 |
Description  |
|
|
BACKGROUND OF THE INVENTION
1. Field of the Invention
The present invention relates to a projection exposure apparatus used for
forming fine patterns in semiconductor integrated circuits, liquid crystal
displays, etc.
2. Description of the Related Art
A projection optical system used in a projection exposure apparatus of the
type described above is incorporated in the apparatus after high-level
optical designing, careful selection of a vitreous material, superfine
processing of the vitreous material, and precise assembly adjustment. The
present semiconductor manufacturing process mainly uses a stepper in which
a reticle (mask) is irradiated with the i-line (wavelength: 365 nm) of a
mercury-vapor lamp as illuminating light, and light passing through a
circuit pattern on the reticle is focused on a photosensitive substrate
(e.g., a wafer) through a projection optical system, thereby forming an
image of the circuit pattern on the substrate. Recently, an excimer
stepper that employs an excimer laser (KrF laser of wavelength 248 nm) as
an illuminating light source has also been used for evaluation or research
purposes.
Generally speaking, in order to faithfully transfer a fine reticle pattern
to a photosensitive substrate by exposure using a projection optical
system, the resolution and focus depth-of-field (DOF) of the projection
optical system are important factors. Among projection optical systems
which are presently put to practical use, those which are designed for the
i-line include a projection optical system having a numerical aperture
(NA) of about 0.6. In general, for a given wavelength of illuminating
light, as the numerical aperture of the projection optical system is
increased, the resolution improves correspondingly. However, the focal
depth (DOF) decreases as the numerical aperture NA increases. The focal
depth is approximately given by DOF=.+-..lambda./(2.times.NA.sup.2), where
.lambda. is the wavelength of illuminating light.
Incidentally, the resolution is improved by increasing the image-side
numerical aperture NAw (cf. the object-side numerical aperture NAr) of the
projection optical system. Increasing the image-side numerical aperture
NAw is the same as increasing the pupil diameter, i.e., increasing the
effective diameter of an optical element, e.g., lens, which constitutes
the projection optical system. However, the focal depth DOF decreases in
inverse proportion to the square of the numerical aperture NAw.
Accordingly, even if a projection optical system of high numerical
aperture can be produced, the required focal depth cannot be obtained;
this is a large problem in practical use.
Assuming that the wavelength of illuminating light is 365 nm of the i-line
and the numerical aperture NAw is 0.6, the focal depth DOF decreases to
about 1 .mu.m (.+-.0.5 .mu.m) in total range. Accordingly, a resolution
failure occurs in a portion where the surface unevenness or the curvature
is greater than DOF within one shot region (which is about 20 by 20 mm to
30 by 30 mm square) on the wafer.
In regard to these problems, super-high resolution techniques have been
proposed, for example, a phase shift method such as that disclosed in
Japanese Patent Application Post-Exam Publication No. Sho 62-50811, and a
SHRINC (Super High Resolution by Illumination Control) method disclosed,
for example, in WO92/03842, Japanese Patent Application Disclosure (KOKAI)
No. Hei 04-180612 and Japanese Patent Application Disclosure (KOKAI) No.
Hei 04-180613 (corresponding to U.S. Ser. No. 791,138 filed on Nov. 13,
1991). With these techniques, however, advantages such as an improvement
in the resolution and an increase in the focal depth can be effectively
obtained when a circuit pattern to be transferred is a periodic pattern
having a relatively high density. However, substantially no effect can be
obtained for discrete patterns (isolated patterns) such as those called
"contact hole patterns" in the present state of the art.
In order to enlarge the apparent focal depth for isolated patterns, e.g.,
contact holes, an exposure method has been proposed in, for example, U.S.
Pat. No. 4,869,999, in which exposure for one shot region on a wafer is
carried out in a plurality of successive exposure steps, and the wafer is
moved along the optical axis of the projection optical system by a
predetermined amount during the interval between each pair of successive
exposure steps. This exposure method is called FLEX (Focus Latitude
Enhancement Exposure) method and provides satisfactory focal depth
enlarging effect for isolated patterns, e.g., contact hole patterns.
However, the FLEX method indispensably requires multiple exposure of
contact hole images which are slightly defocusd. Therefore, a resist image
obtained after development inevitably lowers in sharpness (steepness of
the rise of the edge of the resist layer).
