 |
Description  |
|
|
BACKGROUND OF THE INVENTION
1. Field of the Invention
The present invention relates to apparatus for projecting an image of a
pattern formed on a photo mask onto a substrate through a projection
optical system. More particularly the present invention relates to an
exposure apparatus for use in the optical lithography of semiconductor
devices such as IC and VLSI.
2. Description of the Prior Art
Apparatus for optically printing a circuit pattern described on a mask on a
wafer by exposure is known and widely used. As a type of the apparatus
there has recently been proposed a minifying projection type exposure
apparatus generally called stepper. This exposure apparatus is now coming
into wide use in the technical field of semiconductor devices.
The stepper uses a projection lens through which an image of a pattern on a
mask can be projected on a wafer at a minification of 1/5 or 1/10. The
wafer is placed on a stage for two-dimensional movement and exposed to the
pattern image. After one exposure the stage is moved a determined distance
and then the next exposure is carried out. In this manner, exposure is
repeated many times on the wafer while moving the stage at the same and
constant pitch after every exposure.
The projection lens used in this type of stepper is required to be bright
and have a high resolving power. To satisfy the requirement the lens has a
large numerical aperture and a very short depth of focus. The depth of
focus is only in the order of several .mu.m although it is variable
depending on the type and construction of the lens. Therefore, in
projecting the pattern image from the mask onto the wafer it is essential
to precisely focus the image on the wafer. To attain the necessary
accurate focusing there has been proposed a focus detection apparatus for
detecting the position of the wafer surface. An example of such a focus
detection apparatus is disclosed in Japanese Patent Application laid open
No. 42,205/1981. In this prior art apparatus, an image-forming light beam
is obliquely projected on the wafer surface and detection is made for the
position at which the reflected beam from the wafer surface is received.
The position of the wafer in the direction along the optical axis of the
projection lens can be detected by the detection of the reflected
beam-receiving position.
To this end, the apparatus is preadjusted in such manner that when the
pattern image is correctly focused on the wafer, the reflected
beam-receiving position lies just at the origin of a coordinate.
Therefore, this type of known detection apparatus needs a very careful
adjustment at its manufacture. After the adjustment has exactly been made
once, man can focus the pattern image on the wafer simply by moving the
wafer up or down in the direction of the optical axis of the projection
lens until the beam-receiving point gets in the origin. Thus, man can
projects a pattern image always in focus on the wafer.
However, the atmosphere in which the projection optical apparatus is used
is variable and the apparatus is sensitive to the change of the
atmospheric conditions such as temperature and atmospheric pressure.
Although it is very small in amount, the position of focus of the
projection lens (the position of the focal plane) shifts in the optical
axis direction with the variation of temperature and pressure. This
phenomenon is known as focus shift. Due to the focus shift it is difficult
to keep the focused relation between the focal plane and the wafer
surface. In practice it has been impossible to attain the necessary high
accuracy of the alignment (focusing) merely by moving the wafer up or down
for alignment with the aid of the oblique incidence type focus detection
apparatus according to the prior art. This is an important drawback of the
prior art apparatus.
SUMMARY OF THE INVENTION
Accordingly it is the primary object of the present invention to overcome
the above-mentioned drawback of the prior art and to provide apparatus
which enables to project an image of a pattern on a photo mask or reticle
onto a substrate such as wafer at a high accuracy of focusing.
It is another object of the invention to provide a focusing apparatus for
the high precision alignment between photo mask and substrate which allows
to monitor a mark on the photo mask and a mark on the substrate always in
a focused state and which can keep the photo mask and the substrate in a
positional relation accurately focused relative to the projection optical
system.
To attain the objects, the apparatus according to the present invention is
provided with a projection optical system and an image-forming optical
system. The projection optical system is disposed to project an image of
an alignment mark formed on a photo mask or reticle onto a
light-reflective substrate and also to reverse-project onto the photo mask
the light image of the alignment mark projected on and reflected by the
substrate. The image-forming optical system forms an overlap image from
the reverse-projected reflected image and the alignment mark on the photo
mask.
