JPH0943862A - Projection aligner - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a projection aligner constituted so that a precise focusing position is detected and a high-resolution pattern image is easily obtained by illuminating a previously selected acutal element pattern with focus detection light and detecting a focusing state. SOLUTION: A reticle pattern 83 formed on a reticle is illuminated with an exposure light beam transmitted through a half mirror 80. Then, luminous flux reflected from the pattern 83 is reflected on the mirror 80 and guided to a focal plane detection optical system 81. The optical system 81 is constituted of an objective lens 13, an aperture diaphragm 82 arranged on the back focal plane thereof and a CCD array sensor 86 through a relay lens 15. Then, the image of the pattern 83 is formed on the sensor 86. Besides, a similarly constituted focal plane detection optical system 81 is provided left. By such constitution, the precise focusing position is detected and the high-resolution pattern image is easily obtained even in the case that the precise focusing position cannot be measured even when a TTLAF mark provided outside an actual element area is used.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a projection exposure apparatus,
In particular, in the field of semiconductor device manufacturing, the present invention relates to a projection exposure apparatus called a stepper having an automatic focus adjustment function, a so-called autofocus function, when a reticle circuit pattern is repeatedly reduced and projected and exposed on the surface of a semiconductor wafer.

[0002]

2. Description of the Related Art In recent years, semiconductor elements, LIS elements, super LS
Due to the demand for finer patterns and higher integration of I-elements,
Imaging (projection) with high resolution in projection exposure equipment
Since the optical system is required, the high NA of the imaging optical system
The depth of focus of imaging optics is becoming shallower.

[0003] In addition, the wafer has a planar processing technology,
Some thickness variation and bending must be tolerated. Usually, for the correction of the wafer bending, the wafer is placed on a wafer chuck that is processed so as to guarantee the flatness on the order of submicron, and the back surface of the wafer is vacuum-sucked to correct the flatness. However,
The thickness variation within one wafer, the suction method, and the deformation of the wafer caused by the progress of the process cannot be corrected no matter how much the plane of the wafer is corrected.

[0004] For this reason, the effective depth of focus of the optical system is further reduced because the wafer has irregularities in the screen area where the reticle pattern is reduced and exposed.

Therefore, in the reduction projection exposure apparatus, an effective automatic focusing method for matching the wafer surface with the focal plane (the image plane of the projection optical system) has become an important theme.

As a method of detecting a wafer surface position in a conventional reduction projection exposure apparatus, a method using an air microsensor,
There is known a method (optical method) in which a light beam is incident on a wafer surface in an oblique direction without passing through a projection exposure optical system and a positional deviation amount of the reflected light is detected.

On the other hand, in this type of projection exposure apparatus, a change in the ambient temperature of the projection optical system, a change in the atmospheric pressure, a temperature rise due to a light beam applied to the projection optical system, or a temperature rise due to heat generation in an apparatus including the projection optical system. Causes the focus position (image plane position) to move, which must be corrected. Therefore, the focus position of the projection optical system is measured by measuring the ambient temperature change and the atmospheric pressure change with a detector, or measuring part of the temperature change and the atmospheric pressure change in the projection optical system with the detector. Calculated and corrected.

However, in this method, since the focus position of the projection optical system is not directly measured, the projection optical system is determined from the detection error of the detector that measures temperature and atmospheric pressure, and the amount of temperature change and atmospheric pressure change. There is a drawback that it is difficult to detect the focus position of the projection optical system with high accuracy due to an error included in the calculation formula which is an approximate formula when the focus position is calculated and corrected.

As a method of overcoming such a problem, a so-called through-the-lens auto-focus system (TTL) is used in which the exposure lens is directly passed through to detect the focus surface.
AF) has been devised.

FIG. 12 is a schematic diagram of a projection exposure apparatus disclosed in Japanese Patent Laid-Open No. 1-286418. In FIG. 12, 7 is a reticle, which is held on the reticle stage 70. The circuit pattern on the reticle 7 is transferred to the wafer 9 on the xyz stage 10 by the reduction projection lens 8.
The image is reduced to ⅕ and imaged, and exposure is performed. FIG.
In No. 2, a reference plane mirror 17 whose upper surface and the mirror surface of the wafer 9 are substantially aligned is arranged at a position adjacent to the wafer 9.

Further, the xyz stage 10 has an optical axis direction (z) of the projection lens 8 and an in-plane (x, y) orthogonal to this direction.
It can be moved with, and of course it can be rotated around the optical axis.

The reticle 7 is illuminated within the screen area where the circuit pattern is transferred by the illumination optical system shown by elements 1 to 6 in FIG.

