U.S. patent application number 10/032462 was filed with the patent office on 2002-05-09 for projection exposure apparatus.
This patent application is currently assigned to NIKON CORPORATION. Invention is credited to Endo, Kazumasa, Sasaya, Toshihiro, Ushida, Kazuo.
Application Number | 20020054282 10/032462 |
Document ID | / |
Family ID | 18157870 |
Filed Date | 2002-05-09 |
United States Patent
Application |
20020054282 |
Kind Code |
A1 |
Sasaya, Toshihiro ; et
al. |
May 9, 2002 |
Projection exposure apparatus
Abstract
A high-performance projection exposure apparatus excellent in
durability and reproducibility, which can adjust optical
characteristics which are rotationally asymmetric with respect to
the optical axis of projection optical system and which remain in
the projection optical system. The projection exposure apparatus
comprises an illumination optical system, a projection optical
system and an optical means. The illumination optical system
illuminate a first object, and the projection optical system
projects an image of the first object onto a second object under a
predetermined magnification. The optical means is set between the
first object and the second object, and has rotationally asymmetric
powers with respect to an optical axis of the projection optical
system. Consequently, the optical means can correct an optical
characteristic rotationally asymmetric with respect to the optical
axis of the projection optical system, remaining in the projection
optical system.
Inventors: |
Sasaya, Toshihiro;
(Yokohama-shi, JP) ; Endo, Kazumasa;
(Kawasaki-shi, JP) ; Ushida, Kazuo; (Setagaya-ku,
JP) |
Correspondence
Address: |
OLIFF & BERRIDGE, PLC
P.O. Box 19928
Alexandria
VA
22320
US
|
Assignee: |
NIKON CORPORATION
|
Family ID: |
18157870 |
Appl. No.: |
10/032462 |
Filed: |
January 2, 2002 |
Related U.S. Patent Documents
|
|
|
|
|
|
Application
Number |
Filing Date |
Patent Number |
|
|
10032462 |
Jan 2, 2002 |
|
|
|
09714945 |
Nov 20, 2000 |
|
|
|
09714945 |
Nov 20, 2000 |
|
|
|
08360515 |
Dec 21, 1994 |
|
|
|
Current U.S.
Class: |
355/53 |
Current CPC
Class: |
F21Y 2101/00 20130101;
F21W 2131/103 20130101; G02B 27/0025 20130101; F21V 5/04 20130101;
G02B 13/24 20130101; G03F 7/70308 20130101; G03F 7/70241 20130101;
G02B 3/06 20130101 |
Class at
Publication: |
355/53 |
International
Class: |
G03B 027/42 |
Foreign Application Data
Date |
Code |
Application Number |
Dec 22, 1993 |
JP |
323721/1993 |
Claims
What is claimed is:
1. A projection exposure apparatus comprising: an illumination
optical system for illuminating a first object; a projection
optical system for projecting an image of said first object
illuminated by said illumination optical system onto a second
object under a predetermined magnification; and an optical means
set between said first object and said second object, having
rotationally asymmetric powers with respect to an optical axis of
said projection optical system, for correcting an optical
characteristic rotationally asymmetric with respect to the optical
axis of said projection optical system, remaining in said
projection optical system.
2. A projection exposure apparatus according to claim 1, wherein
said optical means is arranged as rotatable about the optical axis
of said projection optical system.
3. A projection exposure apparatus according to claim 1, wherein
said optical means is arranged as movable along the optical axis of
said projection optical system.
4. A projection exposure apparatus according to claim 1, wherein
said optical means comprises a toric optical member having
different powers in orthogonal directions.
5. A projection exposure apparatus according to claim 4, wherein
said toric optical member comprises first and second toric optical
elements each having different powers in orthogonal directions,
wherein said first and second toric optical elements are arranged
as relatively rotatable about the optical axis of said projection
optical system.
6. A projection exposure apparatus according to claim 4, wherein
said toric optical member comprises first and second toric optical
elements each having different powers in orthogonal directions,
wherein said first and second toric optical elements are arranged
as relatively movable along the optical axis of said projection
optical system.
7. A projection exposure apparatus according to claim 1, wherein
said optical means is set either between said first object and said
projection optical system, inside said projection optical system,
or between said projection optical system and said second
object.
8. A projection exposure apparatus according to claim 1, wherein
said optical means is set at or near the pupil plane of said
projection optical system.
9. A projection exposure apparatus according to claim 1, wherein
said first object is a reticle a pattern to be projected is formed
thereon.
10. A projection exposure apparatus according to claim 1, wherein
said second object is a semiconductor wafer.
11. A projection exposure apparatus comprising: an illumination
optical system for illuminating a first object; and a projection
optical system for projecting an image of said first object
illuminated by said illumination optical system onto a second
object under a predetermined magnification; wherein said projection
optical system comprises a lens, the surface thereof contributes to
imaging performance of said projection optical system and has a
rotationally asymmetric region having rotationally asymmetric
powers with respect to the optical axis of said projection optical
system, in order to correct an optical characteristic rotationally
asymmetric with respect to the optical axis of said projection
optical system.
12. A projection exposure apparatus according to claim 11, wherein
said lens is arranged as rotatable about the optical axis of said
projection optical system.
13. A projection exposure apparatus according to claim 11, wherein
said lens is arranged as movable along the optical axis of said
projection optical system.
14. A projection exposure apparatus according to claim 11, wherein
said lens comprises first and second lens elements each having
different powers in orthogonal directions.
15. A projection exposure-apparatus according to, claim 11, wherein
said first object is a reticle a pattern to be projected is formed
thereon.
16. A projection exposure apparatus according to claim 11, wherein
said second object is a semiconductor wafer.
Description
BACKGROUND OF THE INVENTION
[0001] 1. Field of the Invention
[0002] The present-invention relates to a projection exposure
apparatus for illuminating a first object with light to
reduction-project a pattern on the first object thus illuminated
onto a substrate or the like as a second object:. More
particularly, the invention relates to a projection exposure
apparatus suitable for projecting a circuit pattern formed on a
reticle (mask) as the first object onto a substrate (wafer) as the
second object to effect exposure thereon.
[0003] 2. Related Background Art
[0004] As patterns for integrated circuits become finer, demands
for performance of the projection exposure apparatus used in
printing on wafer are becoming increasingly tougher these days.
[0005] Under such circumstances, a projection optical system is
required to have a higher resolving power, flatness of image plane,
less distortion, etc. Because of those, attempt has been made to
reduce the distortion by shortening an exposure wavelength A,
increasing the numerical aperture NA of projection optical system,
or decreasing the curvature of field. Some examples of such attempt
are those described in U.S. Pat. 5,260,832, Japanese Laid-open
Patent Application No. 5-173065, etc.
[0006] Also, Japanese Laid-open Patent Applications No. 59-144127
and No. 62-35628 describe methods for adjusting only a
magnification error. The former describes a technique that a curved
film, for example a pellicle, which is very thin and which does not
affect imaging performance, is set in an optical path. The latter
describes a technique that a rotationally symmetric plano-convex
lens or a combination of rotationally symmetric plano-convex and
plano-concave lenses is moved along the optical axis to
isotropically adjust the overall magnification on the wafer
surface.
[0007] The high-performance projection optical systems as disclosed
in U.S. Pat. 5,260,832 and Japanese Laid-open Patent Application
No. 5-173065, however, include a lot of constituent lenses, i.e.,
15 to 24 lenses. Particularly, in the case of high-resolution
projection optical systems with numerical aperture NA being at
least 0.4, the number of constituent lenses is very large, i.e., 20
or more. Thus, as the demand performance becomes higher, the
projection optical systems are further increasing the number of
constituent lenses and are becoming very complicated in structure.
Therefore, in order to actually produce these projection optical
systems, to mount them on projection exposure apparatus, keeping
aberrations such as the curvature of field, the astigmatism, the
distortion, etc. within ranges as designed, and then to withdraw
high performance the accuracy of individual lens components and the
accuracy of assembling must be controlled very strictly, which
would raise problems of poor yield, very long production period, or
failing to deliver sufficient performance, etc.
