JPH0361804A - Board alignment device - Google Patents

Abstract

(57) [Summary] This bulletin contains application data before electronic filing, so abstract data is not recorded.

Description

Detailed Description of the Invention [Industrial Application Field] The present invention relates to an alignment device for aligning a flat object to the focal position of an imaging optical system, and in particular for aligning a flat object to the focal position of an imaging optical system. The present invention relates to an alignment device suitable for setting a gap between a tool and a mask in an exposure apparatus for transferring a circuit pattern.

[Prior Art] There is an X-ray exposure device that uses soft X-rays as a device that transfers a pattern on a mask to a photosensitive layer on a wafer with a mask and a semiconductor wafer (hereinafter referred to as wafer) in close proximity. .

In this X-ray exposure system, if the gap between the mask and the wafer, that is, the proxy tee gap, deviates from a predetermined value, a magnification error or distortion will occur in the transferred pattern on the wafer, and when printing a large number of layers, , the newly printed pattern may be misaligned with respect to the layer printed on the wafer. FIG. 1 is a schematic structural diagram of a conventional X-ray exposure apparatus for explaining this.

An electron beam (
(not shown) to generate X-rays from there, and make the generated X-rays enter the mask 3 at a distance of 2, and transfer the pattern on the mask 3 to the resist of the wafer 6''. The leg of the perpendicular drawn from point 2 to the surface of mask 3 is designated as point 4, and the point where this perpendicular passes through the wafer surface is designated as 7. Note that the point 4 usually coincides with the center of the surface of the mask 3. Assume that the gap between the mask 3 and the wafer 6 at a point 5 on the mask 3 is gk, and the distance between the points 4 and 5 is a. At this time, point 5 on the mask 3 is transferred as point 8 on the X halo, but the horizontal shift amount Δ of point 8 with respect to point 5 is approximately Δ=g (a/L)... (1)
It can be expressed as

Here, the horizontal shift amount Δ when the value of g has an error by δg
The deviation amount δ△ from the equation (1) is δΔ#δg (a/L
) ・・・・・・0 that can be expressed by (2) For example, L
= 200+nm, a = 20mm.

If δg is 4 μm, δΔ is 0.4 μm, which results in a deviation of 0.4 μm from the predetermined transfer position. This amount of δΔ is unacceptable in an X-ray exposure apparatus that performs transfer with a submicron line width (1 μm or less).

Traditionally, the gap has been set using two methods: (1) mechanically touching the mask ε wafer and then separating it by the distance of the gap; and (2) using two capacitive gap sensors to measure the mask surface sliding. The following methods are considered: (1) a method of detecting and setting the gap with the eight sides separated horizontally, and (3) a method of detecting and setting the gap using a capacitive gap sensor fixed to the mask. It is being

[Problems to be Solved by the Invention] Among these, method (1) may damage the mask or wafer, and if dust exists between the mask and the wafer, the gap will not reach the specified value, and the wafer The drawback is that the gap setting accuracy is poor due to the elasticity of the mask. The method (2) cannot measure the position of the mask and wafer in the exposed state, so it has the drawback that it cannot measure the gap during exposure. Furthermore, method (3) has the disadvantage that it is time-consuming because the gap sensor must be calibrated for each mask. Furthermore, common drawbacks using capacitive gap sensors include the inability to determine the gap between the mask and the sensor at the position where the mask pattern is transferred, and the poor electrical conductivity of SOS (Silicon on the Fire). It cannot be used for other substrates.

The present invention solves the conventional drawbacks and sets the distance between two substrates such as a mask and a wafer to a target value by adjusting the position of the mask or wafer using an imaging optical system such as an objective lens. It is an object of the present invention to provide a positioning apparatus that employs a method of detecting the same state as that during exposure and transfer and controlling the mask or wafer so that it comes to the focal position of the imaging optical system.

A further object of the present invention is to provide an apparatus that reduces false detections that occur during the process of setting a mask or wafer at the focal position of an imaging optical system.

[Means for Solving the Problems] The alignment device of the present invention includes a light-transmitting first substrate partially having reflective properties, and a light-reflecting second substrate disposed parallel to the first substrate. In the apparatus for aligning the first substrate and the second substrate in a direction perpendicular to the surfaces of the substrates so that they face each other at a predetermined interval, the second substrate is irradiated with a light beam through the first substrate. an imaging optical system arranged to receive reflected light from each of the first substrate and the second substrate; and a first focusing plane of the imaging optical system based on the reflected light;
Focus detection means for outputting a detection signal when the substrate is positioned at either of the first focusing plane and a second focusing plane spaced apart from the imaging optical system by the predetermined distance; a first driving means for relatively moving the first substrate and the imaging optical system in the optical axis direction; and a second driving means for moving the second substrate and the imaging optical system relatively in the optical axis direction. a drive means; a control means for selecting a state in which the first drive means is operated based on a detection signal of the focus detection means; and a state in which the second drive means is operated based on the detection signal; It is something to be prepared for.

In one preferred embodiment of the present invention, the control means includes:
After controlling the first driving means so that the first substrate approaches the first focal plane of the imaging optical system from one direction, the second substrate approaches the second focal plane of the imaging optical system. The second driving means is configured to be controlled so as to approach from one direction.