In a preferred embodiment of the invention, the overlap image formed by the
image-forming optical system is converted into image signal by image
sensor means. The image signal is then used to detect the focusing state
of the projected mark image on the substrate.
In another preferred embodiment, the image-forming optical system includes
means for adjusting the optical system in such manner as to focus the mark
image on a determined image plane. The substrate is provided with a
fiducial mark thereon. An image of the fiducial mark is projected on the
determined image plane through the projection optical system and the
adjusted image-forming optical system. In order to accurately focus the
fiducial mark image on the determined image plane there is provided means
for positioning the substrate in the optical axis direction relative to
the projection optical system.
Other and further objects, features and advantages of the invention will
appear more fully from the following description taken in connection with
the accompanying drawings.
BRIEF DESCRIPTION OF THE DRAWINGS
FIG. 1 schematically shows a first embodiment of the optical exposure
apparatus according to the invention;
FIG. 2 schematically shows an indirect focus detecting apparatus;
FIG. 3 is a block diagram of the control system;
FIG. 4 is a block diagram of the processing circuit shown in FIG. 3;
FIG. 5 is a block diagram of a circuit for the calibration of plane
parallel glass;
FIG. 6 is a flow chart showing the steps of the alignment operation for the
wafer surface and the focal plane of the projection lens;
FIG. 7 is a flow chart showing the steps of the operation for the
calbration of the indirect focus detecting apparatus;
FIG. 8 is a timing chart of the circuit shown in FIG. 4;
FIG. 9 shows an example of the overlap image;
FIGS. 10A and 10B show the waveforms of image signal and its differentiated
signal respectively;
FIG. 11 is a characteristic curve showing the relation between contrast and
focus position;
FIG. 12 is a characteristic curve of the detection signal of the
phase-synchronizing detector circuit;
FIGS. 13A, B and C and FIG. 14 show waveforms of the image signals in the
second embodiment of the invention of which FIG. 13A is of the signal for
rear-focus, FIG. 13B for in-focus, FIG. 13C for front-focus and FIG. 14
for far out-of-focus;
FIG. 15 is a characteristic curve showing the relation between contrast and
focus position in the second embodiment;
FIG. 16 schematically shows a third embodiment of the optical exposure
apparatus according to the invention;
FIG. 17 is a circuit diagram of the control system thereof;
FIG. 18 is a plan view of a reticle;
FIG. 19 is a plan view of a fiducial mark plate;
FIGS. 20 and 21 are flow charts showing the sequence of the operation of
the third embodiment;
FIG. 22 is a view showing the image pickup surface for displaying the
reticle mark;
FIG. 23 shows the waveform of the image signal obtained from the image
pickup surface;
FIG. 24 shows the waveform of the image signal for out-of-focus;
FIG. 25 is a contrast characteristic curve obtained from the image signal;
FIG. 26 is a waveform of the image signal for in-focus;
FIG. 27 is a view showing the image pickup surface for displaying the
reticle mark and the fiducial mark; and
FIG. 28 is a waveform showing an example of the image signal obtainable
from the image pickup surface shown in FIG. 27.
DESCRIPTION OF PREFERRED EMBODIMENTS
Referring first to FIG. 1 showing a first embodiment of the optical
exposure apparatus according to the invention.
In the apparatus, a light source 1 emits rays of light which pass through a
bundle of optical fibers 2 and then enter an illumination condenser lens
3. The pencil of rays is shaped into a beam by the lens 3 and then
reflected by a half-mirror 4 toward an observation optical system (for
example a microscope objective lens) 5. After passing through the
observation optical system 5, the beam of light from the light source 1
illuminates a transparent reticle R having a mark M. The mark M is
light-absorptive and not light-transmissive. An image of the mark M is
projected on a wafer W through a projection lens 6. The projection lens 6
is a lens whose chromatic aberration has been corrected in regard to the
exposure wavelength. The term "exposure wavelength" as herein used means
the wavelength of the light of the light used in the process of the
lithography a circuit pattern on the reticle R onto the wafer W. The
above-mentioned mark M is provided on the patterned surface of the reticle
R. Therefore, when the pattern surface is positioned at a distance d.sub.1
from the pupil 6a of the projection lens 6 and the reticle is illuminated
by the light of the exposure wavelength, an image of the mark M is formed
on an image plane P1 lying at a distance d.sub.2 from the pupil 6a on the
exit side of the projection lens. In order to obviate the variation of the
image plane P1 caused by the chromatic aberration, that is, in order to
keep the conjugation between the reticle R and the image plane P1, the
wavelength of the light from the light source 1 is so predetermined as to
be the same as the exposure wavelength.
