JPH10284415A - Imaging position detecting apparatus and manufacture of semiconductor device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide an imaging position detecting apparatus which can obtain a measured result sufficiently coinciding with a position and shape of a focusing plane caused by test exposure. SOLUTION: A reference pattern 24a is positioned on a conjugate plane of a projection optical system 11 of a reticle 9 pattern, the reference pattern 24a is illuminated from a side opposite to the optical system 11. The apparatus measures a quantity of light flux passed through light transmission regions of the reference pattern, reflected by the reticle pattern, and returned to the light transmission regions of the reference pattern; moves the reference pattern to a plurality of heights in the optical axial direction to measure quantities of light at the respective heights; and finds a best focus height of an image in the projection optical system of the reticle pattern on the basis of light quantity distributions in the optical axial direction. The apparatus includes a means for finding average height of the distributed quantity of light.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image plane detecting apparatus for determining the position and shape of an image plane of a projection optical system, and more particularly to an image forming apparatus that can be obtained when a test exposure is performed on a wafer. The present invention relates to an imaging surface detection apparatus and a method for manufacturing a semiconductor device using an imaging surface detection method, which can determine the position and shape of a surface without performing test exposure.

[0002]

2. Description of the Related Art A projection exposure apparatus is used when a semiconductor element or a liquid crystal display element is manufactured by a lithography process, and there are a projection exposure apparatus and a scanning exposure method. In the projection exposure apparatus, when starting the apparatus, it is necessary to determine the image plane shape and position of the projection optical system with high accuracy, that is, initialize the best focus position of the upper surface of the wafer with respect to the lower surface of the reticle (pattern surface). There is a need. The method of determining the best focus position includes a method of burning a wafer and a method of detecting a spatial image contrast by double imaging.

Further, in a projection exposure apparatus, since an almost flat reticle lower surface (pattern surface) is printed on a substantially flat wafer upper surface, it is required that the image forming surface of the projection optical system is always flat. Therefore, it is necessary to provide a means for confirming that the flatness of the imaging surface has not changed due to aging or other factors. If the flatness is degraded, compensation means must be provided. The shape of the imaging surface is a surface connecting the focal positions at each point in the exposure area. Therefore, in order to determine the shape of the imaging plane, it is essential to determine the focal positions at a plurality of points in the exposure area. At present, the method of obtaining the position and the shape of the imaging plane is performed by a method using wafer printing.

One of the methods by baking a wafer is as follows.
Test marks placed at several places on the reticle are printed by changing the wafer focus position and exposure position, developed, and the optical microscope is used to determine the midpoint of the defocus range where the resist image is resolved, The focus position at each position is determined by the value obtained from the test mark, the surface determined by the focus position at each point in the exposure area is used as the image plane, its shape is determined, and the focus positions at each point in the exposure area are averaged. There is a way to make it the best focus.

As a stricter method, a CD for finding the best focus position using an electron microscope instead of an optical microscope
There is a focus method. Best focus positioning by the CD focus method is to print test marks on a reticle such as an isolated mark or line and space by changing the wafer focus position and exposure position, and measure the line width of the resist image remaining after development with an electron microscope. A defocus range within a range of, for example, ± 10% with respect to a correct printing line width is obtained, a middle point thereof is set as a focal position at each measurement point, and a plane determined from the focal position of each point in the exposure area is defined as an image plane. Then, the shape is obtained, and the focus positions at each point in the exposure area are averaged to obtain the best focus.

Here, an example of a conventional method of measuring the image plane shape by the CD focus method and a method of measuring the best focus value will be described in detail. The confirmation of the best focus value at an arbitrary point in the exposure field of the projection optical system is performed as follows as an example. First,
Line-and-line installed anywhere on the reticle
A space pattern is printed on the wafer coated with the positive resist. Each time printing is performed, the position is shifted in the focal direction (the position is accurately grasped by the autofocus system), and the position is also shifted in the plane perpendicular to the optical axis so that the patterns do not overlap. The direction of the line and space pattern is 0, 45,
90 and 135 degrees are common. The measurement positions in the exposure area are also 0, 45, 90, 13 around the optical axis.
In general, it is several points each in the direction of 5 degrees.
When such a pattern is used, the difference in the image plane shape depending on the pattern direction can also be measured.

At this time, the cross-sectional shape of the resist image at a certain position in the focal direction is as shown in FIG. The bottom line width of the resist image at each focal position is measured by an electron microscope, and a polynomial approximation is performed by data processing. The bottom line width is ± the wafer projection line width of the line and space on the reticle (the value obtained by multiplying the line width on the reticle by the magnification of the projection optical system).
Two focus positions such as 10% are obtained, and the middle point is set as the focus position at an arbitrary position. Here, the bottom line width at the focal position is approximately equal to the wafer projection line width of the line and space on the reticle.
The exposure must be set (see FIG. 8). Through such a process, the image plane shape is measured for each pattern direction. Among them, at each measurement position, a meridional direction pattern which is a pattern located on a straight line passing through the optical axis and extending in a direction perpendicular to the straight line, and a meridional direction pattern located on a straight line passing through the optical axis and parallel to the straight line It is divided into sagittal direction patterns that extend in the direction, and the meridional image surface shape and sagittal image surface shape that are the respective image surface shapes are obtained, and the averaged image surface shape (hereinafter, the image surface shape of a general projection optical system) Will be called).

