JP2815010B2 - Projection optical device and imaging characteristic adjustment method - Google Patents

Description

Description: BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a projection optical apparatus and an image forming apparatus.
Method for adjusting semiconductor properties, especially manufacturing of semiconductor devices and liquid crystal display devices
Optical apparatus in an exposure apparatus for use
The present invention relates to a method of adjusting the image forming characteristics of the device. BACKGROUND OF THE INVENTION Exposure apparatuses, in particular, reduction projection type exposure apparatuses,
In recent years, what has become indispensable for the production of integrated circuits
coming. In such an exposure apparatus, usually the reticle
A circuit pattern is formed on a semiconductor wafer by I / 5 or 1/10
To obtain a resolution of 1 μm or more in line width.
Are used. In particular, at present, the degree of integration of semiconductor devices is
To improve resolution, maintain a large projection exposure area
There has been a demand for projection lenses that increase power.
You. Generally, such an exposure apparatus is used in a lithography process.
Changes in the thickness of the wafer when forming a predetermined pattern
In order to cope with surface irregularities, the projection lens
Focus detection using a gap sensor that detects the interval
Dispenser is incorporated. The detection signal of this focus detector
Based on the image plane of the projection lens, ie the projection pattern
Autofocusing to bring the image plane of the
Have been done. FIG. 2 shows such a gap sensor.
Schematic configuration of a reduced projection type exposure apparatus, so-called stepper
It is shown. In FIG. 2, the circuit pattern to be projected is
Reticle R on which turns and the like are drawn is reticle stage 2
01. This reticle stage 2
01, the light that has passed through the pattern area Pr of the reticle R
Opening 201 for bright light to enter projection lens 202
a is formed. The opening 201a allows the
An optical image of a circuit pattern drawn on the tickle R is used in a c
The image is projected onto the wafer W. Furthermore, reticks
Vacuum stage around the reticle R
A plurality of holding portions 201b for performing
Reticle stage 201 by these holding portions 201b.
The reticle R is held by suction. Reticle stage 201
Slightly moves in the xy directions shown in the figure, and the center of the reticle R is projected.
Positioning is performed so as to coincide with the optical axis AX of the shadow lens 202.
Done. Next, a step is provided below the projection lens 202.
A two-dimensional moving stage unique to the upper is provided. Ma
The wafer W as the substrate to be projected is
03 is vacuum-adsorbed. This wafer check 20
3 is the optical axis AX direction of the projection lens 202, that is, the Z direction.
It is provided on a Z stage 204 that moves vertically.
This Z stage 204 is a Y stage that moves in the y direction.
On an X stage 206 that moves in the x direction over 205,
It is provided to move up and down by a motor 207.
You. The motors 208 and 209 are respectively connected to the Y stage 20
5, for performing one-dimensional driving of the X stage 206
is there. Further, the X direction and the y direction of the Z stage 204
The laser beam that detects the coordinate position of the stage
Movable mirrors 210 and 211 for the intermediary side are provided respectively.
ing. In the above device, the illumination light (not shown)
When the reticle R is illuminated by scientific means, the pattern area P
r is formed on the image plane of the projection lens 202.
You. The positional relationship between this image plane and the surface of the wafer W,
For detecting the distance between the projection lens 202 and the wafer W
Each of the light emitter 220 and the light receiver 221 is a tip sensor.
Are provided. [0006] Of these, the projector 220 is a projection lens.
The slit light image is oblique so as to form an image on the
And the light receiver 221 is positioned at the image plane.
C) receives the reflected light of the slit light image from W, and
C, the position of W in the Z direction, that is,
It has a function of detecting the interval between C and W. This
A slit-shaped or rectangular light beam onto the wafer W obliquely.
The oblique incident light type focus detector that detects the focus by
Is disclosed in JP-A-56-42205.
It is the same as [0007] Further, the light receiver 221 is provided with a counter from the wafer W.
By synchronously rectifying the photoelectric signal of the emitted light, the wafer W
And outputs a focus signal indicating the surface position. This impatience
Since the point signal is a signal that is synchronously rectified,
Has the same s-curve characteristics as the output characteristics such as
Is input to the control circuit 222. This control circuit 2
22 is a vertical movement of the Z stage 204 based on the focus signal.
Control signal for servo-controlling the motor 207 for
Has the ability to force. By this control, the focus signal
The surface of the wafer W is positioned at a height such that
Adjustment of the Z stage 204 is performed as described above. FIG. 3 shows the gap sensor of the above embodiment.
A detailed configuration example is shown. In FIG.
Aether (hereinafter simply referred to as “LED”) 340
(C) a predetermined wavelength width that does not expose the photoresist on W
The emitted light is output as illumination light. This illumination light
It is condensed by the denser lens 341 and has a narrow rectangular shape.