The Super-FLEX method published in Extended Abstracts (Spring Meeting,
1991) 29a-ZC-8, 9, The Japan Society of Applied Physics, is well-known as
an attempt in increasing the focal depth during projection of a contact
hole pattern without moving the wafer along the optical axis during the
exposure operation, as in the case of the FLEX method. In the Super-FLEX
method, a phase filter having a concentric amplitude transmittance
distribution centered at the optical axis is provided on the pupil plane
(i.e., a Fourier transform plane with respect to the reticle) of the
projection optical system so as to increase the effective resolution and
focal depth of the projection optical system by the action of the filter.
It should be noted that a method wherein the transmittance distribution or
phase difference is changed by filtering at the pupil plane of the
projection optical system to thereby improve the focal depth as in the
case of the Super FLEX method, is generally known as "multifocus filter
method". The multifocus filter is detailed in the paper entitled "Study of
Imaging Performance of Optical System and Method of Improving the Same",
pp.41-55, in Machine Testing Institute Report No. 40, issued on Jan. 23,
1961. The method of improving the image quality by spatial filtering at
the pupil plane is generally called "pupil filter method".
The assignee has proposed as a new type of pupil filter a filter of the
type that blocks light only in a circular region in the vicinity of the
optical axis (this filter will hereinafter be referred to as
"light-blocking pupil filter") in Japanese Patent Application Disclosure
(KOKAI) No. Hei 04-179958 (corresponding to U.S. Ser. No. 76,429 filed on
Jun. 13, 1993). The assignee has further proposed a pupil filter named
"SFINCS" that reduces the spatial coherence of a bundle of image-forming
rays from a contact hole pattern which passes through the pupil plane in
U.S. patent application Ser. No. 128,685 filed on Sep. 30, 1993.
Separately from the above-described pupil filters for contact hole
patterns, pupil filters which are effective for relatively dense periodic
patterns, e.g., line and space (L&S) patterns, have also been reported,
for example, in "Projection Exposure Method Using Oblique Incidence
Illumination I. Principle" (Matsuo et al.: 12a-ZF-7) in Extended Abstracts
(Autumn Meeting, 1991), The Japan Society of Applied Physics, and in
"Optimization of Annular Zone Illumination and Pupil Filter" (Yamanaka et
al.: 30p-NA-5) in Extended Abstracts (Spring Meeting, 1992), The Japan
Society of Applied Physics. These filters are adapted to lower the
transmittance (i.e., the transmitted light intensity) of a circular or
annular region centered at the optical axis (this type of filter will
hereinafter be referred to as "filter for L&S patterns"). In the L&S
pattern filter method, the phase of light passing through the filter is
not changed, unlike the Super FLEX method.
Among the foregoing various pupil filter methods, the Super FLEX method,
the light-blocking pupil filter method and the SFINCS method enable the
resolution and focal depth to be effectively increased with respect to
isolated contact hole patterns among fine patterns which are to be
transferred by exposure. However, for relatively dense patterns, e.g., L&S
patterns, these methods cause the resolution to be undersirably low.
Therefore, when such relatively dense patterns are to be exposed, it is
desirable to unload the pupil filter from the projection optical system or
to exchange it for a filter for L&S patterns.
However, the projection optical system is completed through a combination
of high-level designing and production and strict adjustment to obtain a
favorable projected image, as has been described above. Accordingly, if
the pupil filter, which optically changes characteristics of the
projection optical system, is merely loaded, unloaded or exchanged, the
image-forming characteristics of the projection optical system are
undesirably changed and cannot be favorably maintained.
In the case of an exposure apparatus designed on the premise that it will
be used only for specific patterns, e.g., contact hole patterns, the
projection optical system may be adjusted with a specific pupil filter
incorporated thereinto when the system is set up, as a matter of course.
However, in reality, in production lines for semiconductor devices or the
like, a single exposure apparatus is used for pattern transfer by exposure
at various steps in order to increase the production efficiency in the
present state of art.
SUMMARY OF THE INVENTION
Under these circumstances, an object of the present invention is to provide
a projection exposure apparatus capable of constantly maintaining
favorable image-forming characteristics even when a pupil filter suitable
for projection exposure of isolated patterns, e.g., contact holes, or a
pupil filter suitable for projection exposure of relatively dense
patterns, e.g., a filter for L&S patterns, is loaded, unloaded or
exchanged.