As a best focus measurement method using aerial image contrast, there is a focus position initialization mechanism (focus calibration, hereinafter referred to as FC) disclosed in Japanese Patent Application Laid-Open No. Hei 4-34819. This will be briefly described below. The translucent indicator plate on the wafer stage, which is set at approximately the same height as the wafer surface, is illuminated from below with the exposure wavelength, and the light transmitted through the arbitrary pattern engraved therethrough passes through the projection optical system and the lower surface of the conjugate reticle Reflected again on the same pattern on the conjugate index plate via the projection optical system, and the intensity of light transmitted through the pattern is moved while moving the wafer stage in the focus direction of the projection optical system. By detecting (the position in the focus direction is measured by an autofocus system), the intensity peak position is detected, and a certain device offset is added to detect the position as the best focus position.

By moving the FC mechanism to an arbitrary position in the exposure area on a stage and detecting focus, an image plane can be detected in principle. Since such a measurement method does not require a printing step, it is greatly expected as a very simple method. The conventional FC mechanism requires offset management for the best focus position due to factors not depending on the projection optical system, such as the resist thickness and the detection position offset of the autofocus mechanism.

This is because measurement offset occurs in autofocus due to the fact that the resist is not applied to the FC index plate or the difference in reflectance with the wafer. However, in actuality, if the image plane position is measured by the CD focus method and the average value is used as the best focus, and the difference from the FC detection value at that time is measured in advance as an offset amount, the FC measurement is always performed. Because a constant offset amount is added to the value,
During subsequent device initialization, detection of the best focus is possible.

In addition to the two offset factors independent of the projection optical system, the remaining longitudinal aberration of the projection optical system can be cited as an offset factor depending on the projection optical system. If there is residual longitudinal aberration, the focus measurement by the CD focus method and the measurement result by FC are further deviated, and as a result, the focus offset amount is increased or decreased. But,
Regarding the best focus detection, if there is no change with time of the offset amount, it does not particularly cause an error. Also,
In order to manage the offset with respect to the best focus as the average of the entire image plane, the FC measurement for finding the best focus is performed by using any one of the values in the exposure area whose offset amount is known.
It is enough to go in several places.

[0012]

However, in order to detect an accurate image plane shape using the FC function, measurement at a plurality of points in the exposure area is required. Distribution of non-uniform residual longitudinal aberration within
It has been pointed out that the position and shape of the image plane measured by wafer printing by focusing or the like differ from the image plane detected by aerial image contrast by the FC mechanism. Therefore, the present invention uses an imaging position detection device and an imaging position detection method capable of obtaining a measurement result in which the position and shape of the imaging surface sufficiently match as the imaging characteristics of the imaging surface by the test exposure. It is an object to provide a method for manufacturing a semiconductor device.

[0013]

According to the present invention, there is provided a projection optical system for forming an image of a pattern formed on a reticle; and a projection optical system which is located at a position conjugate with the reticle with respect to the projection optical system. An illumination system for illuminating the reference pattern; projecting a reference pattern image illuminated by the illumination system onto a pattern surface of the reticle via the projection optical system; A detection system for receiving the reflected light from the projector via the projection optical system and the reference pattern and detecting the amount of the received light beam; and moving the reference pattern along the optical axis of the projection optical system. A moving device that moves the reference pattern through the moving device along a plurality of positions in the optical axis direction of the projection optical system, and moves the reference pattern at each position detected by the detection system. A light quantity distribution in the optical axis direction based on the quantity, and a processing system for finding a best focus position of an image of the reticle pattern by the projection optical system based on the light quantity distribution in the optical axis direction. In the apparatus, when obtaining a best focus position of an image of the reticle pattern by the projection optical system based on the light amount distribution in the optical axis direction, the processing system corrects a measurement error component due to aberration remaining in the projection optical system. In order to achieve this, a configuration is provided that includes a correction processing unit that deforms the shape of the light amount distribution into a desired shape.

Further, the present invention provides a projection optical system for forming an image of a pattern formed on a reticle; a detection system for photoelectrically detecting an optical characteristic of an image forming surface of the projection optical system; And a processing system for obtaining a best focus position of the image by the projection optical system based on the detection signal from the projection optical system. In order to correct a measurement error component due to residual aberration in the system, a configuration is provided that includes a correction processing unit that deforms the shape of the detection signal into a desired shape.