The light enters the slit plate 342 having the slit 342a.
It is like that. The light transmitted through the slit 342a is
Reflected by the mirror 343 and reflected by the imaging lens 344
Converged. Further, the converged illumination light is applied to the front surface of the wafer W.
Incident on the surface near the optical axis AX of the projection lens 202,
The light image SI of the slit 242a is formed. LE above
D340, condenser lens 341, slit plate 34
2, by the mirror 343 and the imaging lens 344,
Two light projectors 220 are configured. Next, from the wafer W
The reflected light passes through the imaging lens 345,
Through the parallel mirror 346 and the vibrating mirror 347.
It is led to the lit plate 348. Sand
On the slit plate 348, an enlarged image of the optical image SI described above.
Is formed. The slit plate 348 has a slit 348
a is provided and transmitted through this slit 348a.
Light is received by the photoelectric detector 349.
You. The above imaging lens 345, parallel plate glass 345,
Vibrating mirror 347, slit plate 348, and photoelectric detection
The light receiver 221 shown in FIG.
You. Note that the vibration mirror 347 is
As a result, the enlarged image of the light image SI scans on the slit plate 348.
Parallel to the lit 348a and perpendicular to its longitudinal direction
It is driven so as to make a simple vibration in the direction. Next, the photoelectric detector
The photoelectric signal from 349 is amplified by the amplifier 351.
Later, a synchronous rectification (synchronous detection) circuit (hereinafter referred to as “PSD”)
352). This P
SD352 determines the vibration frequency of the vibration mirror 347.
A reference cycle from an oscillator (hereinafter referred to as “OSC”) 353
Input a wave number signal and synchronize the photoelectric signal with the signal
Accordingly, it has a function of outputting the focus signal FPS. The above-described drive unit 350, amplifier 351, synchronization
2 by the detection circuit 352 and the oscillator 353.
A control circuit 222 is configured. The driving unit 56
Relates to the present invention, which will be described later.
You. By the way, the above-mentioned gap sensor is in focus position.
The position in the Z direction detected as
So that the distance from the projection lens 202 is constant.
Temperature change of the projection lens 202
When the focus changes due to the focus signal FPS,
Even if the focus state is adjusted, the surface of the wafer W is not
It will be a misplaced one. In particular, higher resolution
In order to obtain, the number of gateways (NA) of the projection lens must be large.
Must be combed, but this inevitably leads to a depth of focus
Will be shallow. Depth of focus for obtaining a predetermined pattern line width accuracy
The degree becomes smaller as the minimum pattern to be projected becomes smaller.
Projecting μm line and space patterns
In this case, the depth of focus is about ± 1 μm. However,
When projecting a reticle or mask pattern, the projection optics
The system, especially the projection lens, absorbs a part of the thermal energy of the printing light.
As a result, the optical performance changes,
Is known to fluctuate. FIG. 4 shows the irradiation of printing light in a conventional apparatus.
Timing and displacement of the imaging position of the projection optical system (or
(Displacement of magnification) and ΔZ are shown. In this figure
Here, the printing light irradiation starts at time T1 and ends at T3.
(See FIG. 3A). Projection due to the effect of printing light
The variation ΔZ of the imaging position of the lens occurs from time T1,
Saturates at T2. After that, from time T3 when printing ends
Further, time T4 occurs after approximately the same time as T2-T1 elapses.
To return to the initial state (see FIG. 3B). [0015] The result of such a change of the projection lens itself.
The fluctuation of the image position is caused by the conventional projection lens described above and the wafer.
It is a focus detector that uses a gap sensor that detects the distance
Can not be corrected by the This variation is
It can be removed by using an optical focus detection means.
There are various problems. On the wafer to be projected
Has a patterned film layer of about 1 μm,
The film-like film layer is optically transparent like SiO2
Optically opaque materials such as Si, aluminum, etc.
It is. On top of that, on the order of several microns at most
Of the resist film. In addition to this situation,
When light is used for such optical focus detection means,
This point must be kept in mind because
No. There is also a means to use printing light of non-photosensitive wavelength of resist
Yes, but in this case the projection lens has poor aberration and good quality
Image formation cannot be performed. [0016] In addition, such a thermal effect causes
Not only fluctuations but also the magnification of the projection optical system
This causes a failure to form an image with good realism. Change
In addition, fluctuations in the imaging characteristics of the projection optical system
It also occurs due to fluctuations and the like. SUMMARY OF THE INVENTION The present invention is directed to such a point.
It was made in view of the thermal effects of the printing light.