[Function] In the substrate alignment device of the present invention, the focus detection means detects the focus positions of the first and second substrates with respect to the two focus planes of the imaging optical system. Ru. For example, the focus detection means may include a first signal indicating whether the first substrate is located at the first focusing plane, and a second signal indicating whether the second substrate is located at the second focusing plane. obtained separately from 1st
When the signal is out of focus, the first driving means moves the first substrate in the optical axis direction relative to the imaging optical system, and when the second signal is out of focus, the second driving means moves the first substrate in the optical axis direction relative to the imaging optical system. The two substrates are moved, and selection of their movement states is performed by the control means.

In this way, since one focus detection means is selectively used for alignment of the first substrate and alignment of the second substrate, there is no possibility that both substrates may be set at incorrect positions due to generation of false signals. There is no collision between the two substrates.

[Example] Next, embodiments of the present invention will be described using the drawings.

FIG. 2 is a block diagram of a positioning device according to an embodiment of the present invention, in which the diameter of the laser beam emitted from the laser light source 9 is changed by a beam diameter converter 10, and an optical deflector 1
1. The light beam deflected by the optical deflector 11 passes through a three-beam optical system 12 and a relay optical system 13, passes through a beam splitter 14, and enters a condensing scanning lens 15 as an imaging optical system of the present invention. The light beam becomes a light beam 20 and a light beam 21 that are imaged on two different surfaces by the condensing scanning lens 15 of the irradiation means having such a configuration, and the light beam 20 is focused on two different flat objects, that is, the light beam 20 is focused on the surface 16a of the mask 16 on the pattern 16b side. (hereinafter referred to as pattern surface 16a), and a light beam 21 coaxial with light beam 20 is used to detect the position of surface 17a of wafer 17. Therefore, when the mask 16 and the wafer 17 form a predetermined gap, the light beam 20 forms an image on the pattern surface 16a of the mask 16, and the light beam 21 forms an image on the surface 17 of the wafer 17.
The image is formed on a.

The focal points of the light beams 20 and 21 are scanned by the action of the optical deflector 11 so as to move within a plane parallel to the pattern plane 16a and the surface 17a. In FIG. 2, the light beams 20 and 21 are incident on the opaque pattern 16b of the mask 16. Since the opaque pattern 16b is formed from a two-layer thin film of titanium and gold, it has a high light reflectance and acts as a mirror surface.

FIG. 3 shows how the incident point of the light beam 20.21 moves to a position away from the pattern 16b. In this case, the light beam 20.21 hits not only the mask 16 but also the wafer 1.
It also reaches the surface 17a of 7. Surface 17a of wafer 17
acts as a mirror surface.

At this time, the transparent pattern surface 16a of the mask 16 also reflects a portion of the incident light. The light reflected by the pattern surface 16a of the pattern 16b1 of the mask 16 or the wafer surface 17a passes through the condensing scanning lens 15 again, and is branched to the beam splitter 14 along the optical path ε at the time of incidence through the beam splitter 25. sent to.

The beam splitter 25 functions to split the incident light beam into a mask focus detection optical system 26M for the light beam 20 and a wafer focus detection optical system 26W for the light beam 21. The focus detection optical system 26M is an optical system for forming an image of the anti-sword beam of the light beam 20 and detecting the focal shift of the light beam 20 formed on the mask 16. 7) The information light of the spot image is sent to the photoelectric conversion unit 2.7M. - The wafer focus detection optical system 26W similarly images the reflected light of the light beam 21 to form a light beam 21 on the wafer 17. Spots of
This is an optical system for detecting the focal shift of the image, and information light of this spot image is sent to the photoelectric conversion section 27W.

23 is a basic oscillator, and a drive circuit 2 for the optical deflector 11
4 and a signal processing unit 28, and the signal processing unit 28 obtains position information of the light beam 20.21 on the mask 16 and the wafer 17 based on this clock signal C1. The output signals of the charging conversion sections 27M and 27W are sent to the processing section 28, and the signal processing section 28 detects the distance between the mask 16, the condensing scan 1, and the nons 15. still,
The focusing scanning lens 15 is configured so that the imaging positions of the two light beams 20 and 21 are separated by a distance of 71 times in the light @force direction.
The numerical aperture (N, A) is sufficiently large and the depth of focus is also smaller than the gap.

Figure 4 shows the optical system in Figure 2 specifically, and the laser light source in Figure 2 is 2.
:ζ is a semiconductor laser, and the beam diameter converter 10 is an amplimeter 1.・However, when implementing the present invention, it is necessary to use a semiconductor laser for L6 as well!
c <, a gas 17 laser such as a helium neon laser can also be used. Furthermore, when a single nose light source that emits a nearly parallel beam is used, a beam expander is used as the beam diameter converter 10.

In FIG. 4, the light beam made parallel by the collimator lens 10 is angularly deflected by a galvanometer serving as an optical deflector 11 which rotates and oscillates within the plane of the paper in the direction of an arrow 33. The deflected parallel light beam is condensed into a small beam by a lens 34 and reaches a beam splitter 35. In this embodiment, the optical deflector 11 shown in FIG. 2 is shown as a galvanometer mirror, but other types of deflectors such as rotating polygon mirrors, ultrasonic deflectors, etc. may also be used.