[0008] Further, the method for correcting the magnification error
as described in Japanese Laid-open Patent Application No. 59-144127
includes a step of curving a very thin film or the like not
affecting the imaging performance of optical system so as to
correct the magnification error by the prism effect thereof, but it
cannot make fine adjustment for a correction amount or a correction
direction of an asymmetric magnification error component with
directionality remaining in the projection optical system. In
addition, because it employs the thin film, the film can be
two-dimensionally held as bonded on a metal frame or the like for
long and narrow exposure areas as in the mirror projection method,
but it is very difficult for such a thin film to be three
dimensionally held and to reveal good reproducibility for
rectangular or square exposure areas. If glass etc. is used instead
of the thin film for holding the shape, it is also difficult to
form a thin and uniform film without affecting the imaging
performance. Further, there are serious problems, e.g., durability
of the film etc. including breakage accident due to heat absorption
or the like of exposure light in actual use of such films etc., a
change in optical performance with heat absorption of exposure
light or with an environmental change, etc.
[0009] Further, Japanese Laid-open Patent Application No. 62-35620
discloses the technique that the magnification error is adjusted
using a rotationally symmetric lens, but only moving the
rotationally symmetric lens along the optical axis can adjust only
the overall magnification on the wafer surface only on an isotropic
basis and cannot adjust the asymmetric magnification error
component with directionality remaining in the projection optical
system.
[0010] Moreover, the methods for correcting the magnification error
as disclosed in Japanese Laid-open Patent Applications No.
59-144127 and No. 62-35620 can basically correct only the
magnification error, but they cannot correct the astigmatism etc.
as off-axial aberrations. Further, it was also difficult for the
methods to handle rotationally asymmetric magnification error
components or distortion components locally remaining at random in
the projection optical system.
SUMMARY OF THE INVENTION
[0011] The present invention has been accomplished taking the above
problems into account. It is, therefore, an object of the present
invention to provide a high-performance projection exposure
apparatus excellent in durability and reproducibility, which can
adjust, without a very strict control of the accuracy of individual
components and the accuracy of assembling, optical characteristics
which are rotationally asymmetric with respect to the optical axis
of projection optical system and which remain in the projection
optical system, for example rotationally asymmetric off-axial
aberration components (astigmatism, curvature of field, etc.),
rotationally asymmetric magnification error components, etc.
Further, an auxiliary object of the invention is to provide a
projection exposure apparatus which can satisfactorily deal with
correction of rotationally asymmetric distortion etc. locally
remaining at random on a rotationally asymmetric basis in the
projection optical system.
[0012] The above object and other objects will be further apparent
from the following description.
[0013] Provided according to the present invention is a projection
exposure apparatus comprising an illumination optical system for
illuminating a first object, a projection optical system for
projecting an image of the first object illuminated by the
illumination optical system onto a second object under a
predetermined magnification, and an optical means set between the
first object and the second object, having rotationally asymmetric
powers with respect to an optical axis of the projection optical
system, for correcting an optical characteristic rotationally
asymmetric with respect to the optical axis of the projection
optical system, remaining in the projection optical system.
[0014] Also provided according to the present invention is a
projection exposure apparatus comprising an illumination optical
system for illuminating a first object, and a projection optical
system for projecting an image of the first object illuminated by
the illumination optical system onto a second object under a
predetermined magnification, wherein the projection optical system
has a lens, the surface thereof contributes to imaging performance
of the projection optical system and has a rotationally asymmetric
region having rotationally asymmetric powers with respect to the
optical axis of the projection optical system, in order to correct
an optical characteristic rotationally asymmetric with respect to
the optical axis of the projection optical system, remaining in the
projection optical system.
[0015] The present invention will become more fully understood from
the detailed description given hereinbelow and the accompanying
drawings which are given by way of illustration only, and thus are
not to be considered as limiting the present invention.
[0016] Further scope of applicability of the present invention will
become apparent from the detailed description given hereinafter.
However, it should be understood that the detailed description and
specific examples, while indicating preferred embodiments of the
invention, are given by way of illustration only, since various
changes and modifications within the spirit and scope of the
invention will become apparent to those skilled in the art from
this detailed description.
BRIEF DESCRIPTION OF THE DRAWINGS
[0017] FIG. 1 is a drawing to illustrate the principle when a toric
lens is a negative cylindrical lens;
[0018] FIG. 2 is a drawing to illustrate the principle when a toric
lens is a positive cylindrical lens;
[0019] FIG. 3 is a drawing to show the effect by the negative
cylindrical lens of FIG. 1;
[0020] FIG. 4 is a drawing to show the effect by the positive
cylindrical lens of FIG. 2;
[0021] FIG. 5 is a plan view to show a state of beam cross section
on a virtual plane shown in FIG. 3;
[0022] FIG. 6 is a plan view to show a state of beam cross section
on a virtual plane shown in FIG. 4;
[0023] FIG. 7 is a drawing to show a geometrical-optic relation of
the negative cylindrical lens shown in FIG. 3;
[0024] FIG. 8 is a drawing to show a geometrical-optic relation of
the positive cylindrical lens shown in FIG. 4;
[0025] FIG. 9 is a drawing to show a geometrical-optic relation of
a projection optical system;
[0026] FIG. 10 is a drawing to show a geometrical-optic relation
where a cylindrical lens as a toric lens is placed between the
projection optical system shown in FIG. 9 and a reticle;
[0027] FIG. 11 is a drawing to show a geometrical-optic relation
where a cylindrical lens as a toric lens is placed in the vicinity
of the pupil of the projection optical system shown in FIG. 9;
[0028] FIG. 12 is a drawing to show the overall structure of an
embodiment according to the present invention;
[0029] FIG. 13 is a plan view to show the structure of a reference
reticle;
[0030] FIG. 14A is a drawing to show a state where a positive
cylindrical lens and a negative cylindrical lens as toric lenses
are placed between the wafer and the projection lens;
[0031] FIG. 14B is a drawing to show a state where a positive
cylindrical lens and a negative cylindrical lens as toric lenses
are placed at or near the position of the pupil of the projection
lens;
[0032] FIG. 14C is a drawing to show a state where positive
cylindrical lenses as toric lenses are placed between the reticle
and the projection lens and between the projection lens and the
wafer, respectively;
[0033] FIG. 14D is a drawing to show a state where a pair of a
positive cylindrical lens and a negative cylindrical lens as toric
lenses are placed between the reticle and the projection lens and
another pair thereof between the projection lens and the wafer;
[0034] FIG. 14E is a drawing to show a state where a pair of a
positive cylindrical lens and a negative cylindrical lens as toric
lenses are placed between the reticle and the projection lens and
another pair thereof at or near the position of the pupil of the
projection lens;
[0035] FIG. 14F is a drawing to show a state where a pair of a
positive cylindrical lens and a negative cylindrical lens as toric
lenses are placed between the reticle and the projection lens,
another pair thereof at or near the position of the pupil of the
projection lens, and another pair thereof between the projection
lens and the wafer;
[0036] FIG. 15 is a structural drawing to show an embodiment in
which some constituent lenses of the projection lens are toric
optical members having a rotationally asymmetric lens power;
[0037] FIG. 16 is a lens constitutional drawing to show an
appearance of the projection lens shown in FIG. 15 when it is seen
along a direction parallel to the plane of FIG. 15; and
[0038] FIG. 17 is a drawing to show the overall structure of an
embodiment using the projection lens shown in FIG. 15 and FIG.
16.
DESCRIPTION OF THE PREFERRED EMBODIMENTS
[0039] As shown in FIG. 1, let fl be a focal length in the
meridional direction (the direction of yy' plane) of a cylindrical
lens 1 having a negative refracting power, which is a kind of toric
lens having different powers in orthogonal directions, d11 be a
distance from the cylindrical lens 1 to a reticle surface 4 (xy
plane) as a first object, and d12 be a position of an image point
(virtual image) formed between the object point (reticle surface 4)
and the cylindrical lens 1 by the cylindrical lens 1 where the
object point is at the center position of the reticle surface 4 (a
position of intersection between the reticle surface and the
optical axis Ax). In this case, the below formulas (1) and (2)
provide an image magnification p1 in the Y direction rotated by
.theta. from the y axis (or in the direction of a plane including
the optical axis Ax and the Y axis) by the cylindrical lens 1, and
the distance d12 between the cyclindrical lens 1 and the image
point position (hereinafter simply referred to as an image
position). Although not shown in FIG. 1, a projection optical
system for projecting a pattern on a reticle onto a wafer is
provided on the opposite side to the reticle surface 4 with respect
to the cylindrical lens, which is the case as to FIG. 2 to FIG. 4
as detailed later.