Make sure that the imaging characteristics are always well adjusted without
It is an object of the present invention to provide a projection optical device capable of
Things. According to the present invention, a predetermined
Project the pattern onto the substrate to be projected via the projection optical system
In the projection optical device, the source of the fluctuation of the imaging characteristics of the projection optical system
Data collection means for obtaining information on factors
Adjusting means for correcting the imaging characteristics and changing characteristics of the imaging characteristics
Of the imaging characteristics including the parameters corresponding to the time constant of
By the model formula of the projection optical system and the data collection means
Adjustment to command adjustment means based on collected data
And calculating means for calculating the amount. Also,
A fixed pattern is projected onto the target substrate via the projection optical system.
Adjustment Method for Adjusting Imaging Characteristics of Projecting Optical Device
Method, the cause of the fluctuation of the imaging characteristics of the projection optical system
Obtaining information, equivalent to the time constant on the change characteristics of the imaging characteristics
Optics for fluctuations in imaging characteristics including changing parameters
Based on the model formula of the system and information on the cause of the fluctuation of the imaging characteristics
And predicts fluctuations in imaging characteristics, and projects based on the prediction results.
Adjusting the imaging characteristics of the shadow optical device; and
did. The change in the optical characteristics of the projection optical system is
Predictive by operating on mathematical formulas representing linear models
And the optical characteristics are adjusted based on this calculation result.
It is. Preferred embodiments of the present invention will now be described with reference to the accompanying drawings.
This will be described with reference to a plane. First, it is necessary to understand this embodiment.
In order to facilitate the layers, an outline thereof will be described. In general
The system is composed of a number of components
And are represented, for example, as shown in FIG. In FIG. 5, Ui is an input variable and Yi is an output variable.
The force variable, Xk, is a state variable. These variables are generally
In general, it changes over time, so the function of time t
Becomes Use these variables to represent a linear system
And a state equation in the form of a recurrence equation,
And an output equation that derives the physical quantity. According to the present invention, the incidence of illumination light,
Changes in the focal point and magnification of the projection optical system due to environmental changes such as
It deals with cases that occur. Therefore, the input variable U
i corresponds to the incidence of illumination light and environmental change, and the output variable Y
i corresponds to the focus variation and the magnification. Soshi
Changes in the characteristics of the individual elements that make up the projection optical system
This corresponds to the variable Xk. Next, the outline of the focus correction will be described.
First, the temperature rise due to the illumination light passes through the projection lens
It depends on the amount of illumination light. The amount of illumination light is
Illumination intensity that passes through the
It is the product of the opening and closing duty ratio of the yatsuta. Of these,
The illumination light intensity is measured by an illuminance sensor (SPD) on the stage.
At the beginning of a series of exposure work, such as when changing the reticle.
Measured. On the other hand, the open / close duty ratio of the shutter
For example, monitor the opening and closing of the shutter for 5 seconds, and
Average duty ratio (shutter open time within 5 seconds / 5
Second). The amount of illumination light obtained as described above
The correction control system 52 shown in FIG.
Focus based on the state and output equations described above.
And the amount of change in magnification are obtained indirectly. And this strange
Based on the momentum, the gap sensor shown in FIGS.
An offset is added. The magnification correction is basically performed in the same manner.
However, the correction value depends on the empty space in the lens chamber of the projection lens.
It is input to the pressure control device 60 that changes the atmospheric pressure,
Control is performed. If you change the air pressure,
Since the refractive index of air changes, it is twice that of the projection lens 26 itself.
The rate changes. The focus fluctuation and magnification fluctuation
Control is performed. Note that, as described above,
The focus and magnification of the projection lens change due to changes in atmospheric pressure
However, correction for this is also performed. Shadow due to atmospheric pressure
The sound is corrected using an atmospheric pressure sensor. Next, the state of the focal point after the above correction is shown in FIG.
Or focus as direct fluctuation measurement means shown in FIG.
Detection mechanism (hereinafter, referred to as “TTLFA”) 58, 59
(See FIG. 1). And to TTLFA
When it is recognized that the camera is not in focus better,
Is given to the correction control system 52 shown in FIG.
The parameters of the equations and output equations have been modified.
You. Then, the focus is changed again by the corrected equation.
The movement amount is obtained, and a predetermined correction is performed. As described above, fluctuation correction is performed by the correction control system.
And then check the corrected state using TTLFA
It is not necessarily the case that only one of them
This is because the fluctuation correction may not be performed well. Follow
Therefore, the state in which the fluctuation correction is well performed in both
Control should be performed assuming the best matching state
And Next, the projection optical system will be described with reference to FIG.
A focus detection mechanism (hereinafter, “TTLF”) that performs focus detection through
A ”). In addition, FIG.
This is a state in which the pattern of reticle R is being imaged.
Since the shadow substrate is a reflector 28 and not a wafer,
There is no need to consider strike light, and the color
Illumination light of the photosensitive wavelength with the best difference
Can be used. In FIG. 6, the illumination output from the light source 10 is shown.