Now, the beam splitter 35 separates the condensed beam from the lens 34 into two beams, and the two divided beams are as follows:
After being condensed at positions 40 and 41, they are each reflected by reflecting mirrors 36 and 37, and then combined by a beam splitter 43 so that they share the same optical system. Here, the parallel plate glass 42 is inserted so that the optical path length of the light beam passing through the reflecting mirror 37 is shorter than the optical path length of the light beam passing through the reflecting mirror 36 in an equipolarized manner. The thickness of the glass 42 determines the distance between the imaging positions of the two light beams 20, 21. The light beams that are coaxially superimposed by the beam splitter 43 are turned into substantially parallel light beams again by the lens 44, and are imaged by the condensing scanning lens 15 via the beam splitter 14. Although the above-mentioned lens 44 produces a substantially parallel light beam, the light beam from the reflecting mirror 36 to the lens 4
4 and the optical path length from the reflecting mirror 37 to the lens 44 are different because of the parallel flat glass 42.
The three superimposed beams are not parallel. The light beam exiting the condensing scanning lens 15 becomes a light beam 20 and a light beam 21, each of which is imaged as a minute spot L in different planes. The distance between the imaging planes of the light beams 20 and 21 is determined by the thickness of the parallel flat glass 42, the lens 44 and the condensing scanning 1/1
Since it is determined by the focal length of the lens 15, it can be set to a predetermined value by selecting an appropriate parallel flat glass. This predetermined value is g&II'P. g is the mask 1 to be set
6 and the wafer 17.

If the influence of the setting error of the two optical path lengths between the beam splitter 35 and the beam splitter 43 is considered, C is 1, the thickness of the parallel flat glass 42 may be adjustable; An example is shown in which the running path is differentiated by inserting glass, but other methods, such as reflecting mirror 3
It is also conceivable to create a travel path difference by changing the bending angle of the optical axis between the beam splitter 7 and the beam splitter 43.

Now, the light beams 20 and 21 are driven to scan the mask 16 and the wafer 17 simultaneously in the left and right directions in the plane of the paper as indicated by arrows 45 by the galvano main unit 11.

The light beams 20 and 21 are reflected by the mask 16 or the wafer 17, enter the condensing scanning lens 15 in the opposite direction, are reflected by the beam splitter 14, and reach the lens 47. The reflected light beams 20 and 21 passing through this lens 47 are split into two by a beam splitter 25. beam splitter
The reflected light beam 20.21 reflected by the beam splitter 25 is further split into two by a beam splitter 48M, and images are formed on predetermined imaging planes 53MA and 53MB, respectively.

At this time, the imaging planes 53MA and 53MB are predetermined at positions where the image of the spot light on the mask 16 is formed when the light beam 20 is imaged on the mask 1'6. That is, the imaging plane of the light beam 20 and the imaging planes 53A and 53B are determined to be conjugate.

Of course, the light reflected by the wafer 17 of the light beam 21 is also reflected by the lens 4.
7 and is reflected by the beam splitter 25, but the imaging plane of the luminous flux 21 and the imaging planes 53MA, 53MB are not conjugate, and the spot image of the luminous flux 21 on the wafer 17 is reflected by the imaging planes 53MA, 53MB. The image is formed on a different surface than the

On the other hand, among the light beams 20.21 reflected by the mask 16 and the wafer 17, the light beam that has passed through the beam splitter 25 is further split into two by the beam splitter 48W, and is split into two parts at a predetermined imaging plane 53WA, The image is formed on 53WB.

At this time, the imaging planes 53WA and 53WB are predetermined at positions where a spot image is formed on the wafer 17 when the light beam 21 is imaged on the wafer 17. That is, the imaging plane of the light beam 21 and the imaging planes 53WA and 53WB are determined to be conjugate.

Of course, the light beam 20 reflected by the mask 16 also passes through the beam splitter 25.
The spot image on 6 is formed on a plane different from the image planes 53WA and 53WB.

Now, behind the predetermined imaging plane 53MA, there is a grating member 49M.
A is arranged, and a grating member 49MB is arranged in front of a predetermined imaging plane 53MB. Lattice member 49MA, 49MB
are gratings with the same period ', and a pattern is formed on a transparent flat plate in which light transmitting parts and non-transmissive parts are periodically arranged in a stripe shape.

Further, the distance between the grating member 49MA and the imaging surface 53MA and the distance between the grating member 49MB and the imaging surface 53MB are set to be equal to each other. Therefore, the imaging planes 53MA, 53M
When a spot image is formed on B, a spot image slightly larger than the spot size on the imaging planes 53MA and 53MB is formed on the grating member 49MA and the grating member 49MB. Furthermore, the spot image moves on the imaging planes 53MA, 53MB by the arrows 50MA, 5 due to the -11 vibration across the galvano.
Although it moves back and forth like 0MB, the grid members 49MA, 4
The 9MB is arranged so that the direction in which the gratings are arranged coincides with the reciprocating direction of the spot image.

Now, the light beams of the spot images that have passed through the transmission portions on the grating member 49MA and the grating member 49MB are directed to the condenser lens 51.
MA, the light is focused by the condenser lens 51MB onto the detector 27MA and the detector 27MB, and photoelectrically converted.

Although the detection system for the mask surface has been described above, the detection system for the square surface is also similar.

After the light beam 21 is reflected by the wafer 17, an image is formed in the same way as the reflected light of the light beam 20. These light beams are for detecting the surface of the sea urchin and are directed to a predetermined imaging surface 53WA.