Further, according to the present invention, in a method of manufacturing a semiconductor device having a projection exposure step of projecting and exposing a pattern formed on a reticle onto a photosensitive substrate via a projection optical system, prior to the projection exposure, An imaging plane characteristic detection step of detecting an imaging plane characteristic of the projection optical system; and an adjustment step of adjusting the projection optical system after the detection step, wherein the detection step is performed with respect to the projection optical system. A light beam from a reference pattern disposed at a position conjugate with a reticle is projected onto the pattern surface of the reticle via the projection optical system, and reflected light from a projected image of the reference pattern is projected onto the projection optical system and the reference pattern. And a light amount detecting step of detecting the light amount of the received light beam; and while moving the reference pattern along a plurality of positions in the optical axis direction of the projection optical system, Based on the information obtained in the recording light amount detection step, the light amount distribution in the optical axis direction is obtained, and based on the light amount distribution in the optical axis direction, the best focus position of the image of the reticle pattern by the projection optical system is determined. Calculating a best focus position of an image of the reticle pattern by the projection optical system based on the light amount distribution in the optical axis direction.
In order to correct a measurement error component due to the aberration remaining in the projection optical system, a correction processing step of changing the shape of the light quantity distribution to a desired shape is provided.

Further, according to the present invention, in a method of manufacturing a semiconductor device having a projection exposure step of projecting and exposing a pattern formed on a reticle onto a photosensitive substrate via a projection optical system, An imaging plane characteristic detecting step of detecting an imaging plane characteristic of the projection optical system; and an adjusting step of adjusting the projection optical system after the detecting step. A photoelectric detection step of photoelectrically detecting the optical characteristics of the surface; and correcting the shape of the detection signal photoelectrically detected by the photoelectric detection into a desired shape in order to correct a measurement error component due to aberration remaining in the projection optical system. And a deforming correction processing step.

[0017]

DESCRIPTION OF THE PREFERRED EMBODIMENTS FIG. 1 shows an embodiment of the present invention applied to an image plane position detecting device of a projection optical system of a semiconductor exposure apparatus. The configuration of the present embodiment uses an FC mechanism to measure the focal position at each point in the exposure area, and adds an appropriate signal processing system so that an accurate image plane shape can be obtained. Has the same configuration as the conventional example.

When manufacturing a semiconductor device or a liquid crystal display device, a reticle (or photomask) is illuminated with exposure light, and the pattern of the reticle is exposed to a wafer (or glass) coated with a photoresist via a projection optical system. A projection exposure apparatus that forms an image on a plate or the like is used.
As the exposure light source, a bright line (g-line, i-line, etc.) of a mercury lamp or an excimer laser (KrF, ArF, etc.) is used. In this embodiment, an example using a mercury lamp will be described. In the following description, the Z axis is set in the optical axis direction of the projection optical system, and two directions orthogonal to the Z axis and orthogonal to each other are defined as an X axis and a Y axis.

FIG. 1 shows a projection exposure apparatus, and a mercury lamp 1
Is emitted by the elliptical reflecting mirror 2 and collected near the second focal point. A mirror shutter 3 having a surface on the incident side formed as a reflecting surface is disposed near the light condensing point. At the time of exposure, the mirror shutter 3 is retracted to a position where the light from the mercury lamp 1 is not reflected. Light from the mirror 2 passes through the mirror shutter 3. Thereafter, the light beam is converted by the input lens 4 into a substantially parallel light beam, and exposure light having a wavelength selected by a band-pass filter (not shown) (for example, an i-line having a wavelength of 365 nm or a light beam having a wavelength of 436 nm) among the substantially parallel light beams. g-line or the like) is incident on the fly-eye lens 5, and a number of secondary light sources are formed on the exit-side focal plane of the fly-eye lens 5.

An aperture stop 6 is arranged on the focal plane on the exit side of the fly-eye lens 5, and a large number of 2 apertures in the aperture of the aperture stop 6 are provided.
Exposure light from the next light source passes through the condenser lens 7 and the mirror 8 for bending the optical path, and illuminates the pattern area on the reticle 9 in a superimposed manner. The reticle 9 is held on a reticle stage 10, and a pattern of the reticle 9 is projected and exposed to each shot area on a wafer 12 via a projection optical system 11 under exposure light. The reticle position of the reticle stage 10 is measured by a reticle interferometer (not shown).

The wafer 12 is placed on a wafer stage 14 via a wafer holder 13, and the wafer stage 14 is used for positioning the wafer 12 in an XY plane.
It comprises a stage (not shown), a Z stage (not shown) for positioning the wafer 12 in the Z direction, a leveling stage (not shown) for correcting the tilt angle of the wafer 12, and the like. The XY coordinates of the wafer stage 14 are always measured by a laser interferometer (not shown).
The Z coordinate of the wafer stage 14 is measured by an encoder (not shown) or the like.

The movement of the wafer stage 14 in the XY directions, the movement in the Z direction, and the tilt are performed by driving the drive system 32. Therefore, XY in the wafer stage 14
The stage is moved in the X and Y directions by driving the drive system 32, and the Z stage in the wafer stage 14 is driven by the drive system 32.
Is moved in the Z direction by the drive of. Further, the leveling stage in the wafer stage 14 is tilted by the drive of the drive system 32. The drive system 32 is controlled by a processing system 27 described later.