Bright light (the same wavelength as the exposure light) passes through the relay lens 12
Line and space (lattice) pattern PA
And the transmitted light is transmitted through a lens 14 to a half mirror.
-16. This half mirror
Illumination light reflected at 16 is transmitted through a field lens 18
Through the field stop 20 and the relay lens 22 to the mirror 24
Incident on the reticle R. Next, the illumination light transmitted through the reticle R is projected.
The light reaches the reflector 28 via the shadow lens 26, where the light is reflected.
It is. The light reflected by the reflection plate 28 is reflected by the projection lens 26
The light enters the mirror 24 via the ticicle R and is reflected there.
Relay lens 22, field stop 20, field lens
18, the light passes through the half mirror 16 respectively. And projection
A split prism provided at a position conjugate with the pupil of the lens 26
The reflected light split by 30 is transmitted to a relay lens 32
The linear ray sensor 34 transmits through the cylindrical 33
The pupil division image is formed in the different part above
You. FIG. 7 shows the array sensor 34 and the serial
Optical lens 33, relay lens 32, split prism
Part 30, the field lens 18, and the field stop 20
It is shown enlarged. Note that the cylindrical lens 33
The bus bar indicates the arrangement direction of the light receiving elements of the array sensor 34.
doing. This cylindrical lens 33 has a pattern
When the line direction of the reflection image of PA is compressed to increase the amount of light
In addition, it has a function of averaging. Also, the optical element in the part of the TTLFA is shown in FIG.
As shown in FIG. 1, a portion 58 of the alignment optical system is configured.
TTLFA system 59 is constituted by other parts.
You. In FIG. 6, the broken line indicates the connection of the illumination light from the light source 10.
Image rays (principal rays) are shown, and solid lines are reticle R or reflector
28 represents a principal ray of an image forming light flux. Equipment as above
At the angle, the angle formed by the image forming light beam at the point F is open.
If the numerical aperture NA is sin θ1, it is ± θ1. In this example
Divides a 2θl light beam into two in a plane including the optical axis.
The light is split by the beam 30 and the properties of the split light beam are used.
To detect the focus. In FIG. 6, a lattice pattern PA is
Between the condenser lens 12 and the relay lens 14
The reticle R and the reflector 28 in a conjugate position,
10 illuminates the pattern PA and projects it on the reflector 28.
Shadow. The pattern image reflected by the reflector 28 is projected
At a position conjugate to the aperture stop (pupil) eP of the lens 26
The light is split into two by a prism 30. As is well known, each of the divided
Of the light beam is shifted laterally by the front pin and the rear pin.
You. Therefore, the distance between the two divided images is determined by the array sensor.
The focus state is detected by measuring on the
It can be performed. That is, as shown by a solid line in FIG.
If the distance between the two images is L during focusing,
In the case of 間隔, the interval is L 'smaller than L as shown by the dotted line.
In the case of the front focus, it becomes wider than the image interval L. In order to accurately extract the lateral displacement of this image,
Is a Fourier transform of the image on the array sensor 34, for example.
And extract the phase part, and use the Fourier transform transition theorem.
There is a method of obtaining a lateral shift by using First, the array sensor
The intensity of the light image on the
Function f of position x in the column direction, that is, the direction perpendicular to the optical axis
(X). Then, the light image shifts
The Fourier transform F1 of the previous intensity f (x) gives the spatial frequency
S can be expressed as in equation (1). ## EQU1 ## Next, when the light image is displaced by a in the horizontal direction,
The Fourier transform of the case is based on the transition theorem, as shown in equation (2).
Become. ## EQU2 ## Comparing the above equations (1) and (2),
The magnitude of the Fourier component does not change due to the displacement a,
Since only the phase is shifted by 2πas, the phase of the shift is Δθ
And the displacement a are expressed as in equation (3). (Equation 3) By the way, in practice, the integration interval is infinite.
Can not do. Therefore, an appropriate
The integration is performed for the section of the sample length l. Next
In addition, based on the above-mentioned equation of state,
The required correction control system will be described with reference to FIG.
You. In FIG. 1, the illumination light output from the light source 40 is:
C via shutter 42, reticle R, and projection lens 26.
The light is incident on Eha W. Drive of Shutter 42
The movement is performed by the shutter control system 44.
The shutter control system 44 is in contact with the main control system 46.
Has been continued. Next, the main control system 46 controls the X and Y stages.
Connected to the motors 48 and 50
It is also connected to the main control system 52. This correction control system 52
Is a parameter storage unit 52A, a calculation unit 52B, a correction value output
Force section 52C and parameter correction section 52D.
Has been established. Of these, the arithmetic unit 52B includes
Information about ON and OF of the shutter 42 from the control system 46
Parameter corrected by parameter correction unit 52D
Information from the storage unit 52A is input. Arithmetic unit 5
The output of 2B is performed to the correction value output unit 52C.