The grating member 49WA and the grating member 49WB are made of a stripe pattern with the same period, similar to the detection system of the mask surface, and the image is formed on the predetermined image forming surface 53WB.
The distance between 3WA and the grating member 49WA is equal to the distance between the image forming surface 53WB and the grating member 49WB, and the distance between the image forming surface 53
They are placed one after another with respect to WA and 53WB.

The light flux of the spot image from the wafer 17 that has passed through the grating member 49WA% is transmitted through the condenser lens 51W.
The light is collected by the A% dyed optical lens 51WB and photoelectrically converted by the detectors 27WA and 27WB. Of course, the spot image of the light beam 21 transmitted through the beam splitter 25 moves along the image forming planes 53 WA and 53WB along the arrow 50WA.
, 50WB toothpick moves back and forth.

Here, the mask black spot detection optical system 26M in FIG.
M, grating member 49MA, 49MB5 condensing I/lens 51M
The wafer focus detection optical system 26W in FIG. 2 is composed of lenses 47 and 51MB in FIG.
Beam splitter-48W1 Grid member 49WA, 49W
B.

It is composed of condenser lenses 51WA and 51WB.

Next, the principle of focus detection for the mask 16 and wafer 17 will be explained using FIGS. 5 and 6. Since the basic principle of focus detection is the same for both masks and wafers, only focus detection for masks will be described in this embodiment.

FIG. 5(a) is a schematic diagram of the focus detection optical system of the mask, and shows the two light beams separated by the beam splitter 48M in FIG. 4 combined into one. Although the use of the beam splitter 14 is different from that shown in FIG. 4, this is done to simplify the explanation.

The parallel light beam from the laser light source 9 that has entered via the beam splitter 14 is converted into a light beam 20 from the condensing scanning lens 15.
The light is then emitted and is focused on a point 57 as a spot light. At this time, it is assumed that there is a reflective surface of the mask 16, for example, the pattern surface 16a. It may be the surface of the metal film forming the mask 16, the front side or the back side of the transparent part that is the support of the mask 16, or if the thickness of the transparent part is thinner than the depth of focus, an intermediate point therebetween. However, in the case of using a transparent part, it is necessary that no opaque part is formed at the position where the light flux is incident, and the thickness of the support of the mask 16 is 1 to 1. 2 μm is most commonly used, and even if you try to achieve a condensing state with a depth of focus as shallow as possible, the practical limit of the depth of focus is 1 to 2 μm, so in reality, the transparent film on both sides of the mask support is It is observed as a superposition of reflected light beams from In FIG. 5(a), for the sake of simplicity, both sides of the mask support are represented by one reflective surface, which is referred to as a pattern surface 16ah. It is considered that the patterned surface 16a creates a mirror image of the point 57 at a point 59, which is distanced by a distance d from the patterned surface 16a. Therefore, the light beam 20 is transmitted to the pattern surface 16
It is reflected by a, becomes a light beam 62, and enters the condensing scanning lens 15. The light flux 62, that is, the light flux of the spot image is
The beam splitter 14 and the lens 47 form a minute spot image on the grating member 49MB.

Points conjugate to point 57 coincide on predetermined imaging planes 53MA and 53MB. Therefore, the grid member 49MB and the imaging surface 53MA
, 53MB, and the heating distance between the condensing scanning lens 15 and the imaging lens 47 is f l + f 2, respectively. Le = 2 (f z / f +) 2 · d
It is expressed as =・= (3).

Further, as shown in the plan view in FIG. 5(b), the grating member 49MB has light transmitting parts 69a and non-transmissive parts 69b arranged periodically in a grating pattern, and fj, A part 69a and non-transmissive part 69b are periodically arranged. The convergence of the transmission part 69N) is determined by the spot size ε when the spot image is formed on the grating member 49MB as shown in FIG. 5(a).
are defined equally.

Now, when the light beam 20 scans the pattern surface 16a of the mask 16 in the direction of the arrow 45 due to the action of the galvanometer mirror 11, the spot on the grating member 49MB is scanned in the direction of the arrow 64. Therefore, as shown in FIG. 5(b), the grid member 4
The spot 70 on the 9MB moves in the direction of the arrow 71, and almost all of the incident light passes through the transmission section 69a.
Almost all of the light is blocked when passing through the non-transparent part 69b.

In FIG. 6(a), the vertical axis represents the magnitude of the charging output signal corresponding to the amount of transmitted light received by the detector 27MB during the imaging state of FIG. 5(a), and the horizontal axis represents time t. With something that
An AC signal waveform 73 having a frequency corresponding to the moving speed of the spot 70 is obtained. The difference between the maximum value and minimum value of the waveform 73 at this time is assumed to be 2□1.

Next, when the pattern surface 16a is at the position of point 57,
The formed spot images are on predetermined imaging planes 53MA, 53M.
B, and the spot size on the grid member 49MB becomes larger. Therefore, the grid member 49M shown in FIG. 5(b)
The spot size is slightly larger than the width of the transparent part 69a1 and the non-transparent part 69b of B, and the obtained signal waveform is as shown in FIG.
The waveform 75 shown in b) is obtained, and the difference between the maximum and minimum values is 2.2.
becomes smaller than 21. Furthermore, the pattern surface 16a is the point 5
When the spot image reaches the position including the point 60 which is closer to the condensing scanning lens 15 side by d from 7, the spot image becomes the grating member 49M.
The image is formed on A, and the spot size on the grating member 49MB becomes wider. Therefore, the transparent part 6 of the grid member 49MB
The amount of light passing through 9a becomes smaller, and the output signal of the detector 27MB becomes as shown in the waveform 77 shown in FIG. Become.