.beta.1=f1/(d11.multidot.cos.sup.2.theta.+-f1) (1)
d12d11.multidot.f1/(d11cos.sup.2.theta.+f1) (2)
[0040] Similarly, the following formulas (3) and (4) provide an
image magnification .beta.1' and an image position d12' in the X
direction (or in the direction of a plane including the optical
axis Ax and the X axis) perpendicular to the Y direction.
.beta.1=f1/(d11sin.sup.2.theta.+f1) (3)
d12'=d11.multidot.f1/(d11sin.sup.2.theta.+f1) (4)
[0041] Accordingly, an astigmatism amount AS1 is given by the
following formula (5).
AS1=d12-d12' (5)
[0042] It is thus understood that d11 in formula (1) to formula (4)
changes by moving the cylindrical lens 1, which changes the
astigmatism amount from formula (5) and which also changes the
magnifications of formula (1) and formula (3).
[0043] On the other hand, it is also understood that .theta. in
formula (1) to formula (4) changes by rotating the cylindrical lens
1, which changes the astigmatism amount from-formula (5) and which
also changes the magnifications of formula (1) and formula (3).
[0044] Further, as-shown in FIG. 2, let f2 be a focal length in the
meridional direction (or in the direction of yy' plane) of a
cylindrical lens 2 having a positive refracting power, which is a
kind of toric lens, d21 be a distance from the cylindrical lens 2
to the reticle surface 4 (xy plane) as a first object, and d22 be a
position of an image point formed by the cylindrical lens 2 where
the object point is at the center position of the reticle surface 4
(a position of intersection between the reticle surface and the
optical axis Ax). In this case, the below formulas (6) and (7)
provide an image magnification .beta.2 in the Y direction rotated
by .theta. from the y axis (or in the direction of a plane
including the optical axis Ax and the Y axis) by the cylindrical
lens, and the distance d22 between the cylindrical lens 2 and the
image point position (hereinafter simply referred to as an image
position).
.beta.2=f2/(d21.multidot.cos.sup.2.theta.+f2) (6)
d22=d21.multidot.f2/(d21.multidot.cos.sup.2 +f2) (7)
[0045] Similarly, the following formulas (8) and (9) provide an
image magnification .beta.2' and an image position d22' in the X
direction (or in the direction of a plane including the optical
axis Ax and the X axis) perpendicular to the Y direction.
.beta.2'=f2/(d21.multidot.sin.sup.2.theta.+f2) (8)
d22'=d21.multidot.f2/(d21.multidot.sin.sup.2.theta.+f2) (9)
[0046] Accordingly, an astigmatism amount AS2 is given by the
following formula.
AS2=d22-d22' (10)
[0047] It is thus understood that d21 in formula (6) to formula (9)
changes by moving the cylindrical lens 2, which changes the
astigmatism amount from formula (10) and which also changes the
magnifications of formula (6) and formula (8).
[0048] On the other hand, it is also understood that ' in formula
(6) to formula (9) changes by rotating the cylindrical lens 2,
which changes the astigmatism amount from formula (10) and which
also changes the magnifications of formula (6) and formula (8).
[0049] Now, AS1, AS2 expressed by the above formula (5) or formula
(10) is an astigmatism amount which can be corrected by either
cylindrical lens (1, 2).
[0050] In that case, a best focus plane is given by either one of
the following formulas.
(d12+d12')/2 (11)
(d22+d22')/2 (12)
[0051] Since the best focus plane changes depending upon d11, d21,
or .theta., it is clear that an amount of the curvature of field
also changes.
[0052] As described above, it is seen that amounts and directions
of the image magnification, the astigmatism, and the curvature of
field can be adjusted by moving a toric lens such, as a cylindrical
lens along the optical axis or rotating it. Another possible method
is to change the focal length of the toric lens itself, other than
the above method of adjustment.
[0053] Meanwhile, let .theta.=0 in order to estimate an amount of
correction of maximum astigmatism when the cylindrical lens 2 shown
in FIG. 2 is used. The maximum astigmatism in that case is given as
follows.
AS2.sub.max=-(d21).sup.2/(d21+f2) (13)
[0054] It was found that the maximum astigmatism amount AS2.sub.max
to be corrected was not more than 10.sup.-5L from results of
repetitive trial printings and studies with a projection exposure
apparatus for printing a line width of 10 or less microns where L
was a distance between a reticle as the first object and a wafer as
the second object.
[0055] Accordingly, supposing d21<10.sup.-2L,
.vertline.f2.vertline..gtoreq.10L (14)
[0056] from formula (13). The focal length of the positive
cylindrical lens 2 preferably satisfies the above condition of the
range in formula (14).
[0057] For a combination of two or more toric lenses such as the
cylindrical lenses as shown in FIG. 1 and FIG. 2 or for a
combination with another optical element, a new object point is
defined at an image position in a noted direction, of an image
formed by a first toric lens or another optical element from the
object point of reticle 4, a distance is recalculated between the
new object point and the next toric lens or another optical
element, and the distance is put in d11 or d21.
[0058] Next studied is a case wherein the negative cylindrical lens
1 shown in FIG. 1 and the positive cylindrical lens 2 shown in FIG.
2 are arranged in series along the optical axis.
[0059] Here, if directions of generatrices of the two cylindrical
lenses 1, 2 are coincident with each other and if a product of the
image magnifications of the two cylindrical lenses is 1, that is,
if .vertline..beta.1.multidot..beta.2.vertline.=1, combined powers
of the two cylindrical lenses 1, 2 in all directions become
approximately zero, thus not changing the optical characteristics
such as the magnification and off-axial aberrations (astigmatism,
curvature of field, etc.) at all.
[0060] On the other hand, if the directions of generatrices of the
two cylindrical lenses 1, 2 are perpendicular to each other, they
produces a maximum magnification and maximum off-axial
aberrations.
[0061] Accordingly, it is understood that adjustment can be
achieved by relatively rotating the two cylindrical lenses 1, 2,
for correction amounts or correction directions of asymmetric
magnification error components and off-axial aberration components
with directionality remaining in the projection optical system.
[0062] Where two negative cylindrical lenses 1 shown in FIG. 1 are
arranged in series along the optical axis or where two positive
cylindrical lenses 2 shown in FIG. 2 are arranged in series along
the optical axis, the maximum magnification and maximum off-axial
aberrations can be generated when the directions of generatrices of
the respective cylindrical lenses are coincident with each other;
whereas, they can have substantially the same lens effect as a
single, rotationally symmetric, spherical lens, when the directions
of generatrices of the respective cylindrical lenses are
perpendicular to each other.
[0063] As described above, amounts and directions of the optical
characteristics such as the magnification and off-axial aberrations
(astigmatism, curvature of field, etc.) can be arbitrarily adjusted
by using at least two cylindrical lenses each being a kind of toric
lens and arranging at least one of the cylindrical lenses so as to
be rotatable.
[0064] The above description mainly concerned the adjustment for
the astigmatism and the curvature of field, but the adjustment of
magnification error is next described in detail referring to FIG. 3
to FIG. 7 when the negative cylindrical lens 1 shown in FIG. 1 or
the positive cylindrical lens 2 shown in FIG. 2 is rotated about
the optical axis Ax.
[0065] FIG. 3 shows a state when a bundle of parallel rays in
radius R around the optical axis Ax are let to enter the negative
cylindrical lens 1 shown in FIG. 1. Here, in FIG. 3, a circle 13
represents a locus when the bundle of parallel rays in radius R
around the optical axis Ax pass the reticle surface 4 (xy plane),
while an ellipse 11 a locus when a beam diverged by the cylindrical
lens 1 from the bundle of parallel rays in radius R around the
optical axis Ax, passes the virtual plane (x'y' plane). Also, FIG.
5 shows a state of beam size on the virtual plane (x'y' plane)
shown in FIG. 3.
[0066] On the other hand, FIG. 4 shows a state when a bundle of
parallel rays in radius R around the optical axis Ax are let to
enter the positive cylindrical lens 2 shown in FIG. 2. Here, in
FIG. 4, a circle 13 represents a locus when the bundle of parallel
rays in radius R around the optical axis Ax pass the reticle
surface 4 (xy plane), while an ellipse 12 a locus when a beam
converged by the cylindrical lens 2 from the bundle of parallel
rays in radius R around the optical axis Ax, passes the virtual
plane (x'y' plane). Also, FIG. 6 shows a state of beam size on the
virtual plane (x'y' plane) shown in FIG. 4.
[0067] When the cylindrical lens 1, 2 is rotated about the optical
axis, the ellipse 11 in FIG. 3 or the ellipse 12 in FIG. 4 also
rotates with the rotation.