Above the wafer stage 14, a focus detecting device for detecting the position of the surface of the wafer 12 or the position of a reference plate 24 described later is provided. This focus detection device includes an irradiation system 30 for irradiating light to the surface of the wafer 12 or the surface of the reference plate 24,
4 and a focus detection system 31 for detecting the position of light from the irradiation system 30 reflected on the surface. Then, the focus detection system 31 obtains the position of the surface of the wafer 12 or the surface of the reference plate 24 by detecting the position of the light reflected on the surface of the wafer 12 or the surface of the reference plate 24. In addition,
A detection signal from the focus detection system 31 is sent to a processing system 27 described later.
Is input to

Wafer holder 13 on wafer stage 14
An index plate 24 as a reference plate is attached in the vicinity of. On the upper surface of the index plate 24, a reference pattern 24a having a light transmitting area and a light shielding area is formed. As shown in FIG. 2, the reference pattern 24a of this embodiment
Line and space with four types of directions to reduce measurement errors when there is an arbitrary pattern
Uses a pattern. In FIG. 2, the black line part is a light transmitting area, and the white part is a light shielding area. Further, in this example, an example is shown in which an intensity grid type reference pattern 24a having a light transmitting area and a light shielding area is provided on the reference plate 24, but the reference pattern 24a of the present invention has a step. It is also possible to use a concavo-convex pattern, that is, a phase grating type pattern.

When measuring the image forming position of the pattern on the reticle 9, the mirror shutter 3 is inserted into the optical path. As a result, the main lighting system and the mercury lamp 1
The light emitted from the lens system 20 is collected by reflection on the elliptical mirror 2, reflected by the mirror shutter 3, and
And enters the two-branch light guide 21. On this occasion,
At the entrance end of the two-branch light guide, a light source image having a sufficient NA and light diameter is formed.
The illumination light emitted from 1 is reflected vertically upward by the mirror 22, and then Koehler-illuminates the reference pattern 24a on the index plate 24 from below through the condenser lens 23. At this time, the on-stage index plate 2 having the reference pattern 24a
4 and the optical axis of the projection optical system 11, the illumination light has the same wavelength as the exposure light.
Are reflected by the lower surface of the conjugate reticle 9 via the projection optical system 11 and again form an image at the same position of the reference pattern 24a on the conjugate index plate 24 via the projection optical system 11. Further, the light beam transmitted downward through the reference pattern 24a enters the two-branch light guide 21 via the condenser lens 23 and the mirror 22 disposed immediately below the index plate 24. The light beam is branched by the two-branch light guide 22 and guided to the light receiving sensor 25 to be received. Then, the detection signal detected by the light receiving sensor 25 is input to the processing system 27, and predetermined signal processing is performed.

The FC mechanism of this embodiment is configured as described above. In this embodiment, since the mirror 22, the condenser lens 23, and the index plate 24 of the FC mechanism are installed on the stage 14 to be operated, an example using the flexible two-branch light guide 21 is shown. . However, the means for guiding the illumination light is not limited to one two-branch light guide, and the illumination light may be guided by a relay optical system, or a combination of a plurality of light guides and an optical system may be used. It does not matter if the performance can be obtained.
As a method of splitting the transmitted light beam and the received light beam, a beam splitter or the like may be used instead of the two-branch light guide.

FIG. 3 shows how the amount of received light changes when the height of the wafer stage 14 is moved in the optical axis direction of the projection optical system 11 via the drive system 32. The position where the amount of light is maximum in the graph of FIG. 3 is a state where the image of the reference pattern is re-imaged exactly at the position of the index plate 24, and the state where the contrast is the best, that is, the focus of the re-imaged space image of the reference pattern. Position. This focus position is not necessarily CD
It does not match the focus position by the focus method.

In the focus measurement by the FC mechanism, factors that do not depend on each position in the exposure area, that is, factors such as the resist thickness and the detection position offset of the autofocus mechanism, are determined in advance by the difference from the focus position by the test exposure. By holding the offset as an offset and thereafter correcting the offset each time the focus is measured by the FC mechanism, it is possible to adjust the position to the best focus position that would have been obtained by the CD focus method.

On the other hand, if there is residual aberration in the projection optical system, the focus signal obtained by the FC mechanism undergoes signal shape deformation depending on the residual aberration. That is, the factors caused by the residual aberration of the projection optical system are two-dimensionally distributed depending on each position in the exposure area. Regarding this factor, even if the difference from the focus position by the test exposure is stored as a two-dimensional matrix in advance, the offset cannot be actually corrected. In addition, if the two-dimensionally distributed offset remains, the focus measurement by the FC mechanism does not reflect the residual aberration of the projection optical system. Means not. Therefore, it is necessary to eliminate factors that are two-dimensionally distributed depending on each position in the exposure area as much as possible.

In the following, by performing appropriate processing on the light quantity signal of the FC mechanism in the signal correction processing section 27a in the processing system 27, an image plane shape substantially the same as the image plane shape measured by the CD focus method is measured. Describe possible ways. First, the processing system 27 drives the driving system 32 to shift the height z _j of the index plate 24 in the Z-axis direction for each point P _i (i = 1 to N) in the exposure region, and The light quantity I _{i, j} at the distance z _j (j = 1 to M) is measured via the light receiving sensor 25. That is,
In reality, a signal is processed at a certain pitch (time interval) with respect to the movement of the wafer stage in the focus direction.
Is received, the received light quantity signal distribution has discrete discrete values. This light quantity distribution curve I _i (j)
Is a focus curve for determining the best focus position.