The output of the correction value output unit 52C is
This is performed on the optical device (or the pressure control device 60) 54.
It has become so. Specifically, the parallel plate shown in FIG.
A correction value is given to the driving unit 56 for rotating the glass 346.
Is input and the offset is
It has become possible to be. More specifically, the parallel plate glass 34 described above is used.
6 is disposed in the condensing system after the magnifying lens 345,
The movable portion 56 is configured to be tiltable within a predetermined angle range.
Have been. The degree of inclination of the parallel plate glass 346 varies.
When the light image SI is formed on the slit plate 348,
The center of vibration of the enlarged image is perpendicular to the longitudinal direction of the slit 348a.
It shifts in the intersecting direction (in FIG. 3, the horizontal direction in the drawing).
The shift of the vibration center with respect to the slit 348a is
The point signal FPS is focused, that is, zero on the S-curve characteristic waveform.
The position of the wafer W when determined as the point position is shifted in the Z direction.
It is equivalent to the shifted one. In this embodiment, this parallel
The flat glass 346 and the driving unit 56 control the focus fluctuation.
Correction is performed. Next, the wafer W is
The reflected light from the launch plate 28 is reflected by the projection lens 26 and the reticle R
Is incident on the alignment optical system 58 through the
FIG. 6 shows the details taken out from the part of the
The light is incident on the TTLFA system 59 shown in FIG. This
Of defocus detected by the TTLFA system 59 of FIG.
Is a parameter correction unit 52D of the correction control system 52 described above.
To be entered. Further, the output of the correction value output section 52C described above.
Is a pressure control for controlling the pressure in the lens chamber of the projection lens 26.
The pressure control device 60 is also connected to the control device 60.
Control of magnification change of the projection lens 26 by pressure control
The Lord is coming. In this example,
Part 5 of the alignment optical system provided above the reticle R
8 through the TTLFA system 5
9, but reflected light through the projection lens 26
Any method can be used as long as it can be detected.
Optical means, for example, non-photosensitive laser light is passed through a reticle
System (LSA system) that supplies the wafer W without
It may be used to guide reflected light. Of the above components, first, a shutter control system
44 is a shutter 42 according to a command from the main control system 46.
For controlling the opening and closing of. For information on shutter opening and closing,
The data is input from the main control system 46 to the calculation unit 52B.
Is used as data to determine the amount of illumination light.
It is. The illumination light intensity is determined by the transmittance of the reticle and the lamp.
Provided on the wafer stage because it changes with light intensity
As detected by the optoelectronic device (not shown)
The information detected by this photoelectric element is
Is input to the arithmetic unit 52B as data for obtaining
It is like that. This photoelectric element is indispensable in the present invention.
It is not necessary. Next, the arithmetic unit 52B receives the input illumination light.
The information about the light amount and the information stored in the parameter storage unit 52A are stored.
Equation of state and output method using parameters
By solving the equation (hereinafter collectively referred to as “model equation”)
The focus and magnification are determined. In addition, para
The parameters of the meter storage unit 52A are, as described later,
The parameter is corrected by the parameter correction unit 52D.
ing. Next, the correction value output section 52C is
A specific correction value based on the variation output from B is received.
The output from the optical device 54 and the pressure control device 60, respectively.
You. Next, the projection optical system used in the arithmetic unit 52B will be described.
Model equations, that is, state equations and output equations
Will be explained. The projection lens 26 is configured to
The temperature changes every moment. Therefore, the variables “U, Y, X
i "must be expressed as a function of time
However, here we consider time as discrete,
"U (K), Y (K), Xi (K)". Where K
= 0, 1, 2,.... That is, in this embodiment, as described above,
Measures the duty of the shutter open time at 5 second intervals
So, instead of time, we use K
I have. With respect to K as described above,
If the following equation is expressed in two dimensions, [Equation 5] In these equations (4) and (5), the input change
The number U (K) is the illumination that has passed through the projection lens 26 for 5 seconds.
The amount of light, that is, the amount of energy, as described above
Is expressed as the product of the shutter opening / closing duty ratio and the irradiation light intensity.
Be forgotten. The output variable Y (K) is the focus to be determined.
It is a point or an amount of change in magnification. State variables X1 (K), X
2 (K), X1 (K + 1), X2 (K + 1)
It does not correspond to joint physical quantities. In this embodiment, every 5 seconds
The new focus and magnification change Y (K) is calculated
And the state variables after 5 seconds with respect to X1 (K) and X2 (K)
Are X1 (K + 1) and X2 (K + 1). Next, α, β, a, b, m, and L are projection levels.
4 to reproduce the fluctuation characteristics shown in FIG.
Is a constant (parameter). Of these, parame
Data α and β are time constants based on the change characteristics inherent in the projection lens 26.