On a stool 7 there is a XY stage 9 which can be moved two-dimensionally by
an actuator 8. Mounted on the XY stage 9 is a Z stage 11 on which a
light-reflective wafer W is placed. The Z stage can be moved up and down
in the direction along the optical axis of the projection lens 6 by a
second actuator 10. The wafer W has a thin layer of photosensitive
material coated on the surface.
In the position shown in FIG. 1, the photosensitive surface of the wafer W
is a little distant from the image plane P1 in the direction away from the
projection lens 6. This positional state is hereinafter referred to as the
rear focus state or briefly as rear-focus. The light beam l.sub.1 for
forming a light image of the mark M enters the projection lens 6 and is
focused on the image plane P1 through the lens 6. After focused on the
image plane P1, the beam is reflected toward the projection lens 6 by the
surface of the coated photosensitive layer or by the surface of the wafer
itself. Hereinafter, the surface of the coated photosensitive layer and
the surface of the wafer itself will be referred to as the surface of the
wafer W collectively.
Transmitted again through the projection lens 6, the reflected light
l.sub.2 passes through the reticle R at its transparent part around the
mark M and enters the light-receiving surface of a television camera or a
one- or two-dimensional solid state image sensor 12 through the
observation optical system 5 and the half-mirror 4. The light-receiving
surface of the television camera 12 is disposed conjugated with the
underside surface, i.e., the patterned surface of the reticle R.
Therefore, the light image of the mark M observed through the optical
system 5 is focused always on the light-receiving surface of the
television camera 12 as indicated by a beam l.sub.3. Further, since the
patterned surface and the image plane P1 are conjugate with each other
relative to the projection lens 6, when the image of the mark M is
projected on the wafer W using the exposure wavelength of light, the image
plane P1 and the light-receiving surface of the television camera 12 are
also conjugate with each other. In the rear focus state, therefore, the
reflected beam l.sub.2 is focused on a plane P2 a little distant from the
light-receiving surface of the television camera 12 through the
observation optical system 5.
By moving the wafer W upwardly from the position shown in FIG. 1, it is
possible to obtain the state in which the surface of the wafer W and the
image plane P1 are get in coincidence with each other (the state will
hereinafter be referred to as the in-focus state). In this state, the beam
l.sub.1 and the reflected beam l.sub.2 take the same optical path between
reticle R and wafer W. Consequently, the reflected beam l.sub.2 is focused
just on the light-receiving surface of the television camera 12.
If the surface of the wafer W lies at a position nearer to the projection
lens 6 than the image plane P1 (this state will hereinafter be referred to
as the front focus state), then the optical path of the reflected beam
l.sub.2 will again deviated from the optical path of the beam l.sub.1.
Therefore, in such front focus state, the reflected light l.sub.2 is
focused on an imaginary plane P3 receded from the light-receiving surface
of the television camera 12.
The embodiment of exposure apparatus shown in FIG. 1 further comprises a
projector 13 and a light receiver 14. The projector projects a beam of
light l.sub.4 for forming an image of a pin hole or slit. The beam l.sub.4
is obliquely directed to the image plane P1 and reflected by the surface
of the wafer W as a reflected beam l.sub.5. The receiver 14 receives the
reflected beam l.sub.5 to detect the position of the wafer W in the
vertical direction, i.e., in the direction along the optical axis of the
projection lens 6. According to the position of the wafer W in the
vertical direction, the reflection point of the reflected beam by the
wafer surface varies. For example, in the state of in-focus, the beam is
reflected as the reflected beam l.sub.5 ' whereas in the state of
rear-focus it is reflected as the reflected beam l.sub.5. The
light-receiver 14 photo-electrically detects such difference in reflection
point. The projector 13 and the light receiver 14 constitute together a
focus detection system for detecting the state of focusing of the wafer
surface to the image plane P1. A concrete form of the focus detection
system is shown in FIG. 2.