[0068] As shown in FIG. 7, letting .DELTA.R1 be a change amount of
the beam size in the y' direction (in the direction of a plane
including the optical axis Ax and the y' axis) being the meridional
direction on the virtual plane (x'y' plane) by the negative
cylindrical lens 1, and e1 be a distance between the negative
cylindrical lens 1 and the virtual plane (x'y' plane), the
following relation holds.
.DELTA.R1=-R.multidot.e1/f1 (15)
[0069] Similarly, as shown in FIG. 8, letting .DELTA.R2 be a change
amount of the beam size in the y' direction (in the direction of a
plane including the optical axis Ax and the y' axis) being the
meridional direction on the virtual plane (x'y' plane) by the
positive cylindrical lens 2, and e2 be a distance between the
positive cylindrical lens 2 and the virtual plane (x'y' plane), the
following relation holds.
.DELTA.R2=-R.multidot.e2/f2 (16)
[0070] As shown in FIG. 5 or FIG. 6, let R.sub.1, R.sub.2 be a
y'directional length (a half of the major axis in FIG. 5 or a half
of the minor axis in FIG. 6) represented by a solid line on the
virtual plane (x'y' plane). Then, they are given by the following
formulas.
R.sub.1=R(1-e1/f1) (17)
R.sub.2=R(1-e2/f2) (18)
[0071] Since the x'-directional length is R in either case, the
ellipse 11 and ellipse 12 indicated by the solid lines in FIG. 5
and in FIG. 6 can be expressed by the following formulas.
x'.sup.2/R.sup.2+y'.sup.2/[(1-e1/f1).multidot.R].sup.2=1 (19)
x'.sup.2/R.sup.2+y'.sup.2/[(1-e2/f2).multidot.R].sup.2=1 (20)
[0072] As described, where there is an asymmetric magnification
error for example as shown in FIG. 6 inside the projection optical
system, the beam size as shown in FIG. 6 can be arbitrarily changed
from ellipse to circle by rotating the cylindrical lens 1 of FIG. 3
having the optical characteristics as shown in FIG. 5, whereby the
asymmetric magnification error can be adjusted. Conversely, where
there is an asymmetric magnification error for example as shown in
FIG. 5 inside the projection optical system, the beam size as shown
in FIG. 5 can be arbitrarily changed from ellipse to circle by
rotating the cylindrical lens 2 of FIG. 4 having the optical
characteristics shown in FIG. 6, whereby the asymmetric
magnification error can be adjusted.
[0073] Here, when the negative cylindrical lens 1 as shown in FIG.
1 was used and when the distance is L between a reticle as the
first object and a wafer as the second object, results of
repetitive trial printings and studies with a projection exposure
apparatus for printing a line width of 10 or less microns showed
that a correction amount of maximum magnification error was
preferably not more than 10.sup.-4(=100 ppm).
[0074] Modifying the above formula (1) showing the relation between
the focal length f1 of cylindrical lens 1 and the magnification
.beta.1 of cylindrical lens 1, the following formula is
obtained.
f1=(-.sup.d11.multidot..beta.1)/(.beta.1-1) (21)
[0075] Converting the above correction amount of maximum
magnification error, 10.sup.-4 (=100 ppm), into .beta.1,
.beta.1=0.9999 (or 1.0001). Accordingly, supposing
d11.ltoreq.10.sup.-2L,
.vertline.f1.vertline..ltoreq.10.sup.2L (22)
[0076] from formula (21). Thus, it is preferred that the focal
length of the negative cylindrical lens 1 satisfy the condition of
the range of above formula (22).
[0077] The above description showed an example for correcting the
magnification error by rotating a toric lens (cylindrical lens)
about the optical axis, but it is apparent from the above formulas
(1), (3), (6), and (8) that the magnification error can also be
corrected by shifting a toric lens (cylindrical lens) along the
optical axis. In this case, it is more preferable that the above
formula (22) be satisfied.
[0078] Incidentally, the above description concerned that the
magnification error was able to be corrected using a toric lens
(cylindrical lens), but amounts and directions of the optical
characteristics such as the magnification error can also
arbitrarily be adjusted by using at least two cylindrical lenses
each being a kind of toric lens and arranging at least one of the
cylindrical lenses so as to be rotatable.
[0079] Thus, a possible arrangement is such that the negative
cylindrical lens 1 shown in FIG. 1 and the positive cylindrical
lens 2 shown in FIG. 2 are arranged in series along the optical
axis of the projection optical system and that they are arranged as
rotatable relative to each other. In this case, because the
negative cylindrical lens 1 has the optical characteristics as
shown in FIG. 5 and the positive cylindrical lens 1 has the optical
characteristics as shown in FIG. 6, it is understood that a beam
size formed by these cylindrical lenses (1, 2) becomes a
combination of the beam sizes shown in FIG. 5 and FIG. 6 and that
the combined beam size can be arbitrarily changed from ellipse to
circle by relatively rotating them, whereby the asymmetric
magnification error can be corrected.
[0080] Further, where-the projection optical system has the
asymmetric magnification error for example as shown in FIG. 5 or
FIG. 6, the beam size as shown in FIG. 5 or FIG. 6 can be
arbitrarily changed from ellipse to circle by arranging at least
two or more cylindrical lenses in series along the optical axis and
arranging at least one of the cylindrical lenses as rotatable,
thereby enabling to adjust the asymmetric magnification error.
[0081] Where two or more toric lenses (cylindrical lenses) are
combined or where a toric lens is combined with another optical
element, pursuit may be done under such an assumption that a light
beam obtained when a noted beam passes the first toric lens
(cylindrical lens) or another optical element, is considered as a
new beam and that the new beam enters the next toric lens
(cylindrical lens) etc.
[0082] In a combination of two toric lenses (cylindrical lenses),
where the negative cylindrical lens 1 as shown in FIG. 1 and the
positive cylindrical lens 2 as shown in FIG. 2 are set as close to
each other, the total lens power in each direction becomes nearly
zero, thus, the beam shape is not changed, when the directions of
generatrices of the lenses are coincident with each other; whereas,
a change of the shape becomes maximum, when the directions of
generatrices of the lenses are perpendicular to each other.
[0083] Further, where two negative cylindrical lenses 1 as shown in
FIG. 1 are arranged in series along the optical axis, or where two
positive cylindrical lenses 2 as shown in FIG. 2 are arranged in
series along the optical axis, a maximum magnification and maximum
off-axial aberrations can be generated when the directions of
generatrices of the cylindrical lenses are coincident with each
other; whereas, they can have the same lens effect as a single,
rotationally symmetric, spherical lens, when the directions of
generatrices of the cylindrical lenses are perpendicular to each
other.
[0084] As described, using at least two cylindrical lenses each
being a kind of tonic lens and arranging at least one of them as
rotatable, amounts and directions of the optical characteristics
such as the magnification and off-axial aberrations (astigmatism,
curvature of field, etc.) can be arbitrarily adjusted.
[0085] The above formula (14) and formula (22) can be expressed in
a general form as follows, where fA is a focal length of a
cylindrical lens effective for correction of astigmatism and fD a
focal length of a cylindrical lens effective for correction of
magnification error.
.vertline.fA.vertline..gtoreq.10L (23)
.vertline..sup.fD.vertline..gtoreq.10.sup.2L (24)
[0086] It is preferred that the above formula (23) be satisfied for
effectively correcting the astigmatism, and that the above formula
(24) be satisfied for effectively correcting the magnification
error. It should be, however, noted that the focal length (fA, fD)
of cylindrical lens in this case is not limited to a single
cylindrical lens, but may be applied to a combination of a
plurality of toric lenses such as cylindrical lenses, or toric
reflecting members. Namely, the focal length (fA, fD) of
cylindrical lens becomes a combined focal length of the plurality
of cylindrical lenses in the case of a combination of the plurality
of toric optical members.
[0087] Outside the relation of formula (23) or formula (24), a
toric component is too strong, which affects other aberrations
causing a problem. For example, in the case of correction of
astigmatism, the curvature of field or the magnification error is
degraded, or in the case of correction of magnification error, the
telecentricity or the astigmatism is degraded. Therefore, the
correction of asymmetric aberrations can be effectively made within
the above ranges.
[0088] Incidentally, the above formula (23) or formula (24) showed
the range of optimum focal length of toric optical member, and
further the range of optimum focal length of toric optical member
is next studied from another point of view.