And the parameters a and b correspond to the coefficient terms on the characteristic.
And m is the characteristic of each projection even if the amount of illumination is the same.
Considering that each lens is slightly different,
This is a unique constant determined for each size. L is the lamp
It corresponds to the light intensity. Next, the parameters α, β, a,
A specific method for obtaining b and m will be described. these
The parameters of the pseudo exposure operation before the actual exposure operation
The measurement can be performed by performing
Using the above-mentioned TTLFA or the like during or during the operation
Can be measured. In this embodiment, α, β, a,
b and m are determined in advance from experiments, and “L” is
As described later, use TTLFA measurement results
A method of estimating the probability will be described. Α, β, a, b, and m are reserved as follows.
To be determined. First, a reflector just below the projection lens 26
28, and the projection lens 26 is
The interval between the reflectors 28 is kept constant. Then, illuminate
When the light passes through the shadow lens 26, the output of the TTLFA changes.
Is obtained as shown in FIG. The amount of change Δ in FIG.
The rising curve of Z is represented by the sum of exponential functions. An example
For example, [0055] Is represented by In this equation (6), 1 / T
1, 1 / T2 correspond to the parameters α and β, respectively, and A and B
Correspond to parameters a and b, respectively. The above parameters
The meter is for focus variation and for magnification variation.
And each one is asked for. These parameters
Are stored in the parameter storage unit 52A shown in FIG.
The calculation unit 52B is used for the calculation of the model expressions of the expressions (4) and (5).
It is. Next, the overall operation of the above embodiment will be described.
I will tell. First, the calculation unit 52B of the correction control system 52 (see FIG. 1)
Information on the amount of illumination light in
Will be explained. First, exposure is step and repeat
In the case of the port method, as shown in FIG.
It is performed in the shape of a circle. Then, as shown in FIG.
The illumination light amount is the illumination light transmitted through the projection lens 26 within 5 seconds.
Main control system as bright light time ratio, that is, duty ratio
The signal is output from 46 to the operation unit 52B. This percentage is
Corresponds to the ratio of the opening time of the shutter 42 for 5 seconds.
You. The duty ratio detected at a certain point in time
Is the value 5 seconds before the time of detection, and before that
Is not detected at all. Considering the effects of the past
Don't worry about assuming that the projection optics is a linear system
It is because. The illumination light intensity is as described above.
Illuminance sensor (not shown) provided on stage
And is input to the calculation unit 52B. Next, with reference to FIG.
And will be described specifically. FIG. 3A shows the above-mentioned du.
FIG. 11B shows the calculation of the fluctuation.
The value (predicted value) Yi is shown. First, time t = t0
Then, the exposure has not started yet. Therefore, (5)
The input variables X1 and X2 of the expression are both 0,
The calculated value Yj is 0. Next, at time t = t1, exposure is started.
As a result, the duty ratio DT = D1, and the equations (4) and (5)
Is calculated, Y = Y1, and the value after another 5 seconds
Y1 'is pre-calculated. Next, at time t = t2,
And the duty ratio DT = D2,
The calculation based on equations (4) and (5) is performed, and the last 5 seconds
The sum Y2 with the subsequent value Y1 'is obtained, and the value Y after 5 seconds is further obtained.
2 'is required. As described above, the fluctuation amount Yj repeats every 5 seconds.
Return operation is performed. Thus, the fluctuation amount Yj is five seconds before.
It is based on computed values only,
Absent. Note that these variations are relative to the focus,
Each of the factors relating to the magnification is calculated. As described above, the operation is performed by the arithmetic unit 52B.
The calculated variation is output to the correction value output unit 52C.
It is. The correction value output unit 52C outputs the input fluctuation.
Based on the amount, specific parallel flat glass 346 (see FIG. 3)
And the pressure in the lens chamber of the projection lens 26.
Control values are determined, and the light receiver 54, the pressure control device 60
(See FIG. 1). At this time, there is also a focus variation in pressure control.
Cheating. For this reason, the correction value output unit 52C uses the pressure control.
Degree of inclination of parallel flat glass 346 in consideration of focus fluctuation
Is determined, and the driving unit 56 (see FIG. 3) rises based on this.
An offset is applied. For this reason, ideally the focus change
The dynamic and magnification changes are both corrected to zero.
Parameter setting error in formula or repeated calculation
When a certain amount of time has elapsed due to the accumulated error
The calculated predicted fluctuation amount and the actual fluctuation amount (FIG. 9B)
Error from the actual measured value)
It is not always corrected satisfactorily. Therefore, in this embodiment, the TTLFA of FIG.
To correct the parameters of equations (4) and (5). Less than
The correction method will be described. Chosen by TTLFA
Locking is performed for each wafer exposure or for one wafer lot.
It is performed every time. First, the reflector 28 on the stage
Then, focusing is performed by the gap sensor shown in FIG.