Referring to FIG. 2 the projector 13 comprises a light source 20, a slit
plate 21, a collimator lens 22, a mirror 23, a plane parallel glass 24 and
a focusing lens 25.
The light source 20 emits such wavelength of light to which the
photosensitive material coated on the wafer W is not sensitive (for
example, infrared rays). The light from the light source 20 illuminates
the slit plate 21 which has a slit opening elongated in the direction
perpendicular to the plane of the FIG. 2 drawing. The light passed through
the slit opening is collimated by the lens 22. The collimated light is
reflected to the plane parallel glass 24 by the mirror 23. The plane
parallel glass 24 shifts the optical axis of the bundle of collimated
rays. The collimated rays exiting from the plane parallel glass 24 are
focused by the focusing lens 25 to form an image of the slit of the slit
plate 21 on the image plane P1 of the projection lens 6. The plane
parallel glass 24 has a rotation axis extending in the direction
perpendicular to the plane of the FIG. 2 drawing and can be rotated about
the rotation axis within a determined range of rotational angle by an
actuator 26. By this rotation of the plane parallel glass the position at
which the slit image is formed through the focusing lens 25 is shifted in
the direction normal to the image plane P1. In the position shown in FIG.
2, the surface of the wafer W and the image plane P1 are coincident with
each other, namely in the state of in-focus.
On the other hand, the light-receiver 14 comprises a lens 27, an
oscillating mirror 28, a slit plate 29 and a photo sensor element 30.
The reflected beam l.sub.5 mentioned above enters the lens 27 and then
impinges on the oscillating mirror 28 which reflects the beam to the slit
plate 27 while changing the direction of the reflection by the oscillation
of the mirror. The slit plate 29 is disposed at the position at which the
beam from the lens 27 is focused. The slit plate has a slit 29a elongated
in the direction perpendicular to the plane of the FIG. 2 drawing. The
photo sensor element 30 receives the light passed through the slit 29a and
generates a photoelectric signal. The oscillating mirror 28 has a rotation
axis extending in the direction perpendicular to the plane of the FIG. 2
drawing. An actuator 31 oscillates the mirror 28 about the rotation axis
in the manner of simple harmonic motion at a constant angle frequency and
with a constant amplitude.
With the arrangement of the focus detection system above shown in FIG. 2,
in the state of in-focus, the slit image of the slit plate 21 is once
focused on the surface of the wafer W and the slit image formed on the
wafer surface W is again focused on the slit plate 29 by the lens 27. The
slit image reciprocally moves on the slit plate 29 with the oscillation of
the mirror 28. At the time, the center of oscillation of the slit image on
the slit plate 29 is coincident with the slit 29a. Therefore, if the wafer
surface W is deviated from the image plane P1 of the projection lens 6,
the oscillation center of the slit image also deviates from the slit 29a
leftward or rightward on the plane of the drawing according to the
deviation of the wafer surface. The direction in which the oscillation
center is deviated from the slit 29a represents the state of focus,
front-focus or rear-focus.
FIG. 3 shows the control system used in the above embodiment.
From the television camera 12 as previously shown in FIG. 1, an image
signal is introduced into a signal processing circuit 41 through a camera
control unit (CCU) 40. The processing circuit detects the focusing state
of the overlap image on the light-receiving surface of the television
camera 12 from the input image signal. As previously described, the
overlap image is an image formed by overlapping the light image of the
mark M and the reflected image of the mark M from the wafer W one on
another. In accordance with the detected state of focusing, the processing
circuit 41 generates a detection signal Sa representing the amount by
which the Z stage 11 is to be moved. The detection signal Sa is applied to
a main control unit 42 which includes a microprocessor (hereinafter
referred to as the host MPU).
In accordance with the input detection signal Sa, the main control unit 42
generates driving signals Sb and Sc to bring the focusing state to the
best, that is to say, to increase the contrast of the overlap image up to
the highest level. The driving signal Sb is applied to the actuator 10 to
adjust the position of the wafer in the vertical direction and the driving
signal Sc is applied to the actuator 8 of the XY stage 9 to adjust the
wafer position two-dimensionally.