The light emitting portion of the mercury lamp 1 which is a light source for exposure is located at the first focal point of the elliptical mirror 2, and the light emitted from the mercury lamp 1 is condensed at the second focal point of the elliptic mirror 2. ing. An optical integrator 3 having its light incident surface positioned at the second focal position of the elliptical mirror 2 is provided, and the light exit surface of the optical integrator 3 forms a secondary light source. The light emitted from the optical integrator 3 serving as the secondary light source has its optical axis (optical path) bent at 90 ° by the mirror 5 via the condenser lens 4.

Incidentally, 55 is a filter for selectively taking out light having an exposure wavelength, and 56 is a shutter for controlling exposure. The exposure light reflected by the mirror 5 passes through the field lens 6 and illuminates the screen area on the reticle 7 where the circuit pattern is transferred.
The mirror 5 is configured to partially pass the exposure light, for example, 5 to 10%. The light passing through the mirror 5 reaches a photodetector 50 for monitoring fluctuations in the amount of light from the light source 1 through a lens 52 and a filter 51 that transmits the exposure wavelength and cuts excess light for photoelectric detection. .

In the figure, elements 11 to 12 form a known off-axis autofocus optical system.
Reference numeral 11 denotes a light projecting optical system, and the light flux which is the non-exposure light emitted from the light projecting optical system 11 intersects the optical axis of the reduction projection lens 8. It is assumed that the light is condensed and reflected on a point on the reference plane mirror 17 (or the upper surface of the wafer 9). The light beam reflected by the reference plane mirror 17 enters the detection optical system 12.

Although illustration is omitted, a position detecting light receiving element is arranged in the detection optical system 12, and the position detecting light receiving element and the reflection point of the light flux on the reference plane mirror 17 are arranged to be conjugate. The positional deviation of the reduction projection lens 8 of the reference plane mirror 17 in the optical axis direction is measured as the positional deviation of the incident light beam on the position detecting light receiving element in the detection optical system 12.

The positional deviation of the reference plane mirror 17 measured by the detection optical system 12 from a predetermined reference plane is transmitted to the autofocus control system 19. The auto focus control system 19 includes an xy with the reference plane mirror 17 fixed.
A command to move in the z direction is given to the drive system 20 that drives the z stage 10. Also, when detecting the focus position by TTL, the autofocus control system 19 uses the reference mirror 1
7 is driven up and down in the optical axis direction (z direction) of the projection lens 8 near a predetermined reference position. The position control of the wafer 9 during exposure (the wafer 9 is arranged at the position of the reference plane mirror 17 in FIG. 12) is also performed by the autofocus control system 19.

The focus position detecting optical system of the reduction projection lens 8 will be described.

In FIGS. 13 and 14, 7 is a reticle, 2
A pattern portion 1 formed on the reticle 7 has a light-shielding property. Reference numeral 22 denotes a light shielding portion sandwiched between the pattern portions 21. Here, when the focus position (image plane position) of the reduction projection lens 8 is detected, the xyz stage 10 moves in the optical axis direction of the reduction projection lens 8.

The reference plane mirror 17 is positioned on the optical axis of the reduction projection lens 8, and the reticle 7 is illuminated by the illumination optical systems 1 to 6 shown in FIG.

First, the case where the reference plane mirror 17 is on the focus surface of the reduction projection lens 8 will be described with reference to FIG. The exposure light that has passed through the transmission part 22 on the reticle 7 is
It is condensed and reflected on the reference plane mirror 17 via the reduction projection lens 8. The reflected exposure light follows the same optical path as the outward path, is focused on the reticle 7 via the reduction projection lens 8, and passes through the light transmitting portion 22 between the pattern portions 21 on the reticle 7. At this time, the exposure light does not eclipse the pattern portion 21 on the reticle 7, and the entire light flux is
Passing through the transparent part of.

Next, the case where the reference plane mirror 17 is at a position displaced from the focusing surface of the reduction projection lens 8 is shown in FIG.
4 will be described. The exposure light passing through the transmission portion of the pattern portion 21 on the reticle 7 reaches the reference plane mirror 17 via the reduction projection lens 8, but the reference plane mirror 17
Is not on the focus surface of the reduction projection lens 8, the exposure light is reflected by the reference plane mirror 17 as a spread light beam.

That is, the reflected exposure light follows an optical path different from the outward path, passes through the reduction projection lens 8 and is not condensed on the reticle 7, and is displaced from the focus surface of the reduction projection lens 8 of the reference plane mirror 17. It reaches the reticle 7 as a light beam having a spread corresponding to the amount. At this time, part of the exposure light is vignetted by the pattern portion 21 on the reticle 7, and all the light beams cannot pass through the light transmitting portion 22. That is, there is a difference in the amount of light reflected through the reticle when the focus surface is matched and when it is not.

The optical path after the light flux of the exposure light reflected by the reference flat mirror 17 described in FIGS. 13 and 14 passes through the reticle 7 will be described with reference to FIG.