U. Next, while maintaining this state,
Defocus is detected using TTLFA. This shift amount
If the value is equal to or larger than the allowable value, the parameter correction unit 52D
Then, the parameters are modified. Here, the parameter L in equation (4) is modified.
Will be described. This correction is based on the following equation:
It is done. [Mathematical formula-see original document] Here, G is what is called an adaptive gain.
And is determined so that the sequence Li converges. Yi
Is the result of measurement by the i-th TTLFA, and Y
Are (4) and (5) at the time of measurement as described above.
Is the calculated value. Modification of the above parameters
Is performed by the parameter correction unit 52D of FIG.
The modified parameters are read from the parameter storage unit 52A.
The new parameter is input to the calculation unit 52B and
The operations of equations (4) and (5) are performed. In addition, other than L
The parameters a, b, and m may be modified. As described above, the TTLFA checks
Parameters to sequentially drive them to appropriate values.
I do. Therefore, the interval between checks by TTLFA is gradually reduced.
Can be made longer. Next, collecting information on the amount of illumination light
The following describes another method of embedding. In this method,
If the shutter 42 opens 200 ms, the shutter 42
The open / closed state of the yatsuta 42, that is, the ON / OF state is mainly
It is input from the control system 46 to the calculation unit 52B. Specifically
Indicates that the open / closed state of the shutter 42 shown in FIG.
As shown in (B), sampling is performed every lmS
It is input to the calculating unit 52B. Next, the variation amount Y is calculated by the equations (4) and (5).
Find (K). First, at time t1, the shutter is turned off.
Therefore, no exposure is performed. Therefore the amount of fluctuation
Y (1) is 0. At time t = t2, shutter is ON
And exposure is being performed. Therefore, sampling
The value becomes a logical value "T", and U (K
The variation amount Y (2) is obtained by 2). Next, at time t = t3, the incident light
A fluctuation amount Y (3) is obtained based on the light amount U (K3).
You. The above operation is repeated to determine the amount of fluctuation.
And an offset to the gap sensor of FIG. 3 is added.
It is. It should be noted that this offset is performed every 1 ms.
Good, or every 5 seconds as described above
You may. However, in the above method, the performance
Due to the large number of calculations and parameter setting errors,
Errors can occur. Therefore, T shown in FIGS.
Check the actual image plane position from time to time with TLFA
Then, based on the result, the parameter correction unit 52D
(4) and (5) are corrected, and the error
Try to reduce the occurrence. Further, in some cases, the initial state variable Xk
It is also conceivable that the term value is incorrect. In this case,
The estimation is performed by a so-called observer. The observer is: The configuration is as follows. Thereby, Xk
Is modified to Xka using the measurement result Ye by TTLFA.
Corrected. f is a constant called the observer gain
You. By this modified state variable Xka,
The operations of equations (4) and (5) are performed. Explain concretely
Then, first, focus using TTLFA (see FIGS. 6 and 7).
The actual measured value Ye of the point variation is measured. The interval is a state variable
It is assumed that the value of Xk is updated n times. That is, TT
By LFA, Ye (Ko), Ye (Ko + n), Ye
(Ko + 2n). The value of... Will be known. Next
The average of the parameter U between K = K0 and K = K0 + n
The value Ua is calculated from Ye (Ko) and Y (Ko + n) by the equation (4).
Is obtained by solving Next, using this Ua
And the state variables X, K = K0 + n and K = K0 + 2n,
The estimated values Xa and Ya of the output variable Y are sequentially obtained by the following recurrence formula.
Confuse. [Mathematical formula-see original document] [Mathematical formula-see original document] This Ya (K) is the estimated value of focus variation Yj ′.
Will correspond. In the above method, T
The measurement interval of TLFA does not necessarily have to be a constant value.
No. For example, it may be performed every exposure of one wafer W.
When the focus fluctuation is large, it is complicated, and when it is small, the interval is
The method may be any method that is performed in a large manner. The above
In the embodiment, the description has been made by focusing on the temperature change due to the illumination light.
However, the optical characteristics of the projection lens
Depending on the environmental changes and changes in the σ value of the illumination optical system for exposure,
Even if it changes, it can be similarly corrected. Special
In addition, parameter correction by TTLFA is effective. In the above embodiment, the focal point of the projection lens
And magnification have been corrected, but only one of them may be used.
No. In addition, the alignment system 58 of the TTL system (see FIG. 1)
Reference mark provided on the side of the reflection plate 28
The detection of the distortion causes a change in the magnification of the projection lens.
Can be measured. For this purpose, the alignment system 58 is
R to detect marks at different positions on R
Be killed. Then place the fiducial mark in the image plane of the projection lens
Run, the projected points of the two marks on the reticle R
By measuring the positional relationship, a magnification error can be detected. Follow
If such a measurement is possible,
The system and the fiducial mark work together as a means of measuring fluctuation. Alternatively, see JP-A-59-94032.
