wafer stepper is about 10 mm square to 30 mm square. However, it is smaller than the wafer size (5 inches, 6 inches, 8 inches, etc.).
Therefore, in a wafer stepper or the like, the reticle pattern is exposed on the entire surface of the wafer by repeatedly moving the movable stage on which the wafer is mounted by stepping (or scanning) a fixed distance and projecting and exposing the reticle pattern. At this time, the relative alignment between one shot area (exposed area) on the wafer and the projected image of the reticle pattern by the projection optical system is confirmed by various alignment systems, and the alignment is achieved. The movable stage is positioned two-dimensionally. Normally, the coordinate position of the movable stage is measured successively by a light wave interferometer (laser interferometer), and the deviation between the target position confirmed by the alignment system and the position measured by the laser interferometer (current position) is minimized. The servo control of the stage is performed so as to be zero.

[0005]

In the conventional apparatus as described above, in order to maintain high positioning accuracy, it is necessary to move a heavy movable stage with increased rigidity at a large acceleration. However, the fact that a heavy object moves in the exposure apparatus means that unnecessary vibration (sway) or deformation occurs in the entire apparatus or a part of the apparatus. This vibration can be suppressed to a fairly small value if sufficient rigidity is obtained in the structure of the exposure apparatus. However, if the vibration is coped with, the self-weight of the apparatus becomes many times as large as the current one, and the practicality is poor. For this reason, the conventional apparatus detects that the measured value of the laser interferometer for the movable stage has stayed within a permissible range centered on the target value for a certain minute time, and converges the vibration (shake) ( Was settled).

However, considering the positions where the fixed mirror and the movable mirror of the laser interferometer are mounted, the vibrations and fluctuations are not always converged in all parts in the exposure apparatus. Therefore, even if the image of the reticle pattern is projected and exposed after the fluctuation of the measurement value of the laser interferometer is converged by the conventional method, the relative positions of the reticle, the projection optical system, and the wafer stage during the exposure time. Slightly shifted in the lateral direction (or the optical axis direction). Therefore, each exposure position of a plurality of shot regions arranged on a wafer in a step-and-repeat manner has an error with respect to a design position or a position determined by an alignment result, that is, an arrangement error or a stepping error. Transfer position error occurred. As described above, when the pattern projection image and the wafer stage are slightly shifted during the exposure time, depending on the amount of the shift, the image contrast on the wafer to achieve fine image formation, or strictly speaking, the resist layer on the wafer. The cumulative contrast in the process of forming a latent image is reduced, which may have a significant effect on the resolution. Such deterioration of the resolving power can be considered in the same manner as a camera shake at the time of shutter release in a general photographic camera, an image shake at the time of photographing a fast-moving subject, and the like.

In the case where the measured values of the laser interferometer have convergence, disturbances (vibration of the floor of a factory, movement of an operator or a self-propelled robot in a factory, external devices, If the measured value deviates from the convergence range due to the operation of the exposure, the exposure may be interrupted. However, it is possible that the laser interferometer readings will not change from the convergence range if disturbances are added to the stepper and cause a general sway or movement of other structures within the stepper. .

Therefore, the present invention makes it possible to transfer a pattern with a high resolution in a state where there is no image blurring even if the exposure apparatus is totally or locally deformed, and to finely shift the exposure position. It is an object of the present invention to provide an exposure apparatus in which an alignment error, a stepping error and the like due to the above are significantly reduced.

[0009]

Therefore, in the exposure method or exposure apparatus of the present invention, when exposing the pattern of the mask (reticle) onto the substrate (3), the relative alignment between the mask and the substrate is determined. Detecting means (S1 to S7) for detecting a stress which causes deterioration (deterioration) or a stress or deformation is provided at at least one portion of a portion where a deformation on the device structure occurs, and further, based on an output signal of the detecting means. Exposure is performed in a state where the influence of the stress or the deformation is removed.

More specifically, as described in an embodiment to be described later, the timing of exposure operation (start or interruption) is controlled to wait for exposure based on the detection of stress or deformation.

[0011]

According to the present invention, the exposure is carried out in a state where the influence of the stress or the deformation is removed by using the sensor output for detecting the stress or the deformation of the portion in the apparatus which has an adverse effect on the exposure position. Exposure can be performed without deterioration (deterioration) of the relative position of the pattern to be exposed, the mask, and the substrate.

More specifically, since the timing (start or interruption) of the exposure operation is controlled based on the detection of stress or deformation to control the exposure waiting time, a predetermined waiting time is set in anticipation in advance. Unnecessary waiting operation caused by the above operation is eliminated, and a decrease in throughput can be prevented.

Further, when the position shift of the shot is predicted due to the deformation due to the disturbance during the exposure, the exposure is interrupted by closing the shutter or the like, so that an apparatus having a structure substantially resistant to the disturbance can be obtained.