The exposure light transmitted through the reticle 7 reaches the mirror 5 through the field lens 6. Since the mirror 5 has a transmittance of about 5 to 10% with respect to the exposure light as described above, part of the exposure light reaching the mirror 5 passes through the mirror 5 and passes through the imaging lens 13 and the field stop 14. On the surface of. At this time, the surface of the reticle 7 on which the pattern is present and the field stop 14 are separated by the field lens 6 and the imaging lens
It is in a conjugate position through.

The exposure light that has passed through the opening of the field stop 14 enters the light receiving element 16 through the condenser lens 15.

If necessary, a filter 51 for selectively transmitting only the exposure light is provided on the front surface of the light receiving element 16, and an electric signal corresponding to the amount of the incident exposure light is output.

A method of detecting the focus position (image plane position) of the reduction projection lens 8 using the signal output of the light receiving element 16 will be described below.

The drive system 20 drives the xyz stage 10 on which the reference plane mirror 17 is mounted in the optical axis direction of the reduction projection lens 8 around the zero point of measurement preset by the off-axis autofocus detection system 12. To do.

At this time, the position signal in the optical axis direction of the reference plane mirror 17 (autofocus measurement value z) measured by the autofocus detection system 12 at each position and the exposure light reflected by the reference plane mirror 17 are received. The relationship between the outputs obtained from the focal plane (image plane) detection system 18 by receiving light by the element 16 and converting it into an electric signal is as shown in FIG. At this time, the signal of the detection system 18 eliminates the influence of the fluctuation of the light source 1, so that the light from the light source 1 that has passed through the mirror 5 is detected by the photodetector 50 via the light source light amount monitoring optical system (52, 51). Then, the reference light amount detection system 53 generates a light source light amount monitor signal. Then, the signal of the focal plane detection system 18 is standardized by this monitor signal to correct it.

When the reference plane mirror 17 is located on the focus plane of the reduction projection lens 8, the output of the focal plane detection system 18 shows a peak value. The autofocus measurement value z ₀ at this time is used as the focus position of the reduction projection lens 8 when the wafer 9 is exposed using the reduction projection lens 8.
(Or, the preset focus position is corrected based on the measured value z _0. ) The off-axis auto focus detection system 10, 12, 19 is set at the focus position of the reduction projection lens 8 thus determined.
Set the reference position of. The actual best printing position of the wafer is a value obtained by giving an offset from the reference position by an amount in consideration of the values such as the coating thickness of the wafer and the step amount. For example, when exposing a wafer using a multi-layer resist process, only the uppermost portion of the multi-layer needs to be burned, so that the resist surface of the wafer and the reference position substantially coincide.

On the other hand, when the exposure light reaches the substrate sufficiently with the single-layer resist, the focus of the wafer matches the substrate surface, not the resist surface. In this case, an offset of 1 μm or more is provided between the resist surface and the reference position. It is not rare that there exists. Such an offset amount is specific to the process and is given as an offset different from that of the projection exposure apparatus. It suffices for the apparatus itself to accurately obtain the focus position of the reduction projection lens 8 itself by the above-described one method, and the offset amount is projected on the autofocus control system 19 and the drive system 20 only when necessary. It may be input in advance via a system controller (not shown) of the exposure apparatus.

The detection of the focus position z ₀ may be determined by the peak of the output of the focal plane detection system 18, but various other methods are possible. For example, in order to further increase the detection sensitivity, a slice level SL is set to a certain ratio with respect to the peak output, and the focus value is obtained by knowing the autofocus measurement values z ₁ and z ₂ when the output of this slice level SL is shown. Position

[0034]

[Equation 1] Alternatively, a method of finding the peak position using a differential method may be considered.

The advantage of such a TTL autofocus system is that the ambient temperature change of the projection exposure optical system, the atmospheric pressure change, the temperature rise of the projection optical system due to the exposure light beam, and the accompanying temporal change of the focus occur. The point is that measurement and correction are applied.

[0036]

However, this T
The TLAF method has the following problems in practical use.

In the example of FIGS. 12 and 13, TTLA
Although it is described that the F mark is provided on the optical axis of the exposure lens for autofocusing, such mark arrangement is not allowed in an actual reticle. This is because the transfer characteristic of the exposure lens is generally best on the optical axis, and therefore the actual element pattern on the actual reticle is provided within a predetermined exposure area including the optical axis. Therefore, as the TTLAF mark, a mark (reference mark) that is most likely to produce the output characteristics as shown in FIG. 8 must be determined in advance and placed at a predetermined image height outside the actual element region. (Figure 1
6 KM, KS) This has the following drawbacks.

In the TTLAF method shown in FIG. 12, the detection light passes through the exposure lens twice reciprocally. Therefore, the longitudinal aberration (spherical aberration, astigmatism, curvature of field, axial chromatic aberration, etc.) of the lens occurs twice, and the detection output curve of the focal plane shown in FIG. 8 is easily distorted due to the influence thereof. Further, the heat load is accumulated on the lens as the exposure progresses, but in such a state, the off-axis aberration of the lens tends to vary.