To attain the above-described object, the present invention provides a
projection exposure apparatus having an illuminating optical system for
irradiating a mask having a pattern with illuminating light for exposure,
and a projection optical system which is composed of a plurality of
optical elements and arranged to receive light emanating from the pattern
of the mask and to project an image of the pattern on a photosensitive
substrate with predetermined image-forming characteristics. The projection
exposure apparatus further includes an optical corrector plate inserting
device whereby an optical corrector plate that changes a specific factor
in the image-forming characteristics is removably inserted into a space
defined by a Fourier transform plane of the projection optical system and
a neighboring plane. In addition, an optical element moving device causes
at least one of the optical elements to move relative to the entire
projection optical system in accordance with the insertion of the optical
corrector plate.
In one embodiment of the present invention, the optical corrector plate
inserting device is an exchanging device that exchangeably inserts into
the above-described space one of a plurality of optical filters which are
different in optical action from one other.
One of the optical filters is a Super FLEX pupil filter that changes at
least either the transmittance or phase (amplitude transmittance) of
transmitted light at a part of the Fourier transform plane, or a filter
for L&S patterns.
One of the optical filters is a SFINCS pupil filter that reduces coherence
between light passing through a specific region of the Fourier transform
plane and light passing through the other region of the Fourier transform
plane.
One of the optical filters is a pupil filter that causes no change of the
transmittance, phase or coherence of light passing through the Fourier
transform plane, that is, a pupil filter which provides a state equivalent
to an ordinary state where no pupil filter is present (however, a simple
plane-parallel vitreous material may be inserted).
The projection exposure apparatus of the present invention is provided with
an optical element moving device that causes at least one of the
constituent elements of the projection optical system to move relative to
the entire projection optical system. Accordingly, a variation of the
image-forming characteristics which is caused by loading, unloading or
exchange of a pupil filter can be corrected by the action of the optical
element moving device. Thus, favorable image-forming characteristics can
be obtained at all times.
The above and other objects, features and advantages of the present
invention will become more apparent from the following description of the
preferred embodiments thereof, taken in conjunction with the accompanying
drawings, in which like reference symbols denote like elements.
BRIEF DESCRIPTION OF THE DRAWINGS
FIG. 1 is a schematic view showing the arrangement of one embodiment of the
projection exposure apparatus according to the present invention.
FIG. 2 is a plan view of pupil filters as shown in FIG. 1 and a pupil
filter exchanging mechanism.
FIG. 3 is a sectional view as seen in the direction of the arrow 3--3 in
FIG. 2.
FIG. 4 shows a modification of a movable retaining mechanism as shown in
FIG. 1.
FIG. 5 shows various pupil filters wherein sections (A), (B) and (C) show
examples of SFINCS pupil filters, and section (D) shows one example of a
light-blocking pupil filter formed from a metal plate.
FIG. 6 shows one example of a Fresnel lens-shaped optical corrector plate
which can be disposed on a pupil plane of a projection optical system in
the present invention, wherein section (A) is a plan view and section (B)
is a sectional view.
DESCRIPTION OF THE PREFERRED EMBODIMENTS
The present invention will be described below in detail with reference to
the accompanying drawings.
FIG. 1 shows one embodiment of the projection exposure apparatus according
to the present invention. Referring to the figure, a reticle R has a
pattern to be transferred drawn on a pattern surface defined on the lower
side (projection optical system side) thereof. The pattern is projected on
a substrate to be exposed, e.g., a wafer W, through a projection optical
system PL and transferred thereto by exposure. An illuminating light beam
ILB for illuminating the reticle R is emitted from a light source, e.g., a
mercury-vapor lamp 1. From the emitted light, only the i-line (wavelength:
0.365 .mu.m), for example, is selected through an elliptical mirror 2, an
input lens 4, a short-wavelength cut filter 5 and an interference filter
6. The illuminating light (i-line) emanating from the interference filter
6 then enters a fish-eye lens 7. It should be noted that a light source
used in the exposure apparatus of the present invention is not necessarily
limited to an emission line lamp such as a mercury-vapor lamp. For
example, a beam from a laser light source or the like may be made incident
on the fish-eye lens 7 after being collimated.
The exit-side surface of the fly-eye lens 7 forms a Fourier transform plane
in the illuminating optical system with respect to the reticle pattern,
where a surface illuminant image (i.e., a plane composed of the set of a
plurality of point light sources corresponding to the element lenses of
the fly-eye lens 7) is formed, and where a .sigma. stop 8 that defines the
shape and size of the surface illuminant image is also provided.