[0014]

FIG. 1 is a view showing a schematic configuration of an exposure apparatus (wafer stepper) to which the present invention is applied. The reticle 1 patterned with a chromium layer or the like is placed on a reticle stage that moves two-dimensionally by about several mm on the reticle-side column 1A. The pattern of the reticle 1 is image-formed and projected on the wafer 3 by the projection lens 2. The projection lens 2 is fixed to the lens-side column 2A via a flange, and the distance between the reticle 1 and the projection lens 2 and the distance between the projection lens 2 and the wafer 3 are set to constant dimensions by the columns 1A and 2A.

Illumination of the reticle 1 is performed by an illumination system 4, and light from a light source (such as a mercury lamp) in the lamp house 4A is made uniform by various optical components (such as a fly-eye lens) in the illumination system 4. The illumination is distributed and reaches the reticle 1 via the main condenser lens 4C. The lamp house 4A is provided with an exhaust hose 4B.
This illumination system 4 is a column 4E built on the column 2A.
And is fixed at a predetermined height position. The column 2A holding the projection lens 2 is fixed on the surface plate 8,
Is mounted on a base 10 via an anti-vibration suspension 9, and the base 10 is fixed on a floor 11 in a factory. Generally, a combination of the base 10 and the suspension 9 is called an anti-vibration table.

On the surface plate 8, an X stage 5 on which a holder for vacuum-sucking the wafer 3 is mounted, and a Y stage 6 on which the X stage 5 is mounted and moves on the base 7 in the Y direction. Provided. In FIG. 1, the X stage 5 moves one-dimensionally in the left-right direction, and is driven by rotating a feed screw Sc by a screw rotation motor M0 fixed to the Y stage 6. The Y stage 6 is also provided with a feed screw and a motor, and moves in a direction perpendicular to the plane of FIG.

The laser interferometer L has two length-measuring laser beams L 1, L 2 from the X-axis direction perpendicular to the optical axis of the projection lens 2.
L2 is emitted, and the beam L1 is perpendicularly incident on a movable mirror M1 fixed on the periphery of the X stage 5 (actually, the Z stage thereon). The beam L2 is incident on the fixed mirror (reference mirror) M2 fixed to the lower end of the lens barrel of the projection lens 2 perpendicularly from the X-axis direction. The laser interferometer L causes a beam reflected by the fixed mirror M2 to return and a beam reflected by the movable mirror M1 to interfere with each other, and photoelectrically detects a change in brightness of interference fringes caused by the interference. Moving mirror M
1, that is, the position information (or movement amount) Se of the wafer 3 in the X-axis direction is output.

The position of the wafer W due to the movement of the Y stage 6 is detected by a laser interferometer having the same configuration. The stage is driven by a servo amplifier (differential circuit) Di.
After setting the target position value S0 to the servo amplifier Di, the position information Se of the laser interferometer L is added as a deviation signal to the servo amplifier Di. As a result, the servo amplifier Di drives the drive signal (voltage) S according to the distance between the current position of the stage and the target position.
Is output to the motor M0.

By the way, a part of the surface plate 8 has a mounting portion 13.
A wafer loader mechanism 13 is provided via the terminal c. The wafer loader mechanism 13 includes a cassette holding unit 13B for storing a cassette containing a plurality of wafers, and a transfer arm 13A for transferring one wafer taken out of the cassette to a holder on the X stage 5. Are provided. On the other hand, a reticle loader mechanism 12 is provided on a part of the lens-side column 2A via a mounting portion 12C. The reticle loader mechanism includes a reticle library section 12B for storing a plurality of cases for storing reticles used for exposure one by one, and a transfer arm 12A for delivering the selected reticle to a reticle stage. Is provided.

While the wafer loader mechanism 13 is exposing one wafer, the wafer loader mechanism 13 may continue to operate for returning the exposed wafer to the cassette, preparing a new wafer, and the like. Also, in the reticle loader mechanism 12, after the reticle exchange is completed and the exposure processing on the wafer is started, the motors and movable parts of the respective parts may operate in preparation for the next reticle.

In the typical stepper constructed as described above, the measurement reference of the interferometer L is a fixed mirror M 2 fixed to the lower end of the projection lens 2. This means that the interferometer L is assembled by selecting the position where the lower end side of the projection lens 2 is closest to the pattern projection image of the reticle 1 and where the image shift is most likely to occur. Such a configuration is well known in the art. Certainly, the interferometer L has a high ability to detect a deviation from a relative alignment position between the pattern projection image and the wafer 3, but does not monitor a positional deviation due to vibration of the pattern projection image itself. For this reason, if the position of the fixed mirror M2 of the projection lens 2 is not displaced but is displaced toward the reticle 1, it is determined from the measured value of the interferometer L that the position has been settled. Or image shift.