At this time, if there is a difference between the aberrations in the sagittal direction and the meridional direction, or if there is a large difference between the sagittal direction and the meridional direction due to fluctuations in the aberration,
For example, as shown in FIG. 16, when a sagittal mark (KS) and a meridional mark (KM) are simultaneously arranged as a reference mark at a certain position off the axis, and it is attempted to detect a common focus for them, the best of each of them is detected. The focus is separated into the position ZS and the position ZM as shown in FIG. 17, and as a result, the combined output thereof has a double peak. And
There may be a problem that the center of focus cannot be obtained because there are four intersections with the slice value SL.

According to the present invention, the TT provided outside the actual element region
An object of the present invention is to provide a projection exposure apparatus capable of easily obtaining a high-resolution pattern image by detecting an accurate focus position even when the LAF mark cannot be used to measure the accurate focus position.

[0041]

The projection exposure apparatus of the present invention comprises:
Light from a detection mark provided on the reticle surface illuminated by the focus detection light passes through the exposure lens at least once and is received by the photodetector, and the focus state is detected by the output from the photodetector. In a projection exposure apparatus having a detection system, in order to illuminate the actual element mark that can be used as the detection mark with the focus detection light, select a desired actual element pattern from the actual element pattern in advance, The feature is that the selected actual element pattern is illuminated with focus detection light to detect the focus state.

In particular, the selection is performed by inspecting the actual element pattern of the reticle set in the projection exposure apparatus, and the selection is performed by inspecting the actual element pattern of the reticle before being set in the projection exposure apparatus. It is characterized by being performed by doing.

[0043]

FIG. 1 is a schematic view of a main part of a first embodiment of the present invention.

In this embodiment, in the sequential exposure method (stepper), TTLA using an actual element pattern on the reticle is used.
The schematic configuration of the F method is shown. The reticle illumination system (path from mercury lamp 1 to field lens 6) is shown in Fig. 1.
The description is omitted because it is almost the same as 2. Further, a mechanism for detecting the path through which the exposure light and the TTLAF detection light are focused on the wafer 9 or the reference plane mirror 17 from the reticle 7 through the exposure lens 8 and the height of the wafer stage 10 to perform the Z drive. The same is true for 20 mag.

In this embodiment, the exposure light beam transmitted through the half mirror 80 illuminates the actual element pattern 83 on the reticle. Then, the reflected light flux from this pattern 83 is reflected by the half mirror 80 and guided to the focal plane detection optical system 81. The focal plane detection optical system 81 includes the objective lens 13, an aperture stop 82 placed on the focal plane thereafter, and the relay lens 1.
After that, the pattern 83 is formed by the CCD array sensor 86, and the pattern 83 is imaged on the CCD 86. A focal plane detection optical system 81 having a similar structure is also provided on the left side.

The method for automatically searching for the optimum pattern prior to the focal plane detection will be described below. The hardware configuration is shown in FIG. 1, and the search sequence is shown in FIG.

The CCD array sensor 86 transfers the two-dimensional image data of the pattern (reticle pattern) 83 to the read image determination circuit 100 in step 1 of FIG. The image judgment circuit 100 analyzes the sent image data and calculates the optimum pattern and the degree of coincidence which have already been known by then (step 2). The optimum pattern here means a pattern in which the output curve in FIG. 8 is the smoothest and the focus center can be easily calculated even if the lens 8 has a difference in aberration in the sagittal direction and the aberration in the meridional direction, for example.

Since the degree of coincidence with the optimum pattern is low, TTLA
When it is determined that the pattern is inappropriate for F measurement (step 3), the signal is sent to the controller 84.
The controller 84 commands the detection optical drive system (drive system) 85 to determine another actual element pattern on the reticle (step 4), and sets the focal plane detection optical system (detection system) 81 on the reticle 7. Along the xy direction. The search drive of the detection system 81 is performed according to the movement mode already set in the drive system 85. When the optimum pattern is found by performing such pattern searching several times, the controller 84 detects the optimum pattern at that position.
Fix 1 Then, the focal plane detection optical system 18 is instructed to start the focal plane measurement operation (step 5). This measurement operation is almost the same as in FIG.

As an auto-focus signal, the transmitted light flux of the reticle pattern 83 is once focused on the reference mirror 17 (at the same height as the wafer surface) by the action of the exposure lens 8 and then reflected and exposed again. It is preferable to use the total transmitted light amount when passing through the pattern 83 and the pattern 83 (that is, the total integrated light amount on the CCD 86). That is, CC
D86 performs the same function as the light receiving element 16 of FIG. In this embodiment, in order to correct the output fluctuation of the light source 1, the reference light amount detection system 53 including the mirror 87 is configured in another path. The specific operation is the same as that of FIG. The focal plane detection optical system 18 calibrates the signal of the CCD 86 based on this signal. Alternatively, since the image of the pattern 83 is formed on the CCD 86, the contrast of this image may be measured or the rising of the pattern edge portion may be used as the evaluation amount.