Applied to the actuator 31 of the oscillating mirror 28 is an AC signal of
a certain frequency from an oscillator 43. In accordance with the
frequency of the AC signal, the actuator 31 oscillates the mirror 28 in
single harmonic motion. The photoelectric signal from the photo sensor
element 30 is amplified by an amplifier 44. The amplified photoelectric
signal and the AC signal from the above oscillator 43 are introduced into
a phase synchronous detection circuit (PSD) 45 which carries out a
synchronous detection on the photoelectric signal using the AC signal as
the reference signal. The signal detected by the PSD45 is then sent to a
low-pass filter (LPF) 46 to remove the higher harmonic component from the
signal. The output signal from LPF46 is applied to the main control unit
42. This output signal, namely, detected signal Sd is a so-called S curve
signal. When the oscillation center of the slit image is just at the slit
29a, the signal is zero. When the oscillation center is shifted from the
slit 29a, for instance, leftward as viewed on the plane of the FIG. 2
drawing, the signal becomes positive in polarity. When the oscillation
center is shifted in the opposite direction, it becomes negative in
polarity.
The main control unit 42 also applies to the actuator 26 a driving signal
Se for determining the rotation angle of the plane parallel glass 24. By
the way it is to be understood that the driving signal Sb is generated in
accordance not only with the detection signal Sa but also with the
detected signal Sd. The selection of the signal Sa or Sd according to
which the driving signal Sb is to be generated, is suitably made by the
host MPU in the main control unit 42. In addition to the selection, the
host MPU carried out also operations necessary for the generations of the
driving signals Sb, Sc, Se and the controls of the output timings of these
driving signals thereby totally controlling the sequence of the apparatus
as a whole.
FIG. 4 shows an embodiment of the signal processing circuit 41.
In the circuit shown in FIG. 4, a high speed analog-digital converter (ADC)
50 receives the image signal from CCU 40 and digitally converts the level
of the image signal. A sampling pulse generator 51 receives a horizontal
synchronizing signal from CCU 40. In accordance with the input horizontal
synchronizing signal, the generator 51 generates, during one horizontal
scanning period of the television camera 12, a pulse signal SP composed of
pulses corresponding to the respective horizontal picture elements. In
response to each pulse of the pulse signal SP, ADC 50 carried out sampling
of the image signal and converts the sampled level into a digital value.
Digital data D1 are transmitted to a random access memory (RAM) 52 from
the ADC 50. The number of pulses of the pulse signal SP is counted by an
address counter 53 to generate an address signal AD for the memory RAM 52.
Therefore, the RAM 52 serially renews the addresses in response to the
pulse signal SP and then stores at the renewed address the data D1
transmitted at each sampling. A micro-computer CPU 54 which is controlled
by the host MPU of the main control unit 42, serially reads out the data
D1 stored at the respective addresses in RAM52 and detects the state of
rising and falling of the level of the image signal, i.e. the contrast of
the overlap image. Further, the CPU 54 generates a clear signal CL and a
start signal St. The clear signal CL is used to cancel the data stored at
the respective addresses of RAM52 and also to reset the content of the
address counter 53 to zero or another initial value. The start signal St
is used to start the operation of the signal processing circuit 41. This
signal St is applied to the sampling pulse generator 51 as well as the
address counter 53. The sampling pulse generator 51 generates the pulse
signal SP when it receives the horizontal synchronizing signal after the
input of the start signal St thereto. The address counter 53 generates the
address signal AD in response to the pulse signal SP so long as the
counter is receiving the start signal St. When no start signal St is being
applied to the counter 53, it generates the address signal AD through an
address bus ADB from the CPU54 to allow the CPU 54 to access to the memory
RAM 52. The address counter 53 generates also a read-in completion signal
CR to the CPU 54 when one scanning of the horizontal scanning line is
completed, that is, when the counted number of pulse signals SP reaches a
determined value. The signal CR represents the completion of reading-in of
the data D1 in the amount of one scanning line. For the sake of simple
explanation the following description will be made on the assumption that
the number of picture elements per horizontal scanning line of the
television camera 12 is 1024, the ADC used is a 8-bit converter
(therefore, the data D1 is 8 bits) and the RAM 52 used is a memory having
a capacity larger than 1K bite (1024.times.8 bits). Further, the CPU 54
contains therein a memory or memories for storing the results of
operations, various data such constants, programs, etc.