The illuminating light emanating from the fly-eye lens 7 and passing
through the .sigma. stop 8 illuminates the reticle R via mirrors 9 and 13,
lens systems 10 and 12 and a condenser lens 14. A variable illuminating
field stop (reticle blind) 11 is placed in conjugate relation to the
pattern surface of the reticle R by the action of the lens system 12 and
the condenser lens 14, thereby enabling the reticle illuminating range to
be varied. The lens system 10 is set so that the .sigma. stop 8 (surface
illuminant image) forms a Fourier transform plane with respect to the
reticle blind 11 or the pattern surface of the reticle R.
The bundle of rays transmitted and diffracted by the reticle R is focused
by the projection optical system PL to form a pattern image of the reticle
R on the wafer W. It should be noted that the broken line extending from
the reticle R to the wafer W in FIG. 1 shows the optical path of a bundle
of image-forming rays emanating from one hole pattern on the reticle R. In
this embodiment, the projection optical system PL is designed so that a
pupil plane FTP in the projection optical system PL, i.e., an optical
Fourier transform plane with respect to the reticle R, lies in a hollow
space, and an optical corrector plate is provided on the pupil plane FTP
or a plane neighboring to it. In this example, the optical corrector plate
is one pupil filter PF selected from among a plurality of optical filters,
which will be described later. Although the system shown in FIG. 1 employs
Koehler illumination in which the pupil plane FTP in the projection
optical system PL is conjugate to the position of the surface illuminant
image defined by the .sigma. stop 8 in terms of geometrical optics, it
should be noted that the illumination method is not necessarily limited to
the Koehler illumination.
The projection optical system PL is composed of a multiplicity of lens
systems. Some of the lens systems are retained by respective movable
retaining members 15, 18 and 21, which are joined to driving mechanisms
17, 20 and 23 through support members 16, 19 and 22, respectively, so that
these lens systems are movable relative to the entire projection optical
system PL. These optical element moving devices (i.e., the movable
retaining members, support members and driving mechanisms) are controlled
by an optical system controller 24. The movement of each movable lens
system by the associated optical element moving device is effected mainly
along the optical axis AX of the projection optical system PL. However,
there are cases where it is necessary to move a movable lens system in a
direction perpendicular to the optical axis AX or to rotate it about an
axis other than the optical axis AX (i.e., to tilt the lens system), as
described later. Therefore, it is desirable for each movable lens system
to be movable (rotatable) with multiple of degrees of freedom.
Loading, unloading or exchange of a pupil filter causes variation of
various optical aberrations. Among the aberrations, variations in
spherical aberration in particular can be effectively compensated for by
moving one or a plurality of lens elements retained by the movable
retaining members 18 and 21 in the vicinity of the pupil plane FTP.
Variations in distortion, astigmatism and field curvature can be
effectively compensated for by moving one or a plurality of lens elements
retained by the movable retaining member 15 in the vicinity of the reticle
R. Thus, compensation (correction) for general variations in aberration
can be satisfactorily made by moving each lens element only along the
optical axis.
The optical system controller 24 also effects control for the exchange of a
pupil filter PF. Since an optimum amount of movement (rotation) of each
movable lens element may vary with the pupil filter PF used, the optical
system controller 24 is set so that the amount of movement or rotation of
each movable lens element is optimized synchronously with the exchange of
the pupil filter PF. It should be noted that a command as to which pupil
filter should be used for exposure can be appropriately input to a main
control system 25 by the operator through a console (not shown) or the
like. However, since a type of appropriate pupil filter PF is solely
determined by the type of reticle pattern to be transferred, filters may
be automatically exchanged by reading the name, code or the like of a
reticle R to be used with a bar code reader 29, for example, and
determining a type of pupil filter PF to be used on the basis of the
recognized name or code. In this case also, each movable lens element is
moved to and set at an optimum position according to a selected pupil
filter PF under the control of the optical system controller 24, as a
matter of course.
Incidentally, the wafer W is retained on a holder of a wafer stage WST
which is adapted to move two dimensionally in an XY-plane perpendicular to
the optical axis AX. The position of the wafer stage WST is accurately
measured by a length measuring machine, e.g., a laser interferometer 27. A
wafer alignment sensor 28 detects the position of an alignment mark
(registration mark) formed on the wafer W or a positional error. A stage
controller 26 controls a motor for driving the wafer stage WST on the
basis of the value detected by the wafer alignment sensor 28 and the value
measured by the laser interferometer 27, thereby setting the wafer W to an
accurate exposure position.