In addition, in Figure 6 (a), (b), and (c), the center of vibration of the signal does not change in each state, and light is always on the photoelectric element of the detector throughout the vibration of the spot image. There is. Therefore, even if the amplitude becomes small, that is, it deviates significantly from the focused state, a DC level photoelectric output signal corresponding to the center of vibration can be obtained.In other words, the photoelectric element of the detector is irradiated with bias light. Therefore, the responsiveness of photoelectric conversion of the detector can be improved.

FIG. 7 shows changes in the output signal of the detector 27MB. 7th
The figure is a graph in which the horizontal axis shows the position Z of the pattern surface 16a relative to the condensing scanning lens 15 in the optical axis direction, and the vertical axis (denotes half the difference between the maximum and minimum values of the output signal waveform, that is, the amplitude). Note that the 0 point on the horizontal axis represents the position when the point 57 on which the light beam 20 forms an image matches the pattern surface 16a, and the position when the pattern surface 16a is located closer to the lens 15 than the 0 point is the positive point. (When the pattern surface 16a is at the -d position, the amplitude of the output signal is al. When the pattern surface 16a coincides with the 0 point, the amplitude is It is a2.

Further, as shown in FIG. 6(C), the pattern surface 16a
When is at the +d position, the amplitude is a3.

Therefore, the pattern surface 16a is continuously
5, the change in amplitude becomes like waveform 79.

On the other hand, spot image detection by the grating member 49MA shown in FIG.
Also performed by 7MA. At this time, the detector 27MA
The amplitude change of the output signal is obtained as a waveform 83 that is similar to the waveform 79 and shifted in the Z direction with respect to the waveform 79, as shown by the broken line in FIG. The amplitude at the zero point of the waveform 83 is approximately equal to the amplitude a2 of the waveform 79, and reaches the maximum amplitude a at the position of 10d. Therefore, as shown in FIG. 4, the beam splitter 48M divides the beam into three beams with approximately equal intensity, and the two grating members 49MB and 49MA are placed in front and behind the predetermined imaging planes 53MA and 53MB (in the optical axis direction, in a predetermined path @ If placed apart by e, detector 27MB and detector 2
The amplitude changes of the 7 MA dual optical type output signal are as shown in waveform 79 and waveform 83 shown in FIG.

Next, a signal processing system for automatic focus detection using the above detection optical system will be explained based on the block diagram of FIG.

First, the photoelectric output signals of detectors 27MA and 27MB (FIG. 4) are amplified by preamplifiers 84A and 84B, respectively, and passed through a bandpass filter 85A.

It is man-powered by 85B. This band pass filter 85A,
85B selectively passes the fundamental frequency component of the signal, as shown in FIGS. 6(a), (b), and (c).

The modulation component detection circuits 86A and 86B input the output signals of the bandpass filters 85A and 85B and perform the output signals shown in FIG. 6(a).
(b) Output half the difference between the maximum value and the minimum value of the modulation component of the waveform shown in (c), that is, the amplitude value.

The signal CL from the reference oscillator 23 in FIG.
For 7B, the fs signal output from modulation component detection circuits 86A and 86B is sampled and the timing at which it is held is controlled.

The reason why the sample and hold circuits 86A and 86B are used here is that depending on the type of optical deflector, there is a time when it cannot be used for optical deflection, and the area on the mask where the light beam is reflected specularly is limited. This is because the time to obtain a signal for automatic focus detection is limited during the periodic operation of the optical deflector, and the modulation component detection circuit 86A,
C is sampled while the output signal from 86B is obtained, and held otherwise.

A subtracter 89 calculates the difference between the sampled and held Okayama power signals, and its output signal 90 becomes an S-curve signal 92 as shown in FIG.

In FIG. 9, the vertical axis represents the magnitude of the output signal 90 in FIG. 8, which becomes 05 when the amplitude of the output signal of the detector 27MB and the amplitude of the output signal of the detector 27MA become equal. Further, the horizontal axis in FIG. 9 is set to be the same as the horizontal axis of the graph in FIG.

At this time, the distance between the mask 16 and the condensing scanning lens 15 becomes a predetermined value, and the light beam 20 emitted from the condensing scanning lens 15
The relative positions of the mask 16 and the condensing scanning lens 15 in the optical axis direction may be controlled so that the zero point 0 of the S-curve signal 92 is obtained as soon as the beam is focused on the pattern surface 16a of the mask 16.

Here, in FIG. 8, modulation component detection circuits 86A and 86B
was supposed to detect the difference between the maximum and minimum values of the signal, but
The same effect can be obtained by extracting the magnitude of the modulation component by combining full-wave rectification or half-wave rectification with a low-pass circuit.

Although not shown in FIG. 8C, the output values of the sample and hold circuits 87A and 87B are added, and the subtracter
If you divide the output value of 9, even if the reflectance of the mask changes, the 9
The shape of the S-curve signal 92 in the figure remains unchanged, ie, automatic gain control can now be performed. Due to such automatic gain control, the straight line part 94 of the S curve signal 92 near the zero point O in FIG. 9 always has a constant slope, and the distance between the mask 16 and the condensing scanning lens 15 is Understood, this is advantageous when performing gap control.