The apparatus as shown in FIG. 1 includes various moving parts, and various kinds of parts are from large to small. Further, this type of exposure apparatus is housed in an environmental chamber except for the lamp house 4A. The chamber has a vibration source such as a compressor blower to keep the environmental conditions (temperature, humidity, cleanliness, etc.) of the exposure apparatus constant. Among the movable parts inside the stepper and the vibration sources outside the stepper, the factors that greatly affect at the time of exposure are the X stage 5 and the Y stage 6.
Exercise. The stages 5 and 6 are the heaviest of the movable parts and move at a high speed in a stepping movement, so that the acceleration becomes high, and the reaction thereof causes the illumination system 4, the holding part (reticle stage) of the reticle 1, the projection lens 2, and the reticle loader. 12, the wafer loader 13 and the like are vibrated. These structures are all columns 1A, 2A, 4E
However, the vibration (sway) and convergence of the entire system change due to the rigidity and natural vibration of each column and structure.

As described above, the reticle loader 12 and the wafer loader 13 may operate independently of the exposure operation in order to improve the throughput. Furthermore, the vibration of the floor 11 generated by the movement of the worker around the stepper or the movement of the self-propelled robot in the factory may be transmitted to the apparatus main body via the vibration isolator (9, 10). When the self-propelled robot automatically changes the wafer cassette in the wafer loader 13,
Unnecessary force may be applied to the cassette holding unit 13B to cause the wafer loader 13 to shake. In particular, floor vibrations having a long period that are stationary can be absorbed by the vibration isolating tables (9, 10), but cannot be sufficiently absorbed by the pulse-shaped vibrating vibrations and can be absorbed by the apparatus main body. It will be transmitted.

As described above, there are many vibration sources in the environment surrounding the stepper, and it is important to isolate these vibration sources from the stepper body as much as possible. As one of the measures, the reticle loader 12 and the wafer loader 13 are separated from the column 2A and the surface plate 8 of the stepper body,
Holding directly from the floor 11 is also conceivable. in this case,
Since the stepper body may be shaken as a whole on the anti-vibration tables (9, 10), there is a problem in positioning the transfer arms 12A, 13A when transferring the reticle or wafer. For example, when a wafer on the transfer arm 13A is transferred to a holder on the X stage 5, a mechanical configuration such that the wafer on the arm 13 and the holder always have a fixed positional relationship regardless of the swing of the entire stepper, Position correction and the like are required.

However, the portion of the vibration source which must move in the stepper body, especially the wafer stage, cannot be naturally separated from the stepper body. The movement of the wafer stage is the largest factor causing vibration and structural deformation of each part in the stepper main body, and is extremely close to the exposure operation in terms of time. Since the position of the center of gravity of the wafer stage changes with each alignment operation and exposure operation, the movement start position, stop position, movement speed (acceleration), etc. of the stage are obtained to analyze the local part fluctuation in the stepper body. It is possible to predict image shift and image blur, but for that purpose, it is necessary to perform structural analysis computer simulation based on huge data on rigidity and natural vibration of each part in the stepper body .

Accordingly, in a specific embodiment of the present invention, a vibration sensor, a vibration sensor, is provided at a portion in a stepper that causes vibration (sway) or deformation that deteriorates the relative alignment accuracy between the reticle 1 and the wafer 3 during exposure. A detector such as an accelerometer, a strain gauge, or a small interferometer (such as a fiber interferometer) is provided, and the start, interruption, and restart of the exposure operation are controlled according to the detection signal. In another embodiment, the exposure control is performed by monitoring both the signal from the vibration or deformation detector and the expected settling time of the stage. In still another embodiment, a temporal change characteristic (period, change rate, etc.) of a transfer position error at the time of pattern exposure is determined from a detection signal of vibration (shake) or deformation, and an appropriate exposure is performed according to the change characteristic. The exposure time was varied without changing the amount.

FIG. 2 is a block diagram showing a configuration of a signal processing system and a control system commonly used in each embodiment. The same members as those shown in FIG. 1 are denoted by the same reference numerals.
A light source ILS such as a mercury lamp is provided in the lamp house 4A of the illumination system 4 shown in FIG. The light source ILS is driven by the light source control system 20 to give a predetermined lighting power or a predetermined light emission luminance. The illumination light from the light source ILS passes through an interference filter or the like (not shown), and only a specific emission spectrum (g-line, i-line, etc.) for exposure is selected, and a fly-eye lens (optical integrator) for uniforming the illuminance distribution. It is incident on 4G. Fly eye lens 4
Light from a number of secondary light sources formed on the emission side of G enters the main condenser lens 4C via the lens system 4F,
The reticle 1 is illuminated with a uniform illuminance distribution.

A shutter ST provided between the light source ILS and the fly-eye lens 4G is driven by a shutter control system 22 to open and close an optical path. The shutter control system 22 includes a circuit that integrates the light amount based on an output signal from a photoelectric detector 23 that detects the light amount of a part of the illumination light that has passed through the shutter ST, and that the integrated value is a target value ( And a circuit for outputting a closing signal for closing the shutter ST when the exposure time (equivalent to an appropriate exposure amount) is reached. A command to open the shutter ST is sent from the main control system 30. The output signal from the photoelectric detector 23 is also sent to the light source control system 20, and is used for stabilizing the light emission luminance of the light source ILS. Further, the photoelectric detector 2 is also provided at a position where the intensity of illumination light from the light source ILS is always detected independently of the shutter ST.
The output signal is sent to the light source control system 20 and used for power control of the light source ILS.