At this time, the position when the wafer stage 14 is moved in the focus direction is detected by the focus detection system 31 of the focus detection device, and this detection signal is input to the processing system 27. Therefore, the position at each height z _j in the light amount curve distribution I _i (j) obtained by the processing system 27 is associated with each other based on the detection signal from the focus detection system 31 of the focus detection device. As shown in FIG. 1, a sensor 26 for monitoring a change in the light amount of the light source 1 is attached, and a light receiving signal of the light receiving
It is more preferable to reduce the error factor due to the change in the light amount of the light source 1 by performing the division process by the received light amount of No. 6.

Next, the signal correction processing unit 27 in the processing system 27
In a, a discrete focus curve I _i (j) at each height z _j is smoothed by a quadratic curve approximation or a spline interpolation method in order to reduce noise. An example of the graph subjected to the smoothing process is shown by a dotted line in FIG. The signal correction processing section 27a calculates the obtained continuous focus curve I
_{For i} (z), the light amount at each height z in the Z direction is subjected to moving average processing with a certain width Δz, and furthermore, the focus curve I ^* _i which is smoothed by the signal correction processing unit 27a.
(Z) finally becomes as shown by the solid line in FIG. The moving average processing performed by the signal correction processing unit 27a is defined as (z−Δz / 2, z + Δz /
The value obtained by averaging the light amounts of the measurement points within the range 2) is replaced with the value at the height z. Height z that gives the maximum value of focus curve I ^* _i (z) obtained after moving average
= BF is the focal position BF _i (Δz) at each point P _i when the moving average width is Δz.

As described above, the signal correction processing unit 27a
Each point P in the exposure area (imaging plane) of the projection optical system 11_i(I
= 1 to N), each focus obtained by moving average processing
Curve I^* _iFind (z). Then, the processing system 27
The focus position detection unit 27b includes a signal correction processing unit 27a
Each focus curve I obtained by^* _iBased on (z),
Each focus curve I^* _iHeight z = maximum value of (z)
BF is determined as the best focus position. This
As a result, each of the projection optical systems 11 within the exposure area (imaging plane)
Point P_iFind the best focus position at (i = 1 to N)
Can be Then, the focus position detection unit 27b
Each point P in the exposure area (imaging plane) of the optical system 11 _i(I = 1
After determining the best focus position in (1) to (N), the projection optical
A predetermined offset amount that does not depend on the
Adds the position to the scum position to determine the final best focus position.
Each is calculated. Thereby, the focus position detection unit 2
7b finally calculates the image plane shape of the projection optical system 11.
You.

Thereafter, the focus position detecting section 27b
Are displayed on a display unit 28 such as a CRT monitor. The calculation of the image plane shape of the projection optical system 11 need not be performed by the focus position detection unit 27b, and an image plane shape calculation unit may be provided inside the processing system 27 independently of the focus position detection unit 27b. .

Now, the above focus position detecting section 27b
If the result obtained by the above is not good, that is, if the image plane shape as the image plane characteristic (or the image formation characteristic) on the image formation plane of the projection optical system 11 is deteriorated, the focus position detection unit 27b Processing system 2 based on the result obtained by
A correction amount calculation unit 27c provided inside the projection optical system 7 corrects an image plane characteristic (or an imaging characteristic) of the projection optical system 11 on an image plane, such as an optical element such as a lens element constituting the projection optical system 11. The correction amount of the element (L ₁ , L ₂ ) is calculated. Then, based on the calculated result, the correction amount calculation unit 27c transmits an optical element (L ₁ ,
Move L ₂₎ along the optical axis of the projection optical system 11 or the optical element (L _1, L ₂₎ passed along a plane perpendicular to the optical axis of the projection optical system 11 moves, or the optical element (L _1,, By moving L ₂ ) so as to be inclined, the image plane characteristic (or the image forming characteristic) on the image forming plane of the projection optical system 11 is corrected.

As described above, in this example, each point P _i (i = 1) in the exposure area (imaging plane) of the projection optical system 11 is obtained by the moving average processing performed by the signal correction processing section 27a.
Focusing curves I ^* _i (z) obtained for each of the steps (.about.N) are obtained by substantially eliminating measurement error components due to residual aberration of the projection optical system 11. As a result, the image plane shape based on the best focus position at each point P _i (i = 1 to N) in the exposure area (imaging plane) of the projection optical system 11 finally obtained by the focus position detection unit 27b Can substantially match the result of the image plane shape of the projection optical system 11 obtained by the CD focusing method.

As described above, when the step of correcting the image plane characteristic (or the image forming characteristic) on the image forming plane of the projection optical system 11 is completed, the process proceeds to the exposure step (photolithography step). When the wafer 12 as a photosensitive substrate is set on the image forming plane of the projection optical system 11, the illumination optical system (1
8), the reticle 9 is illuminated. Then, the pattern of the reticle 9 is transferred to the wafer 1 via the projection optical system 11.
2 (exposed).