[0089] First, FIG. 9 shows a projection optical system having a
front group GF on the reticle 4 side and a rear group GR on the
wafer 5 side with an aperture stop S in-between. Here, the front
group GF has a focal length of f.sub.GF and the rear group GR a
focal length of f.sub.GR. The projection optical system is
telecentric both on the reticle side and on the wafer side.
[0090] FIG. 10 shows a state where a cylindrical lens having a
positive power as a toric optical member is placed between the
front group GF in the projection optical system shown in FIG. 9,
and the reticle 4. The power of the cylindrical lens 2 is present
in the direction of the plane of FIG. 10 (or in the meridional
direction).
[0091] As shown in FIG. 10, letting f2 be a focal length of
cylindrical lens 2 and D.sub.1 be a distance between the
cylindrical lens 2 and the front group GF (a distance between the
principal points of the two optical systems), a combined focal
length F.sub.1 of the cylindrical lens 2 and the front group GF is
given by the following relation.
F.sub.1=(f2.multidot.f.sub.GF)/(f2+f.sub.GF-D.sub.1) (25)
[0092] Also, letting B.sub.1 be an image magnification of the
projection optical system (GF, GR) and B.sub.1' be an image
magnification of the combined system of the cylindrical lens 2 and
the projection optical system (GF, GR), the following relations
hold.
B.sub.1=-f.sub.GR/f.sub.GF (26)
B.sub.1=-f.sub.GR/F.sub.1=B.sub.1[1+(f.sub.GF.sup.-D.sub.1)/f2]
(27)
[0093] Accordingly, a magnification difference .DELTA.B.sub.1
between magnifications in the sagittal direction and in the
meridional direction of the projection optical system is given as
follows.
.DELTA.B.sub.1=B.sub.1'-B.sub.1=B.sub.1(f.sub.GF-D.sub.1)/f2
(28)
[0094] On the other hand, letting H.sub.1 be the reticle-side
principal point by the combined system of the cylindrical lens 2
and the front group GF, P.sub.1 be a reticle-side focus position by
the combined system of the cylindrical lens 2 and the front group
GF, .DELTA.s.sub.1, be a distance between the focus position
P.sub.1 and the reticle 4, and .DELTA.s.sub.1' be a distance
between the wafer 5 and a position Q.sub.1 of an image of reticle 4
by the combined system of the cylindrical lens 2 and the projection
optical system (GF, GR), the following relations hold.
.DELTA.s.sub.1=(f.sub.GF-D.sub.1).sup.2/(f2+f.sub.GF-D.sub.1)
(29)
.DELTA.s.sub.1'=(B.sub.1').sup.2.DELTA.s.sub.1 (30)
[0095] Here, .DELTA.s.sub.1' means a difference between image
positions in the sagittal direction and in the meridional direction
of the projection optical system, that is, an astigmatism amount
(astigmatic difference).
[0096] Also, letting NA.sub.R be a reticle-side numerical aperture
of the projection optional system and .lambda. be a wavelength of
exposure light, a depth of focus DOFR on the reticle side of the
projection optical system is as follows.
DOF.sub.R=.lambda./(NA.sub.R) (31)
[0097] Then, in order to control the astigmatism amount within the
reticle-side depth of focus, the following formula is derived from
the above formulas (29) and (31).
f2.gtoreq.--(f.sub.GFD.sub.1)+[(NA.sub.R).sup.2(f.sub.GF-D.sub.1).sup.2]/.-
lambda. (32)
[0098] Therefore, it is preferred that the cylindrical lens 2 be
constructed so as to satisfy formula (32), whereby the astigmatism
amount can be controlled within the depth of focus.
[0099] The following formula is a general expression of formula
(32), where .DELTA.f is a power difference in orthogonal directions
of the toric optical member.
.DELTA.f.ltoreq..vertline.-(f.sub.GF-D.sub.1)+[(NA.sub.R).sup.2
(f.sub.GF-D.sub.1).sup.2]/.lambda. (33)
[0100] It is thus understood that the above formula (33) should be
preferably satisfied in use of a toric optical member in order to
control the astigmatism amount by this member within the
reticle-side depth of focus of the projection optical system. It is
needless to mention that the above relations of formulas (32) and
(33) hold for any of projection optical systems having 1:1,
reduction, or enlargement magnification.
[0101] As an example, suppose the reticle-side numerical aperture
NAR of the projection optical system is 0.1, the wavelength
.lambda. of exposure light is 436 nm, f.sub.GF=250 mm, f.sub.GR=250
mm, and D.sub.1=200 mm. From the above formula (32), the focal
length f2 in the meridional direction, of the cylindrical lens
(generally speaking, from above formula (33), the power difference
.DELTA.f in orthogonal directions of toric optical member) is not
less than 5.7.times.10.sup.4 mm, and a magnification correction
amount (magnification difference .DELTA.B.sub.1) which can be
variable in this case becomes not more than 870 ppm
(=8.7.times.10.sup.-4).
[0102] In the above description, formula (33) was derived assuming
that the toric optical member was disposed between the reticle and
the projection optical system, but, because the same relation holds
even where the toric optical member is placed between the
projection optical system and the wafer, the following relation
should be preferably satisfied in that case.
.DELTA.f.gtoreq..vertline.-(f.sub.GR-D.sub.1')+[(NA.sub.W).sup.2(f.sub.GR--
D.sub.1').sup.2]/.lambda..vertline. (34)
[0103] Here, NA.sub.W is the wafer-side numerical aperture of the
projection optical system, and D.sub.1' is a distance between the
toric optical member and the rear group GR (a distance between the
principal points of the two optical systems).
[0104] Next studied referring to FIG. 11 is the range of optimum
focal length of cylindrical lens 2 where a positive cylindrical
lens 2 is placed between the front group GF and the rear group GR
in the projection optical system, in other words, in the vicinity
of the. aperture stop S.
[0105] FIG. 11 shows a state where a cylindrical lens 2 having a
positive power as a toric optical member is placed between the
front group GF and the rear group GR in the projection optical
system shown in FIG. 9. The power of the cylindrical lens 2 is
present in the direction of the plane of FIG. 11 (or in the
meridional direction).
[0106] Here, as shown in FIG. 11, letting f2 be a focal length of
the cylindrical lens 2 and D.sub.2 be a distance between the front
group GF and the cylindrical lens 2 (a distance between the
principal points of the two optical systems), the following
relation holds for a combined focal length F.sub.2 of the front
group GF and the cylindrical lens 2.
F.sub.2=(f2.multidot.f.sub.GF)/(f2.sub.+f.sub.GFD.sub.2) (35)
[0107] Also, letting B.sub.2 be an image magnification of the
projection optical system (GF, GR) and B.sub.2' be an image
magnification of the combined system of the cylindrical lens 2 and
the projection optical system (GF, GR), the following relations
hold.
B.sub.2=-f.sub.GR/f.sub.GF (36)
B.sub.2'=-f.sub.GR/F.sub.2=B.sub.2[1+(f.sub.GF-D.sub.2)/f2]
(37)
[0108] Accordingly, a magnification difference .DELTA.B.sub.2
between the magnifications in the sagittal direction and in the
meridional direction of the projection optical system is as
follows.
.DELTA.B.sub.2=B.sub.2'-B.sub.2=B.sub.2(f.sub.GF-D.sub.2)/f2
(38)
[0109] On the other hand, letting H.sub.2 be the reticle-side
principal point by the combined system of the front group GF and
the cylindrical lens 2, P.sub.2 be a reticle-side focus position by
the combined system of the front group GF and the cylindrical lens
2, .DELTA.s.sub.2 be a distance between the focus position P.sub.2
and the reticle 4, and .DELTA.s.sub.2' be a distance between the
wafer 5 and a position Q.sub.2 of an image of reticle 4 by the
combined system of the projection optical system (GF, GR) and the
cylindrical lens 2, the following relations hold.
.DELTA.s.sub.2=(f.sub.GF).sup.2/(f2+f.sub.GF-D.sub.2) (39)
.DELTA.s.sub.2'=(B.sub.2').sup.2.DELTA.s.sub.2 (40)
[0110] Here, .DELTA.s.sub.2' means a difference between image
positions in the sagittal direction and in the meridional direction
of the projection optical system, that is, an astigmatism amount
(astigmatic difference).