Although not shown in FIG. 1, a TTL alignment system ALG for detecting an alignment mark on the wafer 3 via the projection lens 2 is provided in FIG. The mark position information detected by the alignment system ALG is sent to the main control system 30, and the EGA (enhanced global alignment) of the wafer 3,
It is used to calculate the coordinates of the stepping position of the wafer stage according to an algorithm such as F / F (field-by-field). The calculated coordinate value of the stepping position is sent as a target value to the stage control system 24 including the differential circuit Di shown in FIG. 1, and drives the motor M0 as described above. The stage control system 24 has a function of calculating the amount of deviation between the calculated coordinate value of the stepping position and the current position of the stage measured by the interferometer L. The information ΔPS of the amount of deviation is used for image shift analysis. To the computer 25.
In the conventional apparatus, when the deviation .DELTA.PS falls within a predetermined allowable range and the servo drive of the wafer stage is completed, the stage control system 24 sends a signal indicating that exposure is enabled to the main control system 30. Was. However, in the present embodiment, whether or not exposure is possible is determined by the image shift analysis computer 25, including confirmation that the deviation amount ΔPS has fallen within the allowable range, and when exposure is possible, a signal indicating that fact is sent from the computer 25. EX is sent to the main control system 30.

The image shift analysis calculator 25 calculates a substantial position shift (image shift) between the pattern image projected on the wafer 3 and the wafer 3 among the vibrating portion in the stepper or the structurally deformed portion. , Image blur) are input from n sensors S1 to Sn provided in some of the portions that cause image blurring. In FIG. 2, seven sensors S1 to S7 are prepared, and the sensor S1 is a fly-eye lens 4G.
(Vibration) in the direction orthogonal to the optical axis of the illumination system, or deformation, and the fluctuation or deformation related to the tilt direction of the fly-eye lens 4G from the optical axis are detected, and the magnitude of the vibration (vibration) or deformation is detected. An output signal corresponding to the output signal is generated. Sensor S
2 generates an output signal according to the magnitude of the vibration (fluctuation) and deformation of the main condenser lens 4C, and the sensor S3 produces the displacement and displacement of the reticle stage RS in the X, Y, and Z directions (and the rotation direction). An output signal is generated according to the magnitude of. The sensor S4 generates an output signal corresponding to the magnitude of vibration (fluctuation) of the lens barrel on the upper end side (reticle side) of the projection lens 2. In addition, the sensor S5 outputs a signal corresponding to the magnitude of the vibration (shake) of the lens-side column 2A, and the sensor S6 generates the vibration generated by the reticle loader 12, or the mounting portion 12 supporting the reticle loader 12.
The sensor S7 generates a signal corresponding to a change (strain) in the stress applied to C, and the sensor S7 detects the vibration generated in the wafer loader 13 or the mounting portion 1 supporting the wafer loader 13.
A signal corresponding to a change (strain) in stress applied to 3C is generated.

The above arrangement of the sensors S1 to S7 is an example, and if there is a portion (a chamber or the like) that can be considered as a vibration source, it is preferable to provide a sensor there.
Further, each of the sensors S1 to S7 is not necessarily of the same type, and various types are appropriately selected according to the physical quantity to be detected. For example, when dealing with relatively large vibrations, a micro-position sensor or the like is preferable. When even a minute vibration (micron order) directly affects image shift, a small interferometer (using a semiconductor laser) is used. Good. When detecting structural deformation, a strain gauge (strain gauge) or an interferometer is used.

The relative positional deviation between the reticle 1 and the wafer 3 during exposure is caused by a mark on the periphery of the reticle 1 and a mark on the periphery of the shot area on the wafer 3 projected on the projection lens 2.
Can be detected by using a TTR (through-the-reticle) type alignment system, which detects simultaneously through the reticle. The TTR type alignment system currently in practical use has the ability to relatively quickly detect the relative displacement between the reticle mark and the wafer mark. However, the image shift does not always occur slowly during the exposure time, but may occur rapidly due to release of stress accumulated in a structure in the apparatus. TTR when image shift occurs slowly
It is also possible to deal with this by sequentially detecting the amount of deviation by the alignment system and forming a loop for correcting the position of the reticle stage RS or the wafer stage by feedback control.

Next, a first embodiment of the present invention will be described. In the first embodiment, the image shift analysis computer 2
5 sequentially inputs signals from the sensors Sn at the time of stepping, and determines the projection lens 2 based on the overall correlation of the signals.
, A relative position shift characteristic between an image to be projected on the wafer 3 and a shot area on the wafer 3 is calculated in real time, and the position shift characteristic and the laser interferometer L are calculated.
The exposure to the shot area is started or interrupted based on two of the positioning characteristics (static characteristics) of the wafer stage 5 measured by the method, or only the static characteristics of the wafer stage 5. .