The wafer 12 that has undergone the above-described exposure process (photolithography process) is subjected to a development process, followed by an etching process for removing portions other than the developed resist,
After the etching step, a step of removing unnecessary resist is performed. Then, the steps of exposure, etching, and resist removal are repeated to complete the wafer process. After that, when the wafer process is completed, in the actual assembling process, dicing is performed to cut and divide the wafer into chips for each baked circuit, bonding for providing wiring and the like to each chip, and package for packaging each chip Finally, a semiconductor device such as an LSI is manufactured through the respective steps such as ringing. In the above, an example of manufacturing a semiconductor device such as an LSI by a photolithography process in a wafer process using an exposure apparatus has been described.
A semiconductor device such as a liquid crystal display element, a thin film magnetic head, and an image pickup element (CCD or the like) can be manufactured by a photolithography process using an exposure apparatus.

Here, the smoothing processing performed by the signal correction processing section 27a for noise reduction may be performed before or after the moving average processing, or both. If the width Δz for taking the moving average is changed depending on the resist condition and the illumination condition, the influence of the aberration remaining in the projection optical system on the measurement of the image plane shape by the FC mechanism can be changed. Therefore, by optimizing the width Δz for taking the moving average, it is possible to further reduce the error from the measured value of the image plane shape by the CD focusing method.

Due to recent advances in simulation technology, a resist image simulation and an aerial image intensity distribution simulation when aberrations remain can be easily performed, and their accuracy is considerably high. A sample of aberrations generated in the exposure area of the projection lens and a large number of arbitrarily given aberrations in the image plane are prepared, and the two types of simulations are performed based on the samples, and the simulation is performed at an arbitrary position in the exposure area. The focal position and further the image plane shape can be determined and the correlation can be examined.

According to the previous studies, if the width Δz for taking the moving average of the focus curve I _i (j) or I _i (z) is optimized according to the resist conditions and the illumination conditions, and the image plane shape is measured, A sufficiently practical measurement accuracy of the image plane shape was obtained simply by generating only an offset (offset due to autofocus) independent of the projection optical system for each of the above conditions. An example of the method will be described below.

The difference of the movement between the focal position at each point P _i at the average width of Δz BF _i (Δz), the focal position CD _i determined by CD focus method at that point P _i, Is an offset amount of the focal position at that point P _i. Therefore, the average offset amount OF _M for all points
(Δz) is Becomes

In order to minimize the variation of the offset amount OF _i (Δz) at each point, the variance S ² shown below may be minimized. The dispersion S ² can be minimized by optimizing Δz with respect to a sample of aberration occurring in the exposure area of a large number of projection lenses or an aberration arbitrarily given in the image plane. Alternatively, data may be obtained by performing measurement using an actual projection lens. _Assuming that the optimized value is Δz ^* , the average offset OFM (Δz ^* ) at that time is an offset that does not depend on the aberration of the projection optical system. Further, the offset amount OF _i (Δz ^* ) − OF _M (Δz ^* ) still remaining at each point is a non-optimizable amount for CD _i remaining by the present moving average method. That is, the best focus at each image point by the CD focus method cannot be predicted by the present moving average method. However, if the variance S ² is sufficiently small, this non-optimizable amount can be substantially ignored.
In fact, in the case where only a small aberration as the projection lens does not remain, the variance S ² can be a small value.

[0044] Here, as an application example of the configuration of the present embodiment, a method of determining the residual aberration amount of the projection lens from the relationship of the moving average width Δz and BF _i. Regard the moving average processing before the focus curve I (z), its maximum value BF deviation and the maximum value BF ₀ to the center signal curves left and right symmetry of the image plane ₀ and is largely dependent on the occurrence of longitudinal aberration doing. In the case where the spherical aberration (including that in which the amount of generation is different between the meridional direction and the sagittal direction) remains, the left-right asymmetry is obtained as shown in FIG. 3; On the other hand, when only the field curvature (including a difference in the amount of generation between the meridional direction and the sagittal direction) occurs as the longitudinal aberration, the symmetry maintains perfect symmetry, and the position of the maximum value is the optical axis. It will be off-axis from above.

Therefore, if the longitudinal aberration is 0, the moving average width Δz
BF _i (Δz) does not change due to
_i (Δz) becomes constant in the image plane. Further, the offset factors of the CD focus method and the moving average method are only components that do not depend on the projection optical system. Therefore, if the relationship between the moving average width Δz and BF _i (Δz) is obtained by any configuration of the present embodiment, the function (curve) reflects longitudinal aberration.
In the adjustment of the projection lens, bringing the curve closer to a straight line corresponds to bringing the spherical aberration closer to 0, and bringing the value of the entire BF _i closer to the value on the optical axis reduces the field curvature to 0.
Will be approached.

In this embodiment, the case where the moving average is calculated has been described. The method of calculating the moving average is such that the convolution integral with the intensity distribution (predetermined rectangular function) having a certain width in the Z direction and an intensity value of 1 is obtained. The result is equivalent to the result (slit scan processing). In addition, with respect to a continuous focus curve I _i (z) obtained by simply smoothing the discrete focus curve I _i (j), the center point of the two heights at which the light amount becomes smaller than the maximum value of the light amount distribution. , Each point P
It is also possible to determine the focus position BF _i of by _i. At this time, the slice level of the light amount is selected to be appropriate according to each resist condition and illumination condition. Further, in order to confirm the correlation thus obtained, it is easy to measure the image surface shape using the actually printed resist image and compare it with the image surface shape measured by the FC mechanism.