FIG. 5 shows the processing circuit for processing the above-mentioned
detected signal Sd. This processing circuit is provided within the main
control unit 42.
Referring to FIG. 5, the detected signal Sd is applied to a switch 60. The
function of the switch 60 is to carry out a change-over operation in
response to a change-over signal S1 from the host MPU in the main control
circuit 42. The switch 60 receives also from a differential circuit 61 a
difference signal corresponding to the existing deviation of the detected
signal Sd from zero level (ground potential). In accordance with the
change-over signal S1, the switch 60 applies either the detected signal Sd
or its difference signal to an analog-digital converter (ADC) 62. The ADC
62 converts the input signal, i.e, the signal Sd or its difference signal
into digital data D2 and sends the data D2 to the host MPU and also to a
latch circuit 63. The latch circuit 63 latches the data D2 and the
operation of the latch circuit is controlled by a signal S2 from the host
MPU. The data D2 emitted through the latch circuit 63 is the
above-mentioned driving signal Se. The switch 60 is changed-over by the
host MPU in such manner that for the angle calibration of the plane
parallel 24 the switch selects the difference signal from the differential
circuit 61 and for the detection of the vertical position of the wafer
(focus detection) it selects the detected signal Sd.
The manner of operation of the apparatus is as follows:
As shown in FIG. 9, the light image M' of the mark M of the reticle R is in
the form of an elongate slit with the long side edges Eg1 and Eg2
extending in the direction intersecting the horizontal scanning line SL of
the television camera 12 at right angles. It is not always necessary that
the edges Eg1, Eg2 and the scanning line SL intersect at right angles.
However, for convenience' sake to explanation, we assume that the edges
and the scanning line intersect at right angles as shown in FIG. 9.
Since the observation optical system 5 focuses the light image M' of the
mark M on the light-receiving surface of the television camera 12, under
the state of in-focus the edges Eg1, Eg2 can be image-picked up with very
good contrast whereas the reflected image of the mark M reverse-projected
through the wafer W is in a defocused state at the position of the
reticle. Consequently, on the light-receiving surface of the television
camera 12, the reflected image and the light image M' do not exactly
overlap each other. The reflected image appears as a low contrast image Md
blurred out around the edges Eg1 and Eg2. Since as previously noted, the
mark is not light-transmissive, the reflected image Md appearing around
the light image M' looks darkish and the light image M' itself is also
looks black. In contrast, other part than the light image M' and the
reflected image Md on the scanning line SL looks whitish because of the
illumination light reflected by the wafer W. This means that the degree of
the contrast around the edges Eg1, Eg2 of the light image M' can be used
as a measure of the focusing state. When the reflected image Md and the
light image M' are coincident with each other exactly in position, the
contrast becomes high, which indicates the state of in-focus. Thus, the
focusing state can be detected by detecting the contrast around the edges
Eg1, Eg2 of the light image M'.
FIG. 6 is a flow chart of the program for the alignment (focusing) between
the surface of the wafer W and the image plane P1 of the projection lens
6. The respective steps of the program will hereinafter be described with
reference to FIG. 6.
Step 1
At first the Z stage 11 is set to its initial position. In this embodiment,
the initial position is most lowered position of the wafer W as shown in
FIG. 1. The Z stage is driven in accordance with the driving signal Sb
from the main control unit.
Step 2
The host MPU sends a start instruction to CPU 54 which then emits a clear
signal CL in response to the instruction. Thereby the memory RAM 52 and
the address counter 53 are cleared. Now the address 0 of RAM 52 is
accessible. Further, the CPU 54 emits a start signal St. Thereafter, the
input of horizontal synchronizing signal SH is started. From the time
point of the input a pulse signal SP is generated. In response to the
pulses of the pulse signal, the image signals coming from the CCU 40 are
serially sampled and converted into digital data D1. The data D1 are
serially stored at the respective addresses of RAM 52 starting from the
address of 0. As seen from the histgram in FIG. 8, the data group stored
in RAM 52 describes a level curve gently ascending because the level
becomes the lowest at the portion of the light image M' and there exists a
blurred reflected image Md (low contrast) on the both sides of the light
image M'.