[0111] Then, in order to control the astigmatism amount within the
reticle-side depth of focus of the projection optical system, the
following formula is derived from the above formulas (31) and
(39).
f2.gtoreq.-(f.sub.GF-D.sub.2)+[(NA.sub.R).sup.2(f.sub.GF).sup.2]/.lambda.(-
41)
[0112] Accordingly, the cylindrical lens 2 is preferably
constructed so as to satisfy formula (41), whereby the astigmatism
amount can be controlled within the depth of focus.
[0113] The following formula presents a general expression of
formula (41) as a power difference .DELTA.f in orthogonal
directions of the toric optical member.
.DELTA.f.gtoreq..vertline.-(f.sub.GF-D.sub.2)+[(NA.sub.R).sup.2(f.sub.GF).-
sup.2]/.lambda..Arrow-up bold. (42)
[0114] It is thus understood that the above formula (42) should be
preferably satisfied in use of a toric optical member in order to
control the astigmatism amount by this member within the
reticle-side depth of focus of the projection optical system. It is
needless to mention that the above relations of formulas (41) and
(42) hold for any of projection optical systems having 1:1,
reduction, or enlargement magnification.
[0115] As an example, suppose the reticle-side numerical aperture
NA.sub.R of the projection optical system is 0.1, the wavelength
.lambda. of exposure light is 436 nm, f.sub.GF=250 mm, f.sub.GR=250
mm, and D.sub.2=200 mm. From the above formula (41), the focal
length f2 in the meridional direction, of the cylindrical lens
(generally speaking, from above formula (42), the power difference
.DELTA.f in orthogonal directions of toric optical member) is not
less than 1.43.times.10.sup.6 mm, and a magnification correction
amount (magnification difference .DELTA.B.sub.1) which can be
variable in this case becomes not more than 35 ppm
(=3.5.times.10-5).
[0116] From the results of the above analysis with FIG. 9 to FIG.
11, where the toric optical member is placed between the reticle
and the projection optical system or between the projection optical
system and the wafer, the contribution of correction to the
magnification error can be increased while suppressing the
contribution of correction to the astigmatism; while, where the
toric optical member is placed at or near the pupil of the
projection optical system, the contribution of correction to the
astigmatism can be increased while suppressing the contribution of
correction to the magnification error.
[0117] The toric optical member stated in the present invention may
be replaced by a toric lens having different powers in orthogonal
directions, obtained by polishing a rotationally symmetric
spherical surface more in one direction. A projection optical
system using such a toric lens will be described herein later.
Further, the toric optical member may be a reflecting mirror having
different powers in orthogonal directions, or a distributed index
lens having different powers in orthogonal directions.
[0118] Incidentally, the above description concerned the correction
of rotationally asymmetric aberrations such as the astigmatism, the
curvature of field, the magnification error, etc. using the toric
optical member having different powers in orthogonal directions as
an aspherical surface rotationally asymmetric with respect to the
optical axis of the projection optical system. If rotationally
asymmetric magnification error components or distortion components
locally remaining at random appear in the projection optical system
in addition to these aberrations and magnification error appearing
on a rotationally asymmetric basis, processing such as polishing is
locally applied to a lens surface of a cylindrical lens as a kind
of toric optical member slidable along the optical axis or
rotatable about the optical axis. Locating the thus processed
cylindrical lens between the reticle and the wafer, the
magnification error components and distortion components appearing
at random can be corrected in addition to the correction of the
astigmatism, the curvature of field, and the magnification error
appearing on a rotationally asymmetric basis.
[0119] Further, where the projection optical system has only the
magnification error components or distortion components locally
remaining at random on a rotationally asymmetric basis, the
magnification error components or distortion components appearing
at random can be corrected by applying local processing such as
polishing to an optical element (lens or reflecting mirror) itself
constituting the projection optical system.
[0120] Also, where the projection optical system has only the
magnification error components or distortion components locally
remaining at random on a rotationally asymmetric basis, the
magnification error components or distortion components appearing
at random can be corrected by such an arrangement that a
plane-parallel plate having a certain thickness is subjected to
local processing such as polishing and that the thus machined
plane-parallel plate is placed either between the reticle and the
projection optical system, inside the projection optical system, or
between the projection optical system and the wafer. In this case,
a spherical aberration appears because the plane-parallel plate has
the certain thickness. Then, the projection optical system can be
preliminarily arranged so as to correct the spherical
aberration.
[0121] An embodiment of the present invention is next described in
detail referring to FIG. 12.
[0122] FIG. 12 shows the structure of a projection exposure
apparatus according to the embodiment of the present invention. As
shown in FIG. 12, a reticle 35 held on a reticle stage not shown is
set above a both-side (or single-side) telecentric projection lens
36, and a toric optical member having different powers in
orthogonal directions is provided as optical means having
rotationally asymmetric powers with respect to the optical axis of
the projection lens 36 between the reticle 35 and the projection
lens 36. This toric optical member has, in order from the reticle
side, a negative cylindrical lens 1 having a concave surface facing
the projection lens and a negative power in the direction of the
plane of the drawing, and a positive cylindrical lens 2 having a
convex surface facing the reticle and a positive power in the
direction of the plane of the drawing, wherein the cylindrical lens
1 and cylindrical lens 2 each are arranged as rotatable about the
optical axis of the projection lens 36.
[0123] Further, a wafer 38 mounted on a wafer stage 37 is set at a
position conjugate with the reticle 35 with respect to the
projection lens 36, and the wafer stage 37 is composed of a
two-dimensionally movable XY stage and a Z stage movable along the
optical axis of the projection lens 36.
[0124] Above the reticle 35 there is an illumination optical system
21, 22, 23, 24, 25, 32, 33, 34 for uniformly illuminating the
reticle 35 by Kohler illumination, and the illumination optical
system includes a measurement system 42 for measuring the optical
characteristics of the projection lens, and a first alignment
system 47 for optically performing detection of relative position
between the reticle 35 and the wafer 38 with light of the same
wavelength as exposure light IL described below.
[0125] Also, an off-axis type second alignment system 48 is set
outside the projection lens 36, and the second alignment system 48
optically detects a position of the wafer 38 with light of a
wavelength different from the exposure light IL described
below.
[0126] Specifically describing the embodiment shown in FIG. 12, the
exposure light IL emitted from a light source 21 such as a mercury
lamp is collected by an ellipsoidal mirror 22, then is reflected by
a reflection mirror 23, thereafter is converted into a bundle of
nearly parallel rays by a collimator lens 24, and is incident into
an optical integrator 25 consisting of a fly's eye lens. A shutter
26 is provided near the second focus of the ellipsoidal mirror 22,
and the illumination light IL can be arbitrarily interrupted by
rotating the shutter 26 through a drive unit 27 such as a
motor.
[0127] When the exposure light IL is interrupted by the shutter 26,
the illumination light IL reflected by the shutter 26 is guided in
the direction approximately perpendicular to the optical axis of
the ellipsoidal mirror 22. The thus guided exposure light IL is put
into one end of a light guide 29 by a condenser lens 28.
Accordingly, the exposure light IL emitted from the light source 21
enters either the optical integrator 25 or the light guide 29.
[0128] When the exposure light IL is incident into the optical
integrator 25, there are a lot of secondary light source images
(hereinafter simply referred to as secondary light-sources) formed
on the reticle-side focal plane of the optical integrator 25. A
variable aperture stop 30 is set on the plane where the secondary
light sources are formed. The exposure light IL emitted from the
secondary light sources passes through a half mirror 31 inclined at
45.degree. relative to the optical axis, and thereafter travels via
a first condenser lens 32, a dichroic mirror 33, and a second
condenser lens 34 to illuminate a pattern area on the lower surface
of reticle 35 with uniform illuminance.
[0129] Upon exposure an image of the pattern on the reticle 35 is
formed on the wafer 38 through the toric optical member 1, 2 and
the projection lens 36. In this case, because the secondary light
source plane of the optical integrator 25 is conjugate with the
pupil plane of the projection lens 36, the a value indicating
coherency of the illumination optical system illuminating the
reticle 35 can be changed by adjusting an aperture of the variable
aperture stop 30 set on the secondary light source plane. When a
maximum incident angle of the exposure light IL illuminating the
reticle 35 is .theta..sub.IL and a half angular aperture of the
projection lens 36 on the reticle 35 side is .theta..sub.PL, the
.sigma. value can be expressed by
sin.theta..sub.IL/sin.theta..sub.PL. Here, the .sigma. value is set
in the range of about 0.3 to 0.7.
[0130] Although not shown, an aperture stop is provided at the
pupil position of the projection lens 36, and an aperture of the
aperture stop may be arranged as variable.