FIG. 3A shows a change characteristic (static settling characteristic) of the measurement value of the laser interferometer L for the X direction. The horizontal axis indicates time t, and the vertical axis indicates the measurement position xe. In addition, 0 of the measurement position xe
The dots represent the stepping target position of the wafer stage 5 in the X direction, and ± xs represents the allowable range for the target position. In FIG. 3A, the wafer stage 5 has a characteristic that after moving at high speed toward the target position, decelerates and converges within the allowable range ± xs. The stage control system 24 outputs the deviation amount ΔPS from the target position to the computer 25. In this case, the allowable range of the deviation amount ΔPS is ± xs.
In the mode using only the measured value of the laser interferometer L (the mode using only the static settling characteristic of the wafer stage), the timing is started from the moment the deviation ΔPS is within the allowable range ± xs, Hardware or software that resets the current time to zero when the value deviates from the allowable range ± xs. The counted time tC is set to a predetermined waiting time (for example, 0.1 to 0.5).
(Seconds) or more, the computer 25 issues a command EX to open the shutter ST for the exposure time T0 as shown in FIG.
To the main control system 30.

When each shot area on the wafer 3 is exposed by a step-and-repeat method (or a step-scan method), the center of gravity of the weight of the entire wafer stage 5 moves sequentially in the XY plane on the surface plate 8. Will be done. For this reason, the specific characteristics shown in FIG.
It does not become the same for all of the above shot areas, but slightly changes. That is, the convergence characteristics (amplitude and time) of FIG. 3A change according to the arrangement of the shot areas to be exposed on the wafer 3. Then, the image shift analysis computer 25 receives the array coordinate values of the next shot area stored in the main control system 30 and sets the waiting time tc required for stepping to the next shot area to a more optimal waiting time tc. Process to change to something. In determining the waiting time tc, the wafer stage is stepped by a predetermined distance, and the characteristics as shown in FIG. 3A are fetched into a waveform memory or the like, and are compared with each other in correspondence with the coordinate position at each stepping. It is performed by an experiment or the like. Thus, according to the coordinate value of the shot area on the wafer 3,
By setting the time to wait tc shortly after stepping to be short or long, it is possible to expose the entire shot area in a high resolution state without image blurring while suppressing an extremely low throughput. .

It should be noted that the change of the waiting time tc does not actually need to correspond one-to-one with each arrangement of the shot areas on the wafer 3, and the dimensions of the wafer (5 inches, 6 inches, 8 inches, etc.) are changed. In consideration of this, an appropriate matrix (for example, 30 pixels independent of the size of the shot area) is taken into consideration.
(mm square), and the same waiting time tc may be given to a plurality of shot areas whose centers are present in one section.

As described above, the image shift analysis computer 2
5 sets the optimum waiting time tc internally, and when the waiting time tc elapses, internally sets the signal S as shown in FIG.
H1 is set to logic “1” to control the shutter ST to open. Now, when the image shift analysis computer 25 is used in a mode for analyzing the position shift characteristics based on the information from the various sensors S1 to S7, the computer 25 converts the input information from each of the sensors S1 to S7 into a neuro operation or the like. By processing, the relative position shift characteristics between the projected image and the wafer 3 are obtained almost in real time as shown in FIG. 3C from the state of vibration and stress detected by each of the sensors S1 to S7. FIG.
In FIG. 3C, the vertical axis represents the position shift amount xi generated on the wafer, and its allowable range is ± x as in FIG.
s. As shown in FIG. 3C, the position shift characteristic does not always have convergence, and
The static characteristics of the wafer stage shown in FIG. FIG. 3 (C) is shown together with the time axis of FIG. 3 (A).
Even if it is within s, it may not be within the allowable range ± xs on the position shift characteristic obtained by the neuro operation. Therefore, the computer 25 starts time measurement when the position shift characteristic is within the allowable range ± xs, and
After the state of FIG. 3 has continued for a predetermined waiting time td1, FIG.
, The internal signal SH2 is set to logic “1”. However, after the time T1 elapses, the position shift amount xi becomes within the allowable range ±
As soon as it deviates from xs, the signal SH2 is set to logic "0" immediately. Further, when the waiting time td2 elapses after the position shift amount xi falls within the allowable range ± xs, the signal SH2 is again inverted to logic "1". So Calculator 2
5 is the signal SH1 in FIG. 3B and the signal SH in FIG.
The logical product (AND) with 2 is calculated, and the logical product is "1"
The shutter ST is opened and the exposure is performed only during the period. In this case, the shutter control system 22 in FIG. 2 may be operated in the light amount integration mode or in the timer mode. In the timer mode, a time is measured only when the shutter ST is open, and when the time reaches the target time, exposure of one shot area is completed.