Note that a detector 27WA is used to detect the focus of the light beam 21.

The output signal of 27WB is also processed by a circuit similar to the block in FIG.
The interval between

The above explanation was about focus detection when one light beam is incident on the mask 16, but in reality, in C, the light beam 21 for wafer detection is also incident on the mask 16 at the same time. Therefore, generation of the original signal as described below becomes a problem. The present invention prevents focusing (gap setting) from malfunctioning due to the generation of the original signal.

FIG. 10(a) is a diagram schematically showing how a light flux 20 for mask detection and a light flux 21 for wafer detection are converged through the same optical axis, and the light flux 20 is the same as the light flux 20 for mask detection. The light beam 21 is imaged on the target surface 95, and the light beam 21 is imaged on the target surface 97 of the wafer. Therefore, plane 95 and plane 97
The interval is a predetermined gap amount g& for mask detection. Let us now consider the case where the mask 16 with a mirror surface moves along the optical axis direction so as to approach the condensing scanning lens 15. FIG. 10(b) shows the magnitude of the detection signal of the focus detection system of the mask 16 shown in FIG.'s8 on the vertical axis.
The horizontal axis represents the position of the mask 16 in the optical axis direction. Lens mask 16! Move towards 5 and mask 16
When the mask 16 comes near the surface 96, the reflected light of the light beam 21 is focused near the surface 95, and a pseudo f8 signal like the S curve signal 100 appears, and only when the mask 16 comes near the surface 95 does it become true. An S-curve signal 101 appears. Therefore mask 1
When setting 6, the original signal 1oO is mistakenly converted to the true signal 101.
We must avoid viewing it as such. Therefore, in this embodiment, a comparator is provided to compare the constant level 104 with the magnitude of the S curve signal, and the mask 16 is
(towards the lens 15) (sometimes,
When the S-curve signal exceeds a certain level 104 for the first time, it is ignored as the original signal 100, and only when it exceeds the certain level 104 for the second time does it become a true S-curve signal 101.
The configuration is such that it recognizes that the signal 101 is true and starts using the true signal 101.

Although focus detection for the mask 16 has been described above, detection can be performed for the wafer 17 in the same manner as for the mask 16. However, quest! The major difference in the detection of No. 7 is that detection can be performed only in the transparent portion of the mask 16.

10(a) and (b) in focus detection of the wafer 17.
), since the two light beams 20 and 21 are incident, a longitudinal signal is generated. FIG. 10(C) shows the detection signal of the wafer 17, and when the surface of the wafer 17 is near the surface 97, a true S-curve signal 105 is obtained.
Near the surface 96, the reflected light of the light beam 20 for mask detection enters the focus detection system for the wafer 17, and a vertical signal 106 appears.

Therefore, as in the case of mask detection (it is necessary to distinguish the vertical signal 106 from the true signal), the mask 1 is
When moving closer to 6, a constant level of 10
9 and a comparator that compares the magnitude of the S curve signal,
It is only necessary to detect that the S-curve signal is a true S-curve signal 105 when it exceeds a certain level 109 for the first time.

In addition, the drive device for moving the mask 16 and wafer 17 is
When performing servo control using a curve signal, if the S curve signal becomes zero due to a malfunction in the servo movement, etc., move the wafer 17 away from the negative end in the Z direction again, and then perform true control as described above. Just look for the S curve signal 105.

FIG. 11 is a functional block diagram showing an embodiment of an apparatus for setting the gap between the mask 16 and the wafer 17 using the focus detection system for the mask 16 and the wafer 17 described above. There is a wafer 17 indicated by a broken line under the mask 16,
The mask attitude control unit 116 and the wafer attitude control unit 117 independently adjust the inclination and the position Z in the direction perpendicular to the paper surface IIJa. Three detection optical systems 112L, 1
12R and 112Y each consist of a detection optical system shown in FIG. 2, and two light beams 2 emitted from each detection optical system are
The intervals between the imaging positions of 0.21 are all set to a constant value g.

15L, 15R, and 15Y are, for example, microscopes shown as the condensing scanning lens 15 in FIG. 2 and FIG. With respect to the axes x and y of the orthogonal coordinate system, the condensing scanning lens 15L
The optical axes of C1 and 15R are on the X axis, and the optical axis of the C1 condensing scanning lens f5Y is on the y glaze. The positions of the three lenses are determined to approximately divide the circular mask 16 or wafer 17 into thirds, and the wafer 17 or mask 16 is
They are arranged to be eccentric from the center OI at an equal distance C from each other.

In FIG. 11, the line segment (X axis) connecting the centers of the condensing scanning lenses 15L and 15R does not pass through the center OL of the mask 16 (or wafer 17), but in the present invention, this line segment The condensing scanning lenses 15L and 15 pass through O1.
However, the condensing scanning lens 15 should be placed so that the line segment (X axis) passes through the center 01.
L.

15R, it is still preferable to set the distances from the center O1 to the centers of the respective condensing scanning lenses 15Y, 15L, and 15R to be approximately equal.