With slit on XY stage as disclosed
Of the reticle pattern
The contrast is detected by the photoelectric sensor and the projection lens
The method for obtaining the actual image plane position of the
It is effective. Further, the adjusting means of the present invention includes
If the reticle side of the shadow lens is non-telecentric
If the distance between the reticle R and the projection lens is changed,
Is also effective. As described above, according to the present invention,
In particular, it compensates for the imaging characteristics (focus fluctuation, magnification fluctuation) of the projection optical system.
Correct, high quality reproducible images can be formed.
You. Also, you can always perform accurate focusing
Control the line width of the exposed pattern
The density is increased, and the reproducibility is improved. Ma
Further, according to the present invention, thermal change due to light absorption of the projection optical system is achieved.
Environmental changes such as changes in atmospheric pressure, equipment temperature, etc.
And the release of the structural distortion of the device
What kind of effects can be obtained for focus fluctuation and magnification fluctuation?
Can be.

BRIEF DESCRIPTION OF THE DRAWINGS FIG. 1 is a configuration diagram showing an embodiment of the present invention, FIG. 2 is a perspective view showing a schematic configuration of an exposure apparatus, and FIG. 3 is a configuration diagram showing a detailed configuration example of a gap sensor. 4 is a diagram showing a change in optical characteristics of a projection lens due to incidence of illumination light, FIG. 5 is an explanatory diagram showing a general system, FIG. 6 is a configuration diagram showing a TTLFA, and FIG. FIG. 8 is a configuration diagram illustrating a main part, FIG. 8 is a diagram illustrating the capture of information regarding the amount of incident light, FIG. 9 is a diagram illustrating a variation calculation, and FIG. 10 is a diagram illustrating a variation calculation. [Description of Signs] 25: Projection lens, 28: Reflector, 30: Split prism, 34: Array sensor, 40: Light source, 42: Shutter, 52: Correction control system, 54: Light receiver, 55: Driving unit, 60 ... Pressure control device

────────────────────────────────────────────────── ─── Continuation of the front page (56) References JP-A-60-78454 (JP, A) JP-A-62-136821 (JP, A) (58) Fields investigated (Int. Cl. ⁶ , DB name) H01L 21/027

Claims (1)

(57) [Claims] In a projection optical device that projects a predetermined pattern onto a projection target substrate via a projection optical system, a data collection unit that obtains information on a cause of a change in the imaging characteristic of the projection optical system; and an imaging characteristic of the projection optical system. Adjusting means for correcting, based on the data collected by the data collection means and the model formula of the projection optical system with respect to the fluctuation of the imaging characteristic including a parameter corresponding to the time constant on the change characteristic of the imaging characteristic And a calculating means for calculating an adjustment amount instructed to the adjusting means. 2. 2. The projection optical apparatus according to claim 1, wherein the calculating unit calculates the adjustment amount at predetermined time intervals. 3. 2. The projection optical apparatus according to claim 1, wherein the information on the cause is an amount of light energy incident on the projection optical system. 4. 2. The projection optical apparatus according to claim 1, wherein the model formula is a recurrence formula including a portion that adds information on the cause and a change portion of the imaging characteristic according to the time constant. 3. . 5. The projection optical device according to claim 1, wherein the parameter includes a plurality of parameters corresponding to a plurality of time constants. 6. The projection optical device according to claim 5, wherein the model formula includes a parameter corresponding to a component ratio of the plurality of time constants. 7. In an image forming characteristic adjusting method for adjusting an image forming characteristic of a projection optical apparatus that projects a predetermined pattern onto a projection target substrate via a projection optical system, obtaining information on a cause of a change in the image forming characteristic of the projection optical system Based on a model formula of the projection optical system and information on the cause of the imaging characteristic variation with respect to the variation of the imaging characteristic including a parameter corresponding to a time constant on the variation characteristic of the imaging characteristic, Estimating a change and adjusting an imaging characteristic of the projection optical apparatus based on the prediction result. 8. 8. The method according to claim 7, wherein a change in the imaging characteristic is predicted every predetermined time. 9. 8. The method according to claim 7, wherein the information on the cause is an amount of light energy incident on the projection optical system. 10. 8. The imaging characteristic according to claim 7, wherein the model expression is a recurrence expression including a part for adding information on the cause and a part for changing the imaging characteristic according to the time constant. Adjustment method. 11. The method according to any one of claims 7 to 10, wherein the parameters include a plurality of parameters corresponding to a plurality of time constants. 12. 12. The model formula according to claim 11, wherein the model formula includes a parameter corresponding to a component ratio of the plurality of time constants.
The image forming characteristic adjustment method according to the above section.