The main control system 25 sends commands not only to the optical system
controller 24 but also to the stage controller 26, a shutter controller
31, a .sigma. stop and reticle blind controller 30, etc. to control the
opening and closing operation of a shutter 3 disposed in the vicinity of
the second focal point of the elliptical mirror 2 and to control the
aperture setting of the .sigma. stop 8 or the reticle blind 11.
FIG. 2 is a plan view showing pupil filters PF and a pupil filter
exchanging mechanism. Three different types of pupil filters PF1, PF2 and
PF3 are retained on a rotary plate 41 at regular spacings of 120.degree..
The rotary plate 41 is rotatable about an axis 40 of rotation. It is
assumed that, in the state illustrated in the figure, a Super FLEX pupil
filter PF1, which inverts the phase of light passing through a central
circular region PF1a with respect to the phase of light passing through an
annular region surrounding the circular region PF1a, has been loaded on
the pupil plane in the image-forming optical path of the projection
optical system PL as a pupil filter which is suitable for exposure of
contact hole patterns. The phase inversion may be effected either
continuously or stepwisely.
FIG. 3 is a sectional view of a pupil filter PF and the pupil filter
exchanging mechanism as seen in the direction of the arrow 3--3 in FIG. 2.
The rotary plate 41 is caused to rotate about the axis 40 of rotation by a
rotation driving unit 42, which is retained by a projection optical system
lens mount PL0 (FIG. 4). In addition, an encoder 43 is provided to
accurately measure the rotational position of the rotary plate 41, and it
is assumed that the rotary plate 41 has grating patterns (scale patterns)
cut in the peripheral edge thereof for indexing.
As one example, the Super FLEX pupil filter PF1 is formed from a
transparent flat plate of glass, quartz, etc. which has a transparent
dielectric film formed over its central circular region PF1a. Of the other
two pupil filters, the filter PF2 is arranged such that the transmittance
of a circular region PF2a defined in the center of the filter PF2 is lower
than the transmittance of the surrounding region. The pupil filter PF2 is
used for exposure of L&S (line and space) patterns. The pupil filter PF2
is formed from a transparent flat plate having a light-absorbing member,
e.g., a metal thin film, formed over its central circular region PF2a. It
is even more preferable that, when a pupil filter for L&S patterns is to
be used, the illumination of the reticle R by the illuminating optical
system 1 to 14 should be effected by the so-called annular zone
illumination method. Accordingly, the .sigma. stop 8 in FIG. 1 is
preferably adapted to be compatible with annular zone illumination. More
specifically, the surface illuminant image is partially obscured by the
.sigma. stop 8 so that it is formed into an annular zone shape. Therefore,
it is preferable to prepare a plurality of .sigma. stops 8 having
different aperture configurations and to arrange the system so that the
.sigma. stops 8 can be appropriately exchanged by an exchanging mechanism
similar to that shown in FIG. 2.
The other filter PF3 is formed from a uniform transparent flat plate (i.e.,
plain glass) which gives neither a transmittance difference nor a phase
difference over the entire surface thereof. That is, the filter PF3
provides a state equivalent to a state where no pupil filter is used. The
reason why such a filter PF3 is needed is that the other two pupil filters
PF1 and PF2 are transparent plates each having an optical thickness;
therefore, when neither of the pupil filters PF1 and PF2 is used, the
optical characteristics must be compensated for in a manner such that an
optical thickness equal to that of the pupil filters PF1 and PF2 is
ensured, that is, it is necessary to perform an operation of making the
optical path length uniform.
For the same reason, it is preferable that the optical thicknesses of a
plurality of pupil filters used in the present invention should be
approximately equal to each other. However, in this embodiment each
movable lens element in the projection optical system PL can be adjusted
to an optimum position in accordance with each pupil filter by the optical
element moving device so that variation of the image-forming
characteristics (i.e., increase in aberration) caused by exchange of pupil
filters is minimized. Therefore, the tolerance for variation in optical
thicknesses of a plurality of pupil filters can be markedly increased in
comparison to an arrangement having no movable lens elements.
Further, the optical element moving device makes it possible to compensate
for not only thickness variation among a plurality of pupil filters but
also unevenness of the thickness, particularly the taper component in each
pupil filter. Therefore, the tolerance for errors in production of the
pupil filters themselves can be relaxed considerably. This means that the
production cost of the pupil filters can be reduced to a considerable
extent.