The output signals from the photoelectric detectors of the detection optical systems 112L, 112R, and 112Y are sent to the calculation unit 114 including the circuit shown in FIG.
15 and is used to set the gap between the mask 16 and the wafer 17 via the mask posture control section 116 and the wafer posture control section 117. Note that the attitude control units 116 and 117
is equipped with a mechanism that moves the mask 16 and wafer 17 in the vertical direction when rotated about the coordinate axes X and y, respectively. Before the mask 16 and the wafer 17 are set at a predetermined gap, the imaging positions of the light beams 20 for mask detection emitted from the three detection optical systems have the same Z coordinate value (position in the direction perpendicular to the plane of the paper). It has been adjusted as follows. The mask 16 is connected to a light beam 20 for mask detection of each detection optical system.
To place the mask 16 on the imaging position of
15Y), and the detection optical system 1
At least one light beam 20 of 12L, 112R, and 112Y is set to form an image on the mask 16. It is the first of the S curve signals of each detection processing circuit in the calculation section 114.
This is done by detecting a true S-curve signal 101 as shown in FIG. 0(b). Three detection optical systems
If the true S-curve signal 101 is detected by one detection optical system, then the true S-curve signal 101 is also detected by another detection optical system.
A search operation is performed to detect 1. This search operation is performed by changing the attitude of the mask 16 under the control of the calculation section 114 and the control section 115.

The following Table A shows the mask 16 to be controlled at the time of this search operation.
This figure shows the rotation axis and vertical movement in the rotation direction and Z-axis direction. In addition, Table A&: set.

R and Y are written by omitting 112L, 112R, and 112Y, and indicate the detection optical system in which the true S-curve signal was detected. Since each movement C can be performed independently by the posture side control device 16, each movement C is shown. However, the rotation direction of the rotation axis is shown as the direction of each coordinate axis x, y when viewed from the negative side to the positive side, and as shown in Figure 11, clockwise rotation is CW, counterclockwise rotation. CCW was defined as clockwise rotation.

In addition, the speed of movement of each rotation axis and the speed of movement in the hz-axis direction are determined to have a certain relationship, and the detection optical system that captures the true S-curve signal detects the position of the mask. The posture of the mask is controlled so that nothing escapes ε.

In Table A, when the mask 16 is raised, the detection optical system that first detects the true S-curve signal is 11.
After 2 hours, the mask posture control device 116 rotates the mask 16 around the X axis while stopping the rotation around the X axis.
While performing the CW rotation, the robot ascends in the Z direction. In this way, while the S curve signal is being detected by the detection optical system 112L, the other detection optical systems 112R and 112
Y also captures the true S-curve signal.

When a true S-curve signal is detected by the detection optical system 112R for the second time, at that point the mask posture control device 116 stops the rotation of the mask 16 about the X-axis and starts the CW rotation about the xihh center. While doing so, continue to ascend in the Z direction. As a result, the true S-curve signal is finally detected by the detection optical system 112Y as well.

The above search operation is carried out by the three detection optical systems 112L, 1
ε where 12R, 112Y captures all true S-curve signals
Then, the control unit 115 switches to servo operation using the straight line portion including the zero point of the true S-curve signal.

If the output voltages of the linear portions of the S-curve signals detected by the three detection optical systems 112L, 112R, and 112YC are VL, V, and VY, respectively, then the rotation around the y center is clockwise (CW) (■□- ■, 5), and rotate the X@ center clockwise (CW) by (1/2 (
The movement in the Z-axis direction is performed by an amount proportional to 1/3(vL+vll+VY)C. With the above operations, the setting of the mask 16 is completed.

When setting the wafer 17, the mask 16 is moved by a certain value in parallel to the direction C, plus two, and waits at a position where it will not come into contact with the wafer 17 even if it is tilted significantly. The setting of the wafer 17 is performed in exactly the same manner as the setting of the mask 16. Wafer 1
7 is the detection optical system 112L, 112FL, 112
When the mask 16 coincides with the plane including each image forming point of the light beam 21 from Y, the mask 16 is moved from the waiting position in parallel to the direction of the mass numeric Z and returned to the mask setting position. Finally, the postures of both the mask 16 and the wafer 17 are controlled by servo loop control using an S curve signal. Do it secretly,
The gap between the mask 16 and the wafer 17 is set to a predetermined value g.

This embodiment has been described above, but in short, in this embodiment,
As shown in FIG. 10, when aligning the reflective substrate (which may be the mask 16 or the wafer 17) in the optical axis direction with respect to the two light beams 20 and 21, the focus detection optical system 26M and the photoelectric conversion unit 27M The operation of detecting the mask 16 using the focus detection optical system 26W and the operation of detecting the wafer 17 using the photoelectric conversion unit 27W can be performed selectively or simultaneously.

When setting the mask 16 and wafer 17 under the condensing scanning lens 15 for the first time, the photoelectric conversion units 27M and 27W
are used selectively depending on the board to be set, and both boards receive true S-curve signals (101, 105).
After being supplemented within the range of
Both signals are used simultaneously to drive both boards to a more precise gear value g&:.

Although the embodiments described above are for gap control in a proximity exposure apparatus, the present invention can be similarly applied to a projection exposure apparatus that projects and exposes a pattern of a mask or reticle onto a wafer. In this case, the projection optical system corresponds to the imaging lens 15.

In addition, in the above embodiment, the slit width of the transmitting part and the light blocking part of the grating member was set equal to the minimum size of the spot image, but this means that the amplitude change of the detection signal due to the deviation of the spot image from the focused state is This is because it becomes the largest and the dynamic range becomes wider. However, if the spot image is smaller than the combined width of the transmitting part and the light shielding part of the grating member, that is, the width of one period of the grating, the focal position can be detected as an amplitude change.