FIG. 2 is a plan view showing how an optimum pattern for TTLAF is searched for from the actual element patterns on the reticle 7. The intersection of the X axis and the Y axis coincides with the optical axis of the exposure lens 8. A pair of focal plane detection optical systems 81 and 81 'are provided on the left and right, but at least one of them may be provided.

First, the field of view is aligned with the mark P1.
The mark P1 is a vertical line (meridional mark) arranged on the X axis, but a scribe line exists in the visual field of the detection system. The scribe corresponds to a boundary line when a plurality of real element regions (chips) are formed in the reticle, but has a step as compared with the real element region. Therefore, if auto-focusing is performed using this pattern, the best focus surface of the real element will be displaced by that amount.

Although the image of the reticle pattern is directly determined in this embodiment, the apparatus determines that this is not the optimum pattern. Then, the drive system 85 operates, the optical system moves to the positions of the marks P2, P3, P4, and the mark P5, the same pattern matching is performed, and the best actual element pattern capable of good focal plane detection is selected. are doing.

As described above, in the present embodiment, TT is set in the real element area where the aberration state of the exposure lens 8 is better than in the area outside the real element area.
We are doing LAF. That is, good focus detection is performed by a configuration in which a pattern that can be used as an AF mark is detected in advance from the actual element pattern and the focus position is detected using the pattern.

FIG. 3 is a schematic view of the essential portions of Embodiment 2 of the present invention.

The basic configuration is the same as that of the first embodiment shown in FIG. 1, and in this embodiment as well, the reticle pattern is directly observed in order to find the optimum TTLAF pattern. Unlike the first embodiment shown in FIG. 1 in which the image information of the reticle is used for determination, in this embodiment, the pattern is determined based on the distribution information of the diffracted light.

The illumination luminous flux for detection emitted from the fiber 90 is reflected by the first half mirror 91, and the illumination lens 92
After being reflected by the second half mirror 93, the light is focused on the rear focal plane 94 of the objective lens 96. An aperture stop 95 is placed here to limit the luminous flux. After passing through the objective lens 96 having the rear focal plane 94 as the focal plane, the mirror 9
After being reflected at 7, the reticle pattern 83 is converted into a parallel light beam for Koehler illumination.

The pattern diffraction reflection light from the reticle pattern 83 returns along the same optical path as this illumination light flux. That is, a diffraction pattern is formed on the back focal plane 94 of the objective lens 96. After passing through the half mirror 91 by the illumination lens 92, this diffraction pattern is re-imaged on the two-dimensional sensor (CCD) 98. The diffraction pattern is read by the CCD 98.

The diffraction pattern analysis circuit 110 is the CCD 98.
Measure the distribution of the pattern diffracted light read by
The degree of coincidence is calculated by comparing with the distribution of the diffracted light of the optimum pattern which has already been known.

FIGS. 4A to 4D are explanatory views of examples of actual element patterns and their diffraction patterns. The vertical repeating pattern 111 and the horizontal repeating pattern 112 are the most prominent circuit patterns. The distribution of these diffracted lights is one-dimensionally distributed in the direction orthogonal to the pattern longitudinal direction. Reference numerals 113 and 114 indicate a state in which their diffraction patterns are distributed on the CCD array sensor. 115 is a contact hole array, and its diffraction pattern is 11
It is two-dimensionally distributed like 6. Reference numeral 117 denotes an obliquely repeating pattern in the direction of 45 degrees, and its diffraction pattern 118.
Is distributed diagonally. The fine line widths of the actual elements are arranged as memory cells in the central portion of the chip, and the relatively thick line widths are present as a wiring pattern in the peripheral portion of the chip. The distribution direction, the interval, etc. of the diffraction pattern differ depending on the arrangement direction, pitch, and line width of these actual element patterns, and a desired actual element is selected by reading the difference in distribution.

FIG. 10 shows a reticle pattern searching sequence of the second embodiment shown in FIG. According to step 6, C
The diffracted light pattern image captured by the CD 98 is matched by the diffracted pattern diffraction circuit 110 in step 7. As a result, when the degree of coincidence with the optimum pattern obtained in advance is low and it is determined in step 8 that the pattern is inappropriate for TTLAF measurement, the signal is sent to the controller 84. In step 9, the controller 84 commands the detection optical drive system 85 to determine another actual element pattern on the reticle, and moves the detection system. When the optimum pattern is found by performing such pattern searching several times, the controller 84 fixes the detection optical system at that position. Then, the focal plane detection optical system 81 is instructed,
Enter the measurement operation of the focal plane.