Steps 3 and 4
When the address counter 53 has counted all of 1024 pulses of the pulse
signal at the end of one horizontal scanning period, the address counter
emits a completion signal CR which is monitored by the CPU 54. At the
generation of the signal CR, the CPU 54 terminates the generation of the
start signal St. In this manner, the image signal generated during one
horizontal scanning period is divided into 1024 picture elements. The
respective illuminance levels of these picture elements are converted into
the corresponding digital values and the digital data are stored in the
RAM 52 at the corresponding addresses from 0 to 1024.
Step 5
Among the data group stored in RAM 52, the CPU selects a series of data
corresponding to the edge portions Eg1, Eg2 of the light image M' and
extracts the rising and falling portions of the time-series image signal.
More concretely, data stored at the addresses from 0 to 1024 of RAM 32 are
checked to detect the address Ado corresponding to the edge Eg1 and the
address Ad2 corresponding to the edge Eg2 thereby finding out the address
Ad1 at the middle between Ado and Ad2 as shown in FIG. 10A. The portion of
from address 0 to Ad1 is regarded as the falling section of the image
signal and the portion of from Ad1 to 1024 as the rising section.
Step 6
Regarding the falling section and the rising section the CPU
primary-differentiates the data series in the memory RAM 52. This primary
differentiation may be carried out at high speed, for example, by passing
the data series through a numerical filter. Concretely speaking, as shown
in FIG. 10B, the data of the differentiation waveform are stored at
addresses of RAM 52 other than the addresses 0 to 1024. Although the
envelope of the waveform is shown in FIG. 10B, it is to be understood that
in practice the data series is a series of data dispersed as the data D1
in FIG. 8.
Step 7
The CPU detects the peak value Pv and the bottom (lowest) value Bv on the
differentiated waveform of the falling section and calculates the absolute
value of the difference between the peak value and the bottom value, PB
=.vertline.Pv-Bv.vertline.. The same operation is carried out also on the
rising section. The value PB thus obtained is stored in the memory of the
CPU. The steeper the rising or falling of the image signal is, the larger
the PB value becomes. Therefore, when the image signal rises up and falls
down very gently as in the case shown in FIG. 10, the value of PB is
small.
Steps 8 and 9
The CPU checks the position of the Z state 11 as to whether it has been
elevated up to the limit end (upper limit). If not, then the CPU gives to
the host MPU of the main control unit 42 a data for rising the Z stage by
a certain unit amount, for example, 0.1 .mu.m. Thereby the actuator 10
operates to rise the Z stage by 0.1 .mu.m.
The above operation is repeated starting from the step 2 until the Z stage
reaches the upper limit. Thus, the detection of peak and bottom values is
carried after the elevation of the Z stage by 0.1 .mu.m every time. The
found value of PB as well as the elevated position of the Z stage are
stored in the memory every time.
Step 10
If it is detected at the step 8 that the Z stage has just reached the upper
limit, then there is obtained in the memory of the CPU such data as shown
in FIG. 11. In FIG. 11, the abscissa represents the positions of the Z
stage in the vertical direction at every elevation by 0.1 .mu.m and the
ordinate represents the PB values found at the respective positions of the
Z stage. From the data the CPU detects the position Zo at which the value
of PB is the maximum.
Step 11
After detecting the position Zo at the above step, the CPU sends to the
main control unit 42 a signal for positioning the Z stage at the position
Zo. This signal is the detection signal Sa. In response to the signal Sa,
the main control unit 42 generates a driving signal Sb to lower the stage
11 to the position Zo at which the alignment of the wafer surface with the
image plane P1 is attained. As will be understood from the above, the
position Zo is the position at which the rising and falling of the image
signal is the steepest which means that the light image M' and the
reflected image Md get in coincidence with each other exactly (in the
state of in-focus).