Further, the grating member may be one in which transmitting parts (or light shielding parts) and reflective parts are arranged alternately in the form of slits.

Furthermore, although the spot light was a circular spot in the above-mentioned embodiment, it is not limited to a circular spot, and an elongated strip-shaped spot parallel to the stripes of the lattice member may be used).

In this case, a band-shaped spot can be easily obtained by arranging a cylindrical lens in the illumination optical system, for example.

[Effects of the Invention] As described above, according to the present invention, when aligning two substrates so that they face each other, a focus detection means for receiving and focusing the reflected light from the substrates via the imaging optical system is provided. , it is now possible to selectively switch between the operation of driving one substrate (mask) in the optical axis direction for alignment, and the operation of driving the other substrate (wafer) in the optical axis direction for alignment. did. Therefore, the 52 boards will not be set in incorrect positions due to the generation of false signals, and collisions between the two boards will be prevented.

Furthermore, the alignment detection means can also focus two substrates simultaneously via the imaging optical system, and when setting a strict distance between two substrates, quick alignment can be achieved by switching to simultaneous detection. .

[Brief explanation of drawings]

Figure 1 shows the conventional X
FIG. 2 is a block diagram showing the configuration of an embodiment of the present invention; FIG. 3 is a schematic optical path diagram showing a part of the previous figure under different conditions; Figure 4 is an optical path diagram showing a specific example of the configuration of the optical system in Figure 2;
Fig. 5(a) is a schematic optical path diagram showing an example of the focus detection optical system of the mask, Fig. 5(b) is a schematic plan view showing a partially enlarged grating member, and Fig. 6(a) (b). ) (c) is a line diagram showing the change in photoelectric output over time before and after the focus by the detector of the mask focus detection optical system, and Figure 7 is the position of the mask pattern surface in the optical axis direction with respect to the condensing scanning lens (horizontal axis) FIG. 8 is a block diagram showing a configuration example of the automatic focus detection signal processing system in this embodiment, and FIG. 9 is a diagram showing the relationship between the amplitude (vertical axis) of the photoelectric signal of the detector. An S-shaped characteristic diagram showing the relationship between the amplitude of the output of the signal processing system and the position of the mask pattern surface, FIG. 10(a) shows the convergence status of both the mask detection and wafer detection light beams in the vicinity of the mask surface. 10(b) and 10(c) are diagrams showing the magnitude of each focus detection signal of the wafer and mask with respect to the position in the optical axis direction, and FIG. FIG. 2 is a functional block diagram showing an embodiment of an apparatus for setting a gap between a mask and a wafer. [Explanation of symbols of main parts] 9: Laser light source, 10: Beam diameter converter, 11, Optical deflector, 12: Three-beam optical system, 13: Relay optical system, 14
: Beam splitter-15 = condensing scanning lens, 16: mask, 17: wafer, 20.21: luminous flux, 23 two reference oscillators, 24: drive circuit, 25: beam splitter-26
M: Mask focus detection optical system, 26W: Wafer focus detection optical system, 27M, 27W: Photoelectric conversion unit, 27MA. 27MB, 27WA, 27WB: detector, 28: signal processing section, 34 lens, 35 = beam splitter - 36,
37: Reflector, 42: Parallel flat glass, 43: Beam splitter - 44, 47: Lens, 48M, 48W: Beam splitter - 49MA, 49WA, 49WB: Grid member, 51MA, 51MB, 51WA, 51WB: Condenser lens , 53MA, 53MB, 53WA. 53WB: Image forming surface, 69a: Light transmitting part, 69b: Light non-transmitting part, 70 Nispot image, 84A, 84B: Brian amplifier, 85A, 85B + band pass filter, 86A, 86
B: Modulation component detection circuit, 87A, 87B + sample hold circuit, 89: Subtractor, 112L, 112R, 112
Y: detection optical system, 114: calculation unit, 115: control unit, 1
16: Mask attitude control section, 117: Wafer attitude control section.

Claims (2)

[Claims] (1) The first substrate is arranged so that a light-transmissive first substrate partially having reflective properties and a light-reflective second substrate disposed in parallel with the first substrate face each other at a predetermined interval. and a second substrate in a direction perpendicular to the substrate surface, the apparatus comprises: irradiating the second substrate with a luminous flux through the first substrate; an imaging optical system arranged to receive the reflected light; a first focusing plane of the imaging optical system based on the reflected light;
Focus detection means that outputs a detection signal when the substrate is located on either of the first focusing plane and a second focusing plane that is separated from the imaging optical system by the predetermined distance; a first driving means for relatively moving the first substrate and the imaging optical system in the optical axis direction; and a second driving means for moving the second substrate and the imaging optical system relatively in the optical axis direction. a drive means; a control means for selecting a state in which the first drive means is operated based on a detection signal of the focus detection means; and a state in which the second drive means is operated based on the detection signal; A substrate positioning device characterized by comprising: (2) The control means controls the first driving means so that the first substrate approaches the first focusing plane of the imaging optical system from one direction, and then the second substrate approaches the first focusing plane of the imaging optical system. The second part of the system
2. The apparatus according to claim 1, wherein the second driving means is controlled so as to approach the focal plane from one direction.