[0131] An adjustment plate 39 made for example of a glass plate is
fixed near a wafer holder of wafer stage 37, and a reference
pattern is formed on the surface on the projection lens 36 side, of
the adjustment plate 39. Corresponding to it, a reticle mark RM is
formed at a position conjugate with the reference pattern on the
adjustment plate 39 with respect to the projection lens 36, within
an image field of projection lens 36 and near the pattern area of
reticle 35. As an example, the reference pattern on the adjustment
plate 39 is a cross aperture pattern formed in a light shield
portion, while the reticle mark RM on the wafer 35 is a pattern
obtained by inverting the light and dark portions in a pattern
obtained by multiplying the reference pattern by a projection
magnification of the toric optical member 1, 2 and projection lens
36.
[0132] A condenser lens 41 and a reflective mirror 40 are set below
the adjustment plate 39 of wafer stage 37, and an exit end of the
light guide 29 is fixed at the rear focal plane of condenser lens
41. Since the surface of the exit end of the light guide 29 is
conjugate with the pupil plane of projection lens 36, it is also
conjugate with the variable aperture stop 30. Also, because-the
emission surface at the exit end of the light guide 29 is sized so
that the size of an image projected onto the variable aperture stop
30 is nearly equal to the aperture of variable aperture stop 30,
the reference pattern on the adjustment plate 39 is illuminated at
an illumination a value nearly equal to that for exposure light IL.
Further, in the illumination optical system of exposure light IL, a
light-receiving portion of photomultiplier 42 is set at a position
conjugate with the variable aperture stop 30 with respect to the
half mirror 31. Namely, the light-receiving portion of
photomultiplier 42 is arranged as conjugate with the pupil plane of
projection lens 36 and with the plane of the exit end of light
guide 29. A detection surface of the light-receiving portion is
sized larger than an image of the light-emitting surface of the
exit end of the light guide 29 projected thereon, thereby
preventing a light quantity loss. Therefore, when the reference
pattern on the adjustment plate 39 is illuminated from the bottom
side, most of light emerging from the reference pattern on the
adjustment plate 39 enters the projection lens 36 and toric optical
member 1, 2 no matter where the adjustment plate 39 is located in
the image field of projection lens 36, thus impinging on the
light-receiving surface of photomultiplier 42 through the reticle
mark RM on reticle 35.
[0133] A central processing unit 43 (hereinafter referred to as
CPU) is electrically connected to the photomultiplier 42, and
photoelectrically converted signals output from the photomultiplier
42 are supplied to CPU 43. A mirror for X direction and a mirror
for Y direction not shown are fixed on the upper surface of wafer
stage 37, and, therefore, coordinates of a position on the wafer
stage 37 can be always monitored using a laser interferometer 44
and the two mirrors. The coordinate information from the wafer
stage 37 is supplied through the laser interferometer 44 to CPU 43,
and the CPU 43 can move the wafer stage 37 to a desired coordinate
position through a stage drive unit 45.
[0134] The operation of the present embodiment is next described.
For measuring the rotationally asymmetric optical characteristics
(astigmatism, curvature of field, magnification error, distortion)
rotationally asymmetric with respect to the optical axis of the
projection optical system and remaining in the projection lens 36
and toric optical member 1, 2 because of assembling errors etc., a
reference reticle 35' as shown in FIG. 13 is preliminarily set on
the unrepresented reticle stage. As shown in FIG. 13,
light-shielding patterns of cross chromium or the like are arranged
at predetermined intervals on a two-dimensional basis in a pattern
area of the reference reticle 35'.
[0135] After intercepting the exposure light IL by the shutter 26
through the drive unit 27, CPU 43 moves the adjustment plate 39 on
the wafer stage 37 into the image field of projection lens 36
through the stage drive unit 45. By this, the exposure light IL
(hereinafter simply referred to as illumination light) reflected-by
the shutter 26 is led through the condenser lens 28 and the light
guide 29 into the wafer stage 37. After reflected by the reflective
mirror 40, the illumination light is converted into a bundle of
nearly parallel rays by the condenser lens 41 so as to illuminate
the reference pattern formed on the adjustment plate 39 from the
bottom side. The reference pattern on the adjustment plate 39 is
projected through the projection lens 36 and toric optical member
1, 2 onto the light-shielding patterns of reference reticle 35',
and the photomultiplier 42 photoelectrically detects a matching
condition between the two patterns--through the second condenser
lens 34,. the dichroic mirror 33, the first condenser lens 32, and
the half mirror 31. Then, in order to successively detect
coordinates of positions of plural light-shielding patterns
two-dimensionally arranged in the reference reticle 35' through the
photomultiplier 42, CPU 43 successively moves the wafer stage 37
through the wafer drive unit 45 while always monitoring the
coordinate position of the wafer stage 37 through the laser
interferometer 44. By this, the photomultiplier 42
photoelectrically detects respective matching conditions of the
reference pattern on the adjustment plate 39 with the plural
light-shielding patterns two-dimensionally arranged in the
reference reticle 35', and CPU 43 successively stores the
coordinate positions when matched, in a first memory unit not shown
in CPU 43 through the laser interferometer 44. Further, CPU 43 has
a second memory unit and a first correction amount calculating unit
inside, not shown, wherein the second memory unit preliminarily
stores correlational information about relations between the
optical characteristics (astigmatism, curvature of field,
magnification error, distortion) rotationally asymmetric with
respect to the optical axis of projection optical system and
relative rotation amounts of toric optical member 1, 2.
Accordingly, the first correction amount calculating unit
calculates an optimum amount of relative rotation for the toric
optical member 1, 2 to correct, based on information from the first
and second memory units. Then, based on the correction information
from the first correction amount calculating unit, CPU 43 outputs a
drive signal to the drive unit 46 such as a motor, so that the
drive unit 46 relatively rotates the toric optical member 1, 2 by
the determined correction amount (rotation amount).
[0136] After completion of the above operation, a normal reticle 35
used in actual process is set on the reticle stage not shown, and
CPU 43 changes over the shutter 26 through the drive unit 27. By
this, the exposure light IL illuminates the reticle 35 through the
illumination optical system, whereby an image of the pattern on the
reticle 35 is faithfully transferred through the toric optical
member 1, 2 and projection lens 36 onto the wafer 38. Continuous
exposure transfer with the projection exposure apparatus as
described could accumulate thermal energy due to the exposure light
IL in the projection lens 36, which would change the optical
characteristics of projection lens 36. Thus, during operations of
exposure transfer, the optical characteristics of projection lens
36 are periodically measured-as described above and the toric
optical member 1, 2 is rotated based on the measurement results. On
this occasion, it is more preferable that the above adjustment be
used in combination with the well-known technique to adjust the
magnification of the projection lens 36 itself by controlling the
pressure between constituent lenses of the projection lens 36.
[0137] It is to be desired that it is checked whether the
rotationally asymmetric optical characteristics (astigmatism,
curvature of field, magnification error, distortion) remaining in
the projection lens 36 are corrected in a perfectly optimized state
by an amount of relative rotation of the toric optical member 1, 2.
In this case, more perfect correction can be achieved by repeating
the above-described operations.
[0138] In measuring the magnification error and distortion
remaining in the projection lens 36, the wafer stage 37 is
two-dimensionally moved to obtain coordinate positions of the
respective light-shielding patterns in the reference reticle 35'.
In more accurately measuring the astigmatism and curvature of field
remaining in the projection lens 36, coordinate positions of the
respective light-shielding patterns in the reference reticle 35'
are obtained so as to maximize the contrast of an output signal
from the photomultiplier 42 while moving the wafer stage 37 along
the optical axis of the projection lens 36.
[0139] The projection exposure apparatus of the present embodiment
is fully effective for nonlinear extension or contraction of wafer
38 in the semiconductor fabrication process etc., or for cases
where semiconductors are fabricated by a plurality of projection
exposure apparatus and there are differences of magnification error
and distortion between the projection exposure apparatus.