By the way, image blur due to vibration and stress is as follows.
Generally, it is likely to occur immediately after the completion of the stepping of the wafer stage. Therefore, first, the signal SH1 is set to logic “1”.
It is also possible to check whether or not the calculation has been made, and from the time when the logic becomes "1", the calculator 25 may start the neuro calculation based on the input information of each of the sensors S1 to S7. FIG.
In (C) and (D), after the signal SH2 becomes "1" for the first time after the end of the stepping, since the position shift amount xi slightly deviates from the allowable range ± xs, the exposure is interrupted during this period.

However, in the state where the signal SH2 becomes "1" for the first time and the exposure for one shot area is started (time T1), it is assumed that the image blur due to the vibration and the stress is extremely small. Time T1
The waiting time td2 after elapse may be set shorter than the initial waiting time td1. When the light source ILS shown in FIG. 2 is a pulse light source such as an excimer laser, the light source control system 2
Since there is an oscillation trigger circuit in 0, the signals SH1 and S
Only when the logical product with H2 is true ("1"),
A plurality of pulse oscillation commands may be given to the oscillation trigger circuit.

Since the mechanical shutter SH usually has a constant operation delay (response delay), the signals SH1 and S
It is difficult to close the shutter momentarily when the AND output of H2 becomes "0". Therefore, as a mitigation measure against this, the position shift amount xi must be within the allowable range ± xs
When the position shift amount xi is within the range ± xs, the range ± xs may be narrowed so that the deviation from the inside to the outside of the range is slightly judged.

Conversely, as a technique for reducing the response delay of the shutter, a double blade shutter may be used as shown in FIG. In this type of exposure apparatus, a metal plate or the like is used for the shutter in order to suppress heat generation when blocking high-luminance illumination light from the light source ILS. Such a metal plate has a heavy weight, and when it is configured as a rotary shutter, there is a limit in suppressing its rotary inertia to a small value. Therefore, a sub-shutter, which has poor durability over a long period of time with respect to illumination light but has a reduced speed, is combined with a conventional shutter.

In FIG. 4, the shutter ST1 is at 45 °
This is a conventional rotary shutter in which minute blades are arranged in a circumferential circle every 45 °, and the illumination light IL is in a state of passing through a cutout portion of the rotary shutter. Conventionally, the illumination light IL is blocked only by rotating the shutter ST1 by 45 ° from the state shown in FIG. In such a rotary shutter, transmission and cutoff of the illumination light IL are switched only by rotating the shutter ST1 in one direction. In the present embodiment, the sub-shutter ST2 having a high-speed response is provided in the optical path of the illumination light IL. This sub-shutter ST2 may be equivalent to a focal plane shutter made of a titanium thin plate (0.2 to 0.5 mm in thickness) used in a photographic camera or the like. This type of focal plane shutter has a very short time (operating time) until the shutter is fully opened or fully closed, and is 10 to 5 msec.
It is about. This value corresponds to 1/2 to 1/4 of the operation time of the rotary shutter ST1.

However, since the sub-shutter ST2 has poor durability, the timing of the drive section of each shutter is controlled so as to have an interlocking relationship with the main shutter ST1. First, from the shutter open state of FIG.
When the shutter L is shut off, the start of the closing operation of the main shutter ST1 and the start of the closing operation of the sub-shutter ST2 coincide with each other in terms of timing, or the sub-shutter ST
Immediately after the closing operation of Step 2 is completed, the main shutter ST1
The timing is determined so that is closed.

When the shutter is opened, the main shutter ST1 is opened earlier and then the sub-shutter S is opened.
Open T2. Therefore, the shutter control system includes hardware or software that satisfies such a condition. Also, in the case of FIG.
Since S is on the side of the main shutter ST1, if the sub-shutter ST2 is closed when the shutter ST1 is opened, the blade of the sub-shutter ST2 is irradiated with the illumination light IL, and reflected light is generated therefrom. . Therefore, it is preferable to provide a photoelectric sensor that receives the reflected light, and establish a system that automatically opens the sub-shutter ST2 when the output signal level of the photoelectric sensor increases. The system can be used not only for the shutter opening operation at the start of exposure of the main shutter ST1 and the sub-shutter ST2, but also for the sub-shutter ST2 when the main shutter ST1 is opened halfway due to malfunction.
Can be prevented from being damaged by the illumination light IL.

By the way, in the state as shown in FIGS. 3C and 3D, when the vibration or the stress distortion is settled, the exposure operation may not be substantially started. This is because image blurring (small amplitude) occurs in a relatively short cycle around the permissible range ± xs due to a vibration source or the like in or near the apparatus. Therefore, as another embodiment, when the static setting characteristic of the wafer stage is detected as shown in FIG. 3 (B) and the signal SH1 becomes “1”, the amplitude and cycle of the vibration are rapidly increased from the change in the position shift characteristic. If it is determined that the vibration is a periodic vibration even though it slightly exceeds the allowable range ± xs, the computer 25 sends the light source IL to the light source control system 20 in FIG.
A command to lower the brightness of S by a certain amount, for example, light source I
A command to lower the power supply to the LS is output, and the shutter ST
(Or ST1, ST2) is output.