In this embodiment, the light from the fiber 40 is used as the illumination light flux for measuring the focal plane. The light beam to be received is the same as that in the first embodiment of FIG. 1, and the light beam that is reflected by the reference plane mirror 17 and returns through the reticle pattern 83 is used. After passing through the objective lens 96 and the aperture stop 95, this light flux passes through the half mirror 93 and is once focused on the surface of the field stop 120 by the action of the relay lens 15. This field stop 1
Reference numeral 20 limits the pattern area to be measured on the reticle surface. The shape of the opening may be square, rectangular or circular. The transmitted and divergent light flux becomes a convergent light flux again by the action of the condenser lens 121, and is received by the light quantity sensor 122.
Subsequent measurement operations are the same as in the embodiment of FIG.

In the present embodiment, the light which is transmitted through the half mirror 91 by a part of the fiber 90 is used as the reference light. This light is received by the element 50 and detected by the reference light amount detection system 53 as in the embodiment of FIG.

Incidentally, the wavelength of the light guided from the fiber 90 may be the exposure light or another wavelength for pattern searching. However, exposure light is used for focal plane measurement.

FIG. 5 is a schematic view of the essential portions of Embodiment 3 of the present invention. The basic system configuration is the same as that of each embodiment shown in FIGS.

However, the following points are different. In the embodiment of FIGS. 1 and 3, a direct image of a pattern and diffracted light information are taken in order to search for a reticle pattern, and these are preliminarily acquired by TTL.
Pattern known to have good AF waveform (optimal pattern)
Compared to In the present embodiment, unlike these, the output of the focal plane detection system is taken even at the pattern searching stage. Then, when an output curve as shown in FIG. 8 is obtained, and a curve in which the focus center can be obtained most easily is obtained, that pattern is selected and used as the AF mark after that.

More specifically, the light quantity sensor 122 receives the reflected light from the reference plane mirror 17, and the focal plane detection system 1
The output of 8 is fetched in step 11 of FIG. The ideal shape of the output curve is stored in advance in the AF output curve determination circuit as a template.
Matching between the captured AF output curve and the stored ideal curve is performed. If the degree of matching is not sufficient, the determination is made in step 13, and in step 14, the focal plane detection optical system 81 is moved by the detection optical drive system 85 via the controller 84, and the next actual element pattern 83 is measured. If the degree of matching is good, the output from the focal plane detection optical system 18 is immediately input to the auto focus detection system 19 in step 15, and the reference focus position is set.

FIG. 6 is a schematic view of the essential portions of Embodiment 4 of the present invention. The searching method of the fourth embodiment is almost the same as that of the third embodiment shown in FIG. The feature is that a reticle and a conjugate surface are provided inside the exposure illumination system, and a field stop whose aperture is limited is arranged therein. Then, the opening is moved to select a pattern illumination area on the reticle, thereby selecting a desired actual element pattern.

In FIG. 6, reference numeral 300 is a field stop, and the reticle 7 with respect to the illumination system lens 4 and the field lens 6.
It is placed on the plane conjugate with. Due to the opening of the field stop 300, a part of the actual element pattern of the reticle 7 becomes a light source
Is illuminated by The light passing through the reticle 7 passes through the exposure lens 8 and reaches the mirror 17. The light reflected by the reference plane mirror 17 again passes through the exposure lens 8 and the reticle pattern 83. Then, an image is formed again on the field diaphragm 300 by the action of the field lens 6 and the illumination system lens 4. The light flux that has passed through this opening passes through the half mirror 301 this time, and then is converged on the light amount sensor 122 by the function of the condensing optical system 121 and photoelectrically converted.

The output of the light amount sensor 122 is input to the focal plane detection system 18. The subsequent signal processing is processed according to the sequence of FIG. 11 described above. However, FIG.
Although the focal plane detection optical system was driven in step 14 in the above description, the field stop driving system 302 drives the field stop 300 in the embodiment of FIG.

As the field stop 300 shown in this embodiment, a so-called masking blade used in an actual stepper for limiting the exposure area on the reticle may be used for this purpose. As described above, there is an advantage that the focal plane detection optical system 81 can be simplified by using the image forming system inside the exposure illumination system.

FIG. 7 is a schematic view of the essential portions of Embodiment 5 of the present invention.

In the above-described embodiments, the optimum pattern is selected from the patterns at the coordinates designated in advance on the reticle arranged in the exposure apparatus. On the other hand, Example 5
Now inspect the entire reticle with a scanning laser beam,
The optimum pattern is sought by analyzing the pattern diffracted light as in the second embodiment.

In FIG. 7, reticle library 310 is shown.
The reticle 7 carried out from the reticle 7 is detected by the optimum pattern selection unit 320 before it is sent to the exposure position.