The reference pattern of the index plate according to the present embodiment has four patterns in order to reduce the influence of the directionality of the reticle pattern.
A line-and-space pattern having different directions is used. As a result, the measured light quantity signal has an average image plane shape of the meridional image plane and the sagittal image plane. However, there may be a case where it is desired to obtain an image plane shape by a pattern in an arbitrary direction, such as a difference between a meridional image plane shape and a sagittal image plane shape. In this case, a method is conceivable in which a plurality of FC mechanisms are simply prepared and optical systems and the like are provided so as to overlap each other, such that a reference pattern having a different pattern direction is attached to each FC mechanism. However, as a less wasteful method, an image plane shape by a pattern in an arbitrary direction can be obtained as follows using only one FC mechanism.

That is, as shown in FIG. 5A, the reference patterns of the index plate are 0, 45, 90,
Line-and-spaces LS ₀ , LS ₄₅ , LS ₉₀ , and LS ₁₃₅ oriented in the direction of 135 degrees are arranged in the first, second, third, and fourth quadrants of the reference pattern, respectively. On the other hand,
As for the reticle pattern, as shown in FIG.
A reflection region that sufficiently covers the entire conjugate region of the reference pattern is provided at the center, and a transmission region that sufficiently covers the conjugate region of each line and space pattern of the reference pattern is provided therearound.

At the time of detecting the focal position, as shown in FIG. 6, the position of the reticle stage or wafer stage is set so that the mark area in the direction to be measured is reflected by the reflection area on the reticle surface. That is, when measuring the average image plane shape, so that the entire conjugate region of the reference pattern is completely covered by the reflection region of the reticle pattern,
The measurement is performed by shifting to the position of e in FIG. Also, for example, when measuring the image plane shape in the 0-degree direction, only the conjugate area of the line and space LS ₀ oriented in the arrangement direction in the 0-degree direction is covered by the reflection area of the reticle pattern,
Other line and space LS ₄₅ , LS ₉₀ , LS
_The measurement is performed by shifting to the position of FIG. 6A so that the conjugate area ₁₃₅ is completely covered by the transmission area of the reticle pattern. Similarly, when measuring the image plane shapes in the 45, 90, and 135 degree directions, the measurement is performed by shifting to the positions of b, c, and d in FIG. 6, respectively.

The L of the four types of patterns used
The locations of S ₀ , LS ₄₅ , LS ₉₀ and LS ₁₃₅ are different,
Since it is sufficiently smaller than the exposure area of the projection lens, it can be treated as being measured at almost one point in the exposure area.
By such a method, the positions and shapes of the meridional image plane and the sagittal image plane can be obtained, and therefore, the average image plane can be obtained, and the astigmatic difference (as) due to the pattern direction in the entire exposure area can be obtained. The amount generated can also be determined.

Further, when the relationship between the moving average width Δz and BF _i (Δz) is obtained from these four types of patterns, the function (curve) reflects the longitudinal aberration in each pattern direction. Therefore, for each pattern direction, bringing the curve closer to a straight line is equivalent to bringing the spherical aberration closer to 0, and bringing the overall BFi value closer to the value on the optical axis makes the field curvature closer to 0. And can be used as an index for adjusting the projection lens including the difference due to the pattern direction.

In this patent, the moving average processing is performed for each height z.
It is defined that the value obtained by averaging the contrast of the measurement points within the range of (z−Δz / 2, z + Δz / 2) with the value at the height z is sufficient to obtain sufficient accuracy. I have. However, depending on the resist characteristics in the future, it is a matter of course that optimum weighting may be applied to each measurement point when the average is taken in order to further improve the accuracy. Similarly, with respect to an intensity distribution having an arbitrary width in the Z direction and convolution integration with an intensity distribution having an intensity value of 1, the intensity value within an arbitrary width in the Z direction may be optimally weighted.

[0053]

According to the present invention, an image plane detecting apparatus which can accurately determine the position and shape of an image plane which would be obtained when test exposure is performed on a wafer without performing test exposure. was gotten. Therefore, the quality of the image plane shape of the projection optical system can be determined without actually performing exposure. Moreover, if the pattern of the reticle is finally transferred onto the photosensitive substrate via the projection optical system by using the image position detecting method by the image position detecting apparatus of the present invention, a good semiconductor device can be manufactured. Can be.

[Brief description of the drawings]

FIG. 1 is a sectional view showing one embodiment of the present invention.

FIG. 2 is a plan view showing a reference pattern.

FIG. 3 is a diagram showing a focus curve based on a light amount signal;

FIG. 4 is a diagram showing focus curves before and after moving average processing;

FIG. 5 shows (a) a reference pattern according to another embodiment;
(B) Plan view showing reticle pattern

FIG. 6 is a plan view showing a reference pattern imaged on a reticle pattern in another embodiment.

FIG. 7 is a cross-sectional view showing a reticle cross-section by a CD focus method.

FIG. 8 is a diagram showing a bottom line width distribution curve by a CD focus method.