When the exposure amount is controlled in the light amount integration mode, a decrease in the luminance of the light source ILS increases the exposure time for obtaining an appropriate exposure amount. The long-time exposure is set so that one or more cycles of the vibration of the image blur exist. That is, the exposure time is set so as to include at least one cycle (preferably at least two cycles) of the vibration of the image blur, and the luminance of the light source ILS is reduced so as to obtain the exposure time. Such a method can be similarly used when the exposure amount control is performed in the timer mode.

The decrease in the luminance of the illumination light is reduced by the dimming means (N
A D filter, a mesh filter, etc.) may be inserted into the illumination light path. In this way, even if image blurring occurs, the position of the pattern formed on the resist layer of the wafer 3 in the XY plane has a high probability of being located at the center point of the vibration. There is no extreme deviation.

As described above with reference to FIG. 3, when determining the stabilization of the wafer stage from the reading value of the laser interferometer, depending on the relationship between the structure of the wafer stage and the stepping direction, the waiting time tc or td1 and td2 may be set to different values. For example, in a structure in which the X stage 5 is mounted on the Y stage 6 as shown in FIG. 1, the mass of the movable body when the Y stage 6 moves is the same as when the X stage 5 moves. Naturally, it becomes larger than the mass of the movable body. Therefore, when only the X stage 5 moves and when only the Y stage 6 moves,
The degree of vibration and blur may be different. Therefore, the waiting time tc shown in FIG. 3A or the waiting time tc shown in FIG. 3C when the wafer 3 moves in the X direction and when it moves in the Y direction by the step-and-repeat method or the step-and-scan method. It is desirable that the waiting times td1 and td2 shown in (1) are changed and set in advance so that a more optimal state can be obtained.

In the above embodiment, the exposure condition is changed (long-time exposure) in accordance with the vibration and blur. However, another method for coping with the vibration and the image shift by changing the exposure condition will be described below. One method is to monitor the deviation of the measured value from the laser interferometer L from the target value as shown in FIG. 3 (A), and wait for the waiting time tc (in this case, it may be shorter than 0.1 second). When exposure is performed by opening the shutter ST later, control is performed such that the amount of illumination light to the reticle 1 is increased stepwise or continuously during one-shot exposure. That is, immediately after the start of one-shot exposure, the amount of illumination is kept low, and immediately before the end of exposure, the amount of illumination is increased. In general, image blur due to vibration or stress is sufficiently small after a predetermined time has elapsed after stepping is completed, so that the exposure amount is concentrated in a time region where image blur is determined to be sufficiently small,
The exposure amount is set to be small in a time region where some image blur remains.

FIG. 5A shows an example of the illuminance change characteristic when the illumination light amount on the reticle 1 is continuously changed during one-shot exposure, and FIG. 5B shows the case where the illumination light amount is changed stepwise. An example of the illuminance change characteristic is shown.

In the characteristics of FIGS. 5A and 5B, 1
The magnitude or temporal change of the illumination light amount to be given during the shot exposure can be learned by monitoring the state of image blur caused by vibration or stress by monitoring with the laser interferometer L or various sensors S1 to S7, or The optimal curve is set in advance by a method such as inspecting a resist pattern on a wafer exposed by trial printing or the like.
Or, when exposing shot areas on the wafer one after another,
Even if the state of the image blur (vibration) generated in the previous shot area is learned, the same image blur is predicted to occur in the next shot area, and control is performed to set a curve of the light amount change characteristic. Good.

A continuous and stepwise change in the amount of light can be realized by disposing a rotating ND filter disk whose transmittance changes in the circumferential direction as shown in FIG. 5C in the illumination light path. At this time, if the rotation speed of the ND filter disk is changed,
It is possible to change a time characteristic on a change in light amount during one-shot exposure. Of course, the rotation of the ND filter disk and the opening (or closing) of the shutter ST are performed in synchronization. When a complete light shielding portion BS can be formed in a part of the filter disk as shown in FIG. 5C, this filter disk can be used also as the shutter ST.