The selection unit 320 will be described. Laser 3
The laser beam emitted from 11 is shaped into a predetermined beam diameter by a pinhole 312 and a beam expander 313, and then enters a polygon mirror 314. The polygon mirror 314 rotates in a plane orthogonal to the paper surface, and the reflected light is condensed on the reticle 7 by the action of the mirror 315 and the scanning lens 316, and is scanned in the direction orthogonal to the paper surface. The transmitted diffracted light produces a diffracted light distribution (see FIG. 4) on the back focal plane of the lens 317 by the light receiving lens 317. A CCD array sensor 97 is arranged on this surface to determine the distribution of diffracted light. The analysis circuit 110 measures the degree of coincidence between the optimal diffracted light distribution of the TTLAF mark (actual element) and the measured distribution.

The reticle 7 moves in a direction intersecting with the beam scanning (direction S in the drawing) in synchronization with the beam scanning. As a result, the entire surface of the reticle is inspected. This makes reticle 7
A desired TTLAF real element mark on the entire surface of the reticle is discriminated, and in order to identify the coordinate position of this mark on the reticle, the scanning position of the beam and the moving position of the reticle 6 are calculated based on the output of a synchronization sensor (not shown). To do. Information on the desired actual element mark position is sent to the controller 84. Reticle 7
Is transferred to the exposure apparatus shown on the left side of FIG.
Based on the desired actual element mark position signal sent to the controller 84, the focal plane detection system 81 moves to perform the focal plane detection using the desired actual element mark as in the previous embodiment.

In the above description, the beam scanning system and the diffracted light receiving system may be turned upside down with respect to the reticle 7. Further, it may be configured to illuminate from above and receive light from above, or conversely, to illuminate from below and receive light from below. Oblique incidence optical systems are also within the scope of the present invention. As for the focused state of the beam, in general, focusing on the pattern surface of the reticle (lower surface in the drawing) improves the positional resolution, but this is not necessarily the case.

That is, by intentionally blurring or inspecting with a parallel beam, the diffraction pattern can be measured evenly in a wider region within the beam diameter at one time, which is more effective in some cases. As a special case, it is advisable to match this inspection beam diameter with the size of the field of view for signal acquisition at the time of subsequent focal plane detection.

According to this embodiment, the TT from the front surface of the reticle is
Since the actual element pattern serving as the optimum mark for LAF can be selected, more accurate focal plane detection can be performed.

FIG. 18 is a schematic view of the essential portions of Embodiment 6 of the present invention.

In the above-described embodiments, the final AF detection optical path passes through the exposure lens 8 twice. Example 6
Uses a configuration in which the exposure lens 8 used in Example 5 in FIG. 6 can be passed only once.

That is, in FIG. 6, the light beams emitted from the illumination systems 1 to 6 including the mirror 301 pass through the selected reticle pattern 83, and then are condensed on the reference plane mirror 17 by the exposure lens 8. On the reference flat mirror 17, a mark 400 (reference mark) patterned by a transmissive portion and a non-transmissive portion is formed. As the shape of the reference mark 400, for example, one of the patterns shown in FIG. 4 is used.
Only the light flux passing through both the reticle pattern 83 and the reference mark 400 is condensed by the condensing optical system 81, and the light receiving element 1
Detected at 22. The amount of light changes depending on the in-focus state of both, so the focal plane detection system 81 detects the focal plane.

When the field stop 300 is searched for, the position of the actual element pattern image of the reference mirror 17 is set to the reference mirror 1.
Move on 7. In each of the above-described embodiments, since the reference mirror 17 was a full-face mirror, there was no problem. However, in this embodiment, since the pattern 400 is provided in a part of the reference mirror 17, it is synchronized with the movement of the field stop 300. Then the reference mirror 1
It is necessary to move 7 on the pattern 400 so that the actual element pattern images match.

Further, in the case of the present embodiment, since the shape of the reference mark 110 is fixed, it is inevitably easy to select an actual element pattern having a shape close to this. Then, the AF detection is performed using the selected actual element pattern.

In this embodiment, the reticle is illuminated from above and the light is received on the surface (reference mark surface) substantially conjugate with the wafer, but the scope of the present invention is not limited to this structure. Conversely, the wafer side may be illuminated and the reticle side may receive light.

FIG. 19 is a schematic view of the essential portions of Embodiment 7 of the present invention.

In each of the above embodiments, the light path for the AF detection is such that the light beam passing through the reticle passes through the exposure lens 8 twice and then passes through the reticle again to be received.

On the contrary, in the present embodiment, the detecting light beam reciprocates from the wafer side to the exposure lens 8 and is received by the wafer side.
That is, the focal plane detection optical system 81 shown in FIG. 5 is provided inside the wafer stage 10 in FIG. The reference mirror 17 has a pattern 400 as in the sixth embodiment shown in FIG.
have. The light from the pattern 400 passes through the exposure lens 8 to form an image of the pattern 400 on the lower surface of the reticle, is reflected by the reticle 7 and returns to the reference mirror 17 again to form a re-image of the pattern 400.