In the manner described above, the alignment of the wafer surface W with
the image plane P1 is attained. In this position, the projected image of
the mark M can be focused the best on the wafer surface. After completing
the adjustment of the position of the Z stage in the above manner, the
operation for the calbration of the focus detection system comprising the
projector 13 and the light receiver 14 is carried out in accordance with
the flow chart shown in FIG. 7. The operation is as follows:
The host MPU of the main control unit 42 emits a changeover signal S1 to
make the switch 60 select the difference signal generated from the
differential circuit 61. Subsequent to it the host MPU emits a signal S2
to cancel the latching operation of the latch circuit 63. Thereby, in
accordance with the difference signal the plane parallel 24 rotates in a
manner of feedback control. When the difference signal becomes zero, the
plane parallel 24 stops rotating. FIG. 12 illustrates this phase of
operation. In FIG. 12, the abscissa represent the positions of the wafer
in the vertical direction (Z-direction) and the ordinate the magnitude
(level) of the detected signal Sd. Assuming that it is immediately after
the releasing of latch now and the detected signal Sd has a characteristic
curve as represented by SD1, the level of the detected signal Sd is -Sdo
at the time. However, as previously described, the wafer surface W and the
image point P1 must really be in the in-focus state at the position Zo. In
other words, in the case of characteristic curve SD1, the focus detection
is wrongly made as if the position of in-focus were at such a position of
the wafer W further lowered from the position Zo by the distance -Z1. In
order to correct it, in this case, the differential circuit 61 generates a
difference signal corresponding to the level -Sdo to change the angle of
the plane parallel 24 accordingly. Thereby, the level -Sdo is reduced and
the characteristic curve SD1 is corrected to the characteristic curve SD2.
According to the corrected characteristic curve SD2, as seen in FIG. 12,
the level becomes correctly zero at the position Zo.
The host MPU reads the data D2 from ADC 62 and detects the time point at
which the data D2 has just become zero, that is to say, the time point at
which the difference signal of the differential circuit 61 has just become
zero. However, there may be such case where the noise component of the PSD
45 can not completely filtered off by the LPF 46. In such case, the data
D2 can not always become zero perfectly. Taking into account such case,
the host MPU detects in practice it whether the data D2 has been within a
certain range of levels including zero. When the data D2 is zero or
approximately zero, it means that the characteristic curve SD1 is
completely corrected to SD2.
Subsequently to the above step, the host MPU emits a signal S2 to latch the
data D2 by the latch circuit 63 which continues to latch the data until
the next input of signal S2. During the time, the actuator 26 of the plane
parallel 24 continues to maintain the set angle of the plane parallel
using, as the reference value, the driving signal Se based on the latched
data. Then, the host MPU generates the changeover signal S1 to select the
detected signal Sd.
In this manner, the calibration of the focus detection system shown in FIG.
2 is completed. Now, the exposure of the reticle pattern on the wafer by
projection can be started. At the start of exposure, the host MPU reads
the data D2 of the detected signal Sd and generates a driving signal Sb
for reducing it to zero thereby positioning the wafer at the correct
position for in-focus in the vertical direction.
While the first embodiment has been described and shown to be provided with
an oblique incidence type focus detection system, it is naturally possible
to use other type of detection system, for example, an air micrometer type
focus detection system. This system detects the position of wafer in
Z-direction by blowing air toward the wafer through a small nozzle orifice
and detecting the back pressure. Even using such type of detection system,
the same effect of the invention as the first embodiment has can be
obtained without any problem.
A second embodiment of the invention will be described hereinafter with
reference to FIGS. 13, 14 and 15.
In the second embodiment, there is used, as the mark M on a reticle R, a
lattice pattern comprising plural lines arranged parallel at regular
intervals. Therefore, the mark has the so-called line-and-space forming
periodical structure.
In FIGS. 13A, 13B and 13C a reticle provided with such a lattice mark is
shown in a sectional view taken along the direction of scanning line,
together with the waveform of an image signal Sv and the light intensity
distribution curve Cw on the wafer corresponding to the scanning line from
which the image signal Sv was obtained. Of three figures FIG. 13A is for
the state of rear focus | | |