When the light amount is changed stepwise as shown in FIG. 5B, the power supplied to the light source ILS by the light source control system 20 shown in FIG. Flash up). In FIG. 5B, the light amount is changed in two stages, but may be changed in more stages. Now, as another method of changing the exposure condition, a light source image (secondary light source image or the like) in an illumination optical system (from the light source ILS to the condenser lens 4C) is used.
When a stop that limits the area of the light source image, that is, a so-called illumination aperture stop (hereinafter referred to as a σ stop) is disposed on or near the surface where is formed, the aperture amount of the σ stop is made variable. Alternatively, a change in light amount may be obtained. In the optical path shown in FIG. 2, the σ stop is arranged on the exit side (secondary light source image side) of the fly-eye lens 4G, and can be formed by each of a large number (50 to 200) of element lenses constituting the fly-eye lens 4G. The number of secondary light source images can be
Are made substantially concentrically variable from the peripheral side toward the center side. σ aperture is 0.1 as its set value
It is determined so that about 0.7 is obtained. Where σ = 0.
7 is an area ratio (light source image area / pupil area) when the σ stop is fully opened and all of the secondary light source images of the fly-eye lens 4G are formed in the entrance pupil (Fourier transform plane) of the projection lens 2.
It has become. Therefore, if the value of the σ stop is changed from a smaller value to a larger value during the exposure of one shot, a time characteristic as shown in FIG. 6 can be obtained in terms of light quantity. When the opening amount of the σ stop is changed during the exposure of one shot as described above, while the σ value is small, an image in which the depth of focus is small and the edge of the pattern is relatively steep is obtained. As the σ value increases, an image with a large depth of focus is obtained.

As described above, the method of changing the light amount as the exposure condition is effective in the step-and-repeat method in which the relative positional relationship between the reticle and the wafer is stopped for each shot, and exposure is performed. However, the method cannot be applied when the exposure is performed by the scanning method (relative movement between the reticle and the wafer) in which the light is irradiated in a slit-shaped or arch-shaped illumination area. However, when the light amount changes continuously as shown in FIGS. 5A and 6, it is possible to cope with this by changing the scan speed in synchronization with the change in the light amount (illuminance on the reticle).

[0055]

As described above, according to the present invention, the exposure is performed in a state where the influence of stress or deformation which deteriorates the positioning stability of the substrate at the time of exposure is removed. The effect of improving the arrangement accuracy of a plurality of shot areas is obtained. Further, in the present invention, even during the exposure period, control for removing the influence of the deformation (exposure interruption, etc.) can be performed, so that there is an effect that the resolution of the exposed resist image can be prevented from lowering.

The present invention can be similarly applied to an apparatus using a pulsed laser beam as a light source which does not perform light quantity control by a shutter.

[Brief description of the drawings]

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

FIG. 2 is a block diagram showing a schematic control system configuration of the apparatus shown in FIG. 1;

FIG. 3 is a graph showing a static determination characteristic of a wafer stage and an image blurring (position shift) characteristic when a neuro calculation is performed on the influence of vibration.

FIG. 4 is a perspective view showing a configuration of a double blade shutter.

FIG. 5 is a diagram showing a change characteristic of an illumination light amount and a configuration of a light amount variable filter.

FIG. 6 is a view showing a light amount change characteristic by a σ stop.

[Description of Signs of Main Parts]

 DESCRIPTION OF SYMBOLS 1 Reticle 2 Projection lens 3 Wafer 5, 6 Wafer stage 25 Computer for image shift analysis ST Shutter ILS Light source S1 to S7 Sensor for detecting vibration, distortion, etc.

Claims (10)

(57) [Claims] 1. An exposure method using an exposure apparatus for exposing a predetermined pattern onto a substrate, wherein a stress acting on the exposure apparatus or a deformation of the exposure apparatus due to the stress is detected, and based on a result of the detection. Calculating a relative position shift amount between the predetermined pattern and the substrate due to the stress or the deformation; and starting the exposure in a state where the position shift amount is within a predetermined amount. An exposure method, wherein the exposure is interrupted when the amount exceeds the predetermined amount. 2. The exposure method according to claim 1, wherein the exposure is performed when the position shift amount remains within the predetermined amount for a predetermined period. 3. The exposure method according to claim 2, wherein the predetermined period differs depending on a moving direction of the substrate. 4. The exposure according to claim 2, wherein the exposure is restarted when the position shift amount falls within the predetermined amount for a specific period of time after the interruption of the exposure. Method. 5. The exposure method according to claim 4, wherein the specific period is shorter than the predetermined period. 6. The exposure apparatus according to claim 1, wherein the stress or the deformation is detected at a plurality of locations in the exposure apparatus.
The exposure method according to claim 1. 7. The pattern is formed on a mask, wherein the exposure method is a first exposure method for exposing the mask and the substrate in a stationary state, or a method of relatively moving the mask and the substrate. The exposure method according to any one of claims 1 to 6, further comprising one of a second exposure method for performing exposure while performing exposure. 8. A method for manufacturing a circuit pattern using the exposure method according to claim 1. Description: 9. The method according to claim 8, wherein the circuit pattern includes one of a semiconductor element, a liquid crystal display element, and an original wiring pattern. 10. An exposure apparatus for exposing a predetermined pattern onto a substrate, wherein: a stress acting on the exposure apparatus, or a detecting means for detecting deformation of the exposure apparatus due to the stress; Calculating means for calculating a relative position shift amount between the predetermined pattern and the substrate due to stress or the deformation; and starting the exposure in a state where the position shift amount is within a predetermined amount, and thereafter, the position shift An exposure control unit for interrupting the exposure when the amount is equal to or more than the predetermined amount.