That is, since the actual element pattern of the reticle 7 is expected to have a function merely as a reflecting surface (or a bare surface), a reflection of at least the same size as the image of the pattern 400 formed on the lower surface of the reticle 7 is expected. It is desirable that the pattern is a surface (or a blank surface) pattern. The stage 10 moves in the x and y directions until a real element pattern like this is obtained, and determines the desired pattern position.

At this time, the AF output determination circuit determines the proportion of the reflection surface (or the bare surface) in the image size of the pattern 400 formed on the lower surface of the reticle 7, that is, the amount of reflected light. I am doing it.

AF detection after the desired actual element pattern is determined, the stage 10 is moved in the z direction, and pattern 4
00 and its re-imaging coincidence are measured in the same manner as in the above-mentioned embodiment.

In the case of this embodiment, the TTLAF mark 40 is used.
0 is formed on the reference plane mirror 17, and the reticle surface functions only as a reflecting surface of this image. Therefore, a perfect chrome reflective surface without a pattern can be the optimum pattern of the reticle. If such a reflecting surface is not present on the reticle, an actual element region that does not affect the focal plane measurement as much as possible is selected as the optimum pattern.

[0092]

As described above, according to the present invention, even if the accurate focus position cannot be measured even if the TTLAF mark provided outside the actual element region is used, the accurate focus position can be detected. A projection exposure apparatus that can easily obtain a pattern image with high resolution can be achieved.

In particular, according to the present invention, the following effects can be obtained by providing a means for searching for and selecting the optimum pattern for TTLAF from the actual element patterns and performing the focal plane detection using it.

(B) Since TTLAF can be performed at a low image height with a small aberration of the exposure lens, highly accurate focal plane detection can be performed. In particular, (b) it is less likely to be affected by aberration (thermal aberration) generated by heat when an exposure load is applied, so that a focus change of the lens occurring during that time can be accurately followed and detected.

(C) Since the focus of the actual circuit pattern can be captured, the offset (the detected focal plane and the focal plane measured by actually printing the pattern on the wafer are compared to the case where AF is performed with a separately provided mark). Deviation amount) is extremely small.

[Brief description of drawings]

FIG. 1 is a schematic view of a main part of a first embodiment of the present invention.

FIG. 2 is a diagram showing pattern searching on the reticle shown in FIG.

FIG. 3 is a schematic view of the essential portions of Embodiment 2 of the present invention.

FIG. 4 is an explanatory diagram showing an actual element according to the present invention and its diffraction pattern.

FIG. 5 is a schematic view of the essential portions of Embodiment 3 of the present invention.

FIG. 6 is a schematic view of the essential portions of Embodiment 4 of the present invention.

FIG. 7 is a schematic view of the essential portions of Embodiment 5 of the present invention.

FIG. 8 is an explanatory diagram showing the output of the focal plane detection system according to the present invention.

FIG. 9 is a diagram showing a reticle pattern searching sequence according to the first embodiment of the present invention.

FIG. 10 is a diagram showing a reticle pattern searching sequence according to the second embodiment of the present invention.

FIG. 11 is a diagram showing a reticle pattern searching sequence in the third and fourth embodiments of the present invention.

FIG. 12 is a schematic view of a main part of a conventional projection exposure apparatus.

13 is an explanatory diagram of a part of FIG.

14 is an explanatory view of a part of FIG.

FIG. 15 is an explanatory diagram of a part of FIG.

FIG. 16 is an explanatory view of a layout of a conventional TTLAF mark on a reticle.

FIG. 17 is an explanatory diagram of a collapsed focal plane detection system output.

FIG. 18 is a schematic view of the essential portions of Embodiment 6 of the present invention.

FIG. 19 is a schematic view of the essential portions of Embodiment 7 of the present invention.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 light source 4 illumination system lens 6 field lens 7 reticle 8 exposure lens 9 wafer 10 wafer stage 17 reference plane mirror 18 focal plane detection system 19 auto focus detection system 20 drive system 50 light receiver 53 reference light amount detection system 81 focal plane detection optical system 83 reticle pattern 84 controller 85 detection optical drive system 86 CCD array sensor 111, 112, 115, 117 pattern

Claims (3)

[Claims] 1. At least one light from a detection mark provided on a reticle surface illuminated by focus detection light.
Can be used as the detection mark with the focus detection light in a projection exposure apparatus that has an autofocus detection system that passes through a double exposure lens, receives it with a photodetector, and detects the focus state by the output from this photodetector. In order to make it possible to illuminate a real element mark, a desired real element pattern is selected from the real element patterns in advance, and the selected real element pattern is illuminated with focus detection light to detect the focus state. Projection exposure system. 2. The projection exposure apparatus according to claim 1, wherein the selection is performed by inspecting an actual element pattern of a reticle set in the projection exposure apparatus. 3. The projection exposure apparatus according to claim 1, wherein the selection is performed by inspecting an actual element pattern of the reticle before being set in the projection exposure apparatus.