[Explanation of symbols]

DESCRIPTION OF SYMBOLS 1 ... Mercury lamp 2 ... Elliptical reflecting mirror 3 ... Mirror shutter 4 ... Input lens 5 ... Fly-eye lens 6 ... Aperture stop 7 ... Condenser lens 8 ... Mirror for optical path bending 9 ... Reticle 10 ... Reticle stage 11 ... Projection optical system 12 ... Wafer 13 Wafer holder 14 Wafer stage 20 Lens system 21 Two-branch light guide 22 Mirror 23 Condenser lens 24 Index plate 24a Reference pattern 25 Light receiving sensor 26 Sensor 27 Processing system 27a Signal correction processing unit 27b ... focus position detecting section 27c ... correction amount calculating unit 28 ... display unit 30 ... illumination system 31 ... focus detection system 32, 33 ... driving system L _1, L ₂ ... optical element _{LS 0, LS 45, LS 90} , LS 135 … Line and space pattern

Claims (7)

[Claims] A projection optical system for forming an image of a pattern formed on a reticle; a reference pattern disposed at a position conjugate with the reticle with respect to the projection optical system; and an illumination system for illuminating the reference pattern. Projecting a reference pattern image illuminated by the illumination system onto the pattern surface of the reticle via the projection optical system, and reflecting light from the projection image of the reference pattern via the projection optical system and the reference pattern A detection system that receives light and detects the amount of the received light beam; a moving device that moves the reference pattern along the optical axis direction of the projection optical system; While moving along a plurality of positions in the optical axis direction of the projection optical system, the light amount distribution in the optical axis direction is obtained based on the light amount at each position detected by the detection system. A processing system for determining a best focus position of an image of the reticle pattern by the projection optical system based on a light amount distribution in a direction, wherein the reticle pattern of the reticle pattern is based on a light amount distribution in the optical axis direction. When obtaining the best focus position of the image by the projection optical system, the processing system corrects the shape of the light amount distribution to a desired shape in order to correct a measurement error component due to aberration remaining in the projection optical system. An imaging position detection device comprising a unit. 2. The method according to claim 1, wherein the correction processing unit performs a process of obtaining a moving average of the light amount distribution or a process of performing convolution integration of the light amount distribution and a predetermined function. Image position detection device. 3. The shape of the image plane of the projection optical system, wherein the processing system obtains the best focus height at a plurality of positions in the image plane of the projection optical system orthogonal to the optical axis. The imaging position detecting device according to claim 1, wherein: 4. The reference pattern has a plurality of line-and-space patterns in different directions of pattern arrangement, and the reticle pattern includes a plurality of line-and-space patterns of the reference pattern. A reflection area covering a conjugate area of the line and space pattern in at least one direction of the pattern with respect to the projection optical system; and at least one direction of the plurality of line and space patterns of the reference pattern The line
And a transmission area that covers a conjugate area of the projection optical system of an and space pattern.
4. The imaging position detecting device according to 2 or 3. 5. A projection optical system for forming an image of a pattern formed on a reticle; a detection system for photoelectrically detecting an optical characteristic of an image forming surface of the projection optical system; and a detection signal from the detection system. A processing system for determining a best focus position of the image by the projection optical system based on the above, wherein the processing system determines the best focus position of the image by the projection optical system, the aberration remaining in the projection optical system An imaging position detection device, comprising: a correction processing unit that deforms the shape of the detection signal into a desired shape in order to correct a measurement error component caused by the measurement. 6. A method for manufacturing a semiconductor device, comprising a projection exposure step of projecting and exposing a pattern formed on a reticle onto a photosensitive substrate via a projection optical system, wherein the projection optical system is provided prior to the projection exposure. And an adjusting step of adjusting the projection optical system after the detecting step, wherein the detecting step is conjugate with the reticle with respect to the projection optical system. A light beam from a reference pattern disposed at a position is projected onto the pattern surface of the reticle via the projection optical system, and reflected light from a projected image of the reference pattern is received via the projection optical system and the reference pattern. Detecting a light amount of the received light beam; and moving the reference pattern along a plurality of positions in the optical axis direction of the projection optical system while detecting the light amount. Processing step of obtaining the light amount distribution in the optical axis direction based on the information obtained in step (a), and obtaining the best focus position of the image of the reticle pattern by the projection optical system based on the light amount distribution in the optical axis direction. And the processing step comprises: measuring a measurement error due to aberration remaining in the projection optical system when obtaining a best focus position of an image of the reticle pattern by the projection optical system based on the light amount distribution in the optical axis direction. A method of manufacturing a semiconductor device, comprising: a correction processing step of correcting a shape of the light amount distribution to a desired shape in order to correct a component. 7. A method for manufacturing a semiconductor device, comprising a projection exposure step of projecting and exposing a pattern formed on a reticle onto a photosensitive substrate via a projection optical system, wherein the projection optical system is provided prior to the projection exposure. And an adjusting step of adjusting the projection optical system after the detecting step. The detecting step includes: an optical characteristic of an imaging surface of the projection optical system. And a correction process for deforming the shape of the detection signal photoelectrically detected by the photoelectric detection into a desired shape in order to correct a measurement error component due to aberration remaining in the projection optical system. And a method for manufacturing a semiconductor device.