Since the compensation for unevenness of the taper component cannot
satisfactorily be made by simply moving each movable lens element along
the optical axis as described above, it is preferable to structure a
specific movable lens element so that it is rotatable in a selected
direction (about a selected axis) together with the associated movable
retaining member (15, 18, 21, etc.). The rotation of a specific movable
lens element herein means extremely slight tilt, and the associated
rotating mechanism is only required to have a structure which enables
tilting of the movable retaining member.
The movable retaining members (15, 18, 21, etc.) may be retained directly
by the lens mount of the projection optical system PL through the support
members (16, 19, 22, etc.), as shown in FIG. 1. Alternatively, the movable
retaining members (15, 18, 21, etc.) may be retained, as shown in FIG. 4,
by a projection optical system lens mount PL0 through an intermediate lens
mount 66. In this case, the degree of freedom for movement of the movable
lens elements 50, 51, 52 and 53 further increases. FIG. 4 shows a
modification of each movable retaining mechanism in the arrangement shown
in FIG. 1. Two movable lens elements 50 and 51 are fixed to an inner lens
mount 60, and two movable lens elements 52 and 53 are fixed to another
inner lens mount 61. These two inner lens mounts 60 and 61 are spaced
apart from each other in the direction of the optical axis AX. A pair of
support members 62 and 63 retain the inner lens mount 60 with respect to
the intermediate lens mount 66, and another pair of support members 64 and
65 retain the inner lens mount 61 with respect to the intermediate lens
mount 66. The intermediate lens mount 66 is retained with respect to the
outer lens mount PL0 through support members 67, 68, 69 and 70.
With the above-described arrangement, when the four movable lens elements
50 to 53 are to be slightly moved together along the optical axis AX, the
intermediate lens mount 66 is moved vertically by actuating a driving
mechanism (e.g., a motor, air piston, piezoelectric element, etc.) joined
to each of the support members 67 to 70. When a pair of movable lens
elements 50 and 51 or 52 and 53 are to be moved singly, driving mechanisms
which are joined to the pair of support members 62 and 63 or 64 and 65,
which retain the respective inner lens mounts 60 and 61, should be
actuated independently. The inner lens mount 60 or 61, which has a pair of
movable lens elements fixed thereto, may be adapted to be tiltable.
Further, at least either one of the inner lens mounts 60 and 61 may be
adapted to be capable of moving slightly in a plane perpendicular to the
optical axis AX.
Incidentally, pupil filters applicable to the present invention are not
necessarily limited to the three types. It is also possible to use other
types, described above, of pupil filters, for example, a light-blocking
filter wherein a circular region in the vicinity of the optical axis is
shielded from light, as described above, and a SFINCS pupil filter that
reduces the spatial coherence of a bundle of rays passing through the
pupil plane, as described above. The SFINCS pupil filter enables an
improvement of the focal depth when an image of a contact hole pattern is
formed by reducing the coherence between light passing through a circular
region in the vicinity of the optical axis and light passing through an
annular region at the outer periphery of the pupil plane. The coherence
may be reduced by a method wherein an optical path length not less than
the coherence length (about 25 .mu.m in the case of the i-line having a
wavelength of 365 nm and a wavelength width .DELTA..lambda. of 5 nm) is
given between two bundles of rays which respectively pass through a
central circular region and an annular region surrounding it. In this
case, a transparent substrate in which the central circular portion and
the peripheral annular portion are different from each other in thickness
or refractive index is used as an actual pupil filter plate.
It is also possible to reduce the coherence by a method wherein light
passing through a region in the vicinity of the optical axis and light
passing through an outer peripheral region of the pupil plane differ in
polarization characteristic. An actual pupil filter plate used in this
case comprises a transparent substrate and polarizing plates, a halfwave
plate and quarter-wave plate disposed on the substrate so that the
directions of linear polarization in the central circular portion and the
peripheral annular portion intersect perpendicularly to each other.
In FIG. 5, section (A), (B) and (C) show examples of SFINCS pupil filters.
As shown in FIG. 5(A), a basic SFINCS pupil filter is formed from a
transparent circular substrate having a radius D.sub.0 slightly larger
than the effective pupil radius r.sub.0 in the projection optical system
PL, and has a central circular region PFc with a radius r.sub.1 (r.sub.1
<r.sub.0) and an annular region PFs.sub.1 with an inner radius r.sub.1 and
an outer radius r.sub.0. In order to prevent interference between a bundle
of image-forming rays passing through the circular region PFc and a bundle
of image-forming rays passing through the annular region PFs.sub.1, the
optical path length difference (thickness difference) between the circular
region PFc a | | |