Specifically, first, in order to successively optically detect
coordinate positions of plural wafer marks formed on the wafer
through the second alignment system 48 set outside the projection
lens 36, CPU 43 successively moves the wafer stage 37 through the
stage drive unit 45 while always monitoring the coordinate position
of wafer stage 37 through the laser interferometer 44. By this, CPU
43 successively stores the coordinate positions of respective wafer
marks formed on the wafer, as obtained from the second alignment
system 48 and laser interferometer 44, in a third memory unit not
shown inside CPU 43. Further, CPU 43 has a fourth memory unit and a
second correction amount calculating unit inside, not shown,
wherein the fourth memory unit preliminarily stores correlational
information about relations between the optical characteristics
(astigmatism, curvature of field, magnification error, distortion)
rotationally asymmetric with respect to the optical axis of
projection optical system and amounts of relative rotation of the
toric optical member 1, 2. Accordingly, the second correction
amount calculating unit calculates an optimum amount of relative
rotation for the toric optical member 1, 2 to correct, based on the
information from the third and fourth memory units. Then, based on
the correction information from the correction amount calculating
unit, CPU 43 outputs a drive signal to the drive unit 46 such as a
motor, so that the drive unit 46 relatively rotates the toric
optical member 1, 2 by the determined correction amount (rotation
amount).
[0140] Although the above embodiment shown in FIG. 12 showed an
example to correct the rotationally asymmetric optical
characteristics (astigmatism, curvature of field, magnification
error, distortion) remaining in the projection lens 36 by an amount
of relative rotation of the toric optical member 1, 2, it is
needless to mention that the correction may be made by relatively
moving the toric optical member 1, 2 along the optical axis of
projection lens 36. Also, the embodiment of FIG. 12 showed an
example to automatically correct the rotationally asymmetric
optical characteristics (astigmatism, curvature of field,
magnification error, distortion) remaining in the projection lens
36, but the rotation or movement of the toric optical member 1, 2
can be manually performed.
[0141] Further, the light source 21, ellipsoidal mirror 22, and
collimator lens 24 in the present embodiment may be replaced by a
laser light source such as an excimer laser etc. for supplying a
bundle of parallel rays. Moreover, this laser may be combined with
a beam expander for converting the laser light into a light beam
having a selected beam cross section.
[0142] The embodiment shown in FIG. 12 showed an example in which
the toric optical member 1, 2 is placed between the reticle and the
projection lens, but the present invention is by no means limited
to this arrangement. For example, arrangements as shown in FIGS.
14A to 14F may also be employed.
[0143] FIG. 14A shows an example in which the toric optical member
1, 2 is placed between the projection lens 36 and the wafer 38. As
shown, the toric optical member 1, 2 has, in order from the side of
wafer 38, a negative cylindrical lens 1 with a concave surface
facing the reticle 35 and a positive cylindrical lens 2 with a
convex surface facing the wafer 38. This arrangement can exert
greater contribution on correction of magnification error with
little affecting the astigmatism, as in the embodiment shown in
FIG. 12. Accordingly, this arrangement is effective (as the
embodiment shown in FIG. 12 is-similarly effective) to. cases where
a large magnification error remains in the projection lens 36.
[0144] FIG. 14B shows an example where the projection lens 36 is
composed of a front group 36A and a rear group 36B, and the toric
optical member 1, 2 is placed between the front group 36A and the
rear group 36B, i.e., at or near the pupil position of the
projection lens 36. As shown, the toric optical member 1, 2 has a
negative cylindrical lens 1 with a concave surface facing the wafer
38 and a positive cylindrical lens 2 with a convex surface facing
the reticle 35. This arrangement can exert greater contribution on
the correction of astigmatism with little affecting the
magnification error. Accordingly, this arrangement is effective to
cases where a large astigmatism remains in the projection lens
36.
[0145] FIG. 14C shows an example where the toric optical member 2A,
2B is separately arranged, one on the reticle 35 side and the other
on the wafer 38 side with the projection lens 36 in-between. As
shown, a first positive cylindrical lens 2A with a convex surface
facing the wafer 38 is set between the reticle 35 and the
projection lens 36 and a second positive cylindrical lens 2B with a
convex surface facing the reticle 35 is set between the projection
lens 36 and the wafer 38. Similarly as in the examples shown in
FIG. 12 and FIG. 14A, this arrangement can exert greater
contribution on the correction of magnification error with little
affecting the astigmatism.
[0146] FIG. 14D shows an example of application of FIG. 14C,
wherein negative cylindrical lenses 1A, 1B are combined with
associated positive cylindrical lenses 2A, 2B set on the reticle 35
side and on the wafer 38 side, respectively, with the projection
lens 36 in-between. This arrangement can exert greater contribution
on the correction of magnification error with little affecting the
astigmatism. In this case, out of the first toric optical member
1A, 2A and the second toric optical member 1B, 2B, one is mainly
used to correct the magnification error remaining in the projection
lens 36 while the other is used to correct the magnification error
due to expansion or contraction of wafer 38. Further, if, based on
this arrangement, the first toric optical member 1A, 2A and the
second toric optical member 1B, 2B are arranged so that one of them
has a strong power but the other a weak power, the one toric
optical member with strong power can be used to coarsely adjust the
magnification error with little affecting the astigmatism, while
the other toric optical member with weak power can be used to
finely adjust the magnification error with little affecting the
astigmatism.
[0147] FIG. 14E shows another example of application based on a
combination of FIG. 14A with FIG. 14B. As shown, a first toric
optical member 1A, 2A composed of a negative cylindrical lens 1A
and a positive cylindrical lens 2A is set between the reticle 35
and the projection lens (front group 36A), and a second toric
optical member 1B, 2B composed of a negative cylindrical lens 1B
and a positive cylindrical lens 2B is set between the front group
36A and the rear group 36B (at or near the pupil position of
projection lens 36) in the projection lens 36. According to this
arrangement, the first toric optical member 1A, 2A can adjust the
magnification error with little affecting the astigmatism, while
the second toric optical member 1B, 2B can adjust the astigmatism
with little affecting the magnification error. Namely, the
magnification error and the astigmatism can be corrected
independent of each other.
[0148] FIG. 14F shows an example of a combination of FIG. 14D with
FIG. 14E, which can correct the magnification error and the
astigmatism independently of each other and which can perform
coarse adjustment and fine adjustment of each of the magnification
error and the astigmatism.
[0149] In the above description, the cylindrical lenses 1, 2 etc.
as the toric optical-member are provided separately from the
projection lens 36, but some lenses constituting the projection
lens 36 may be arranged to to have a rotationally asymmetric power.
A projection lens 36 having such structure is next described.
[0150] In the projection lens 36 shown in FIG. 15, one surface of
each of lenses L1, L2 is processed into a toric surface to have a
curvature r.sub.am, r.sub.bm in the direction of the plane of FIG.
15.
[0151] FIG. 16 shows a state where the projection lens 36 shown in
FIG. 15 is observed in the direction of the plane of FIG. 15 (or in
the direction parallel to the plane of FIG. 15).
[0152] The lenses L1, L2 processed into a toric surface have
respective curvatures r.sub.as, r.sub.bs in the direction of the
plane of FIG. 16, as keeping the following relations.
[0153] r.sub.am.noteq.r.sub.as
[0154] r.sub.bm.noteq.r.sub.hs
[0155] The lenses L1, L2 processed into a toric surface are
rotatable about the optical axis Ax and rotatable by the drive unit
46.
[0156] If formula (33) is satisfied by a power difference .DELTA.f
in two mutually orthogonal directions of the toric surface in the
lens L1, L2, the rotationally asymmetric magnification error can be
corrected well while controlling the astigmatism amount within the
reticle-side depth of focus of the projection lens 36.
[0157] Similarly, one surface of lens L8, L9 in the projection lens
36 is processed into a toric surface, and the lenses L8, L9 are
rotatable about the optical axis Ax by the drive unit 46.
[0158] The toric surfaces have respective curvatures r.sub.cm,
r.sub.dm in the direction of the plane of FIG. 15 and respective
curvatures r.sub.cs, r.sub.ds in the direction of the plane of FIG.
16. Also, there are the following relations between the
curvatures.
[0159] r.sub.cm.noteq.r.sub.cs
[0160] r.sub.dm.noteq.r.sub.ds
[0161] If the power difference .DELTA.f in two mutually orthogonal
directions of the toric surface in each lens L8, L9 is selected so
as to decrease the rotationally asymmetric magnification error
given by formula (38), the astigmatism can be corrected well while
suppressing generation of the rotationally asymmetric magnification
error.
[0162] FIG. 17 shows the structure of an entire system using such a
projection lens 36. The drive unit 46 rotates either one pair out
of the pair of lenses L1, L2 and the pair of lenses L8, L9.
[0163] In the projection lens system 36 shown in FIG. 15 and FIG.
16, lenses other than the lenses L1, L2, L8, L9, that is, some of
lenses L3 to L7, L10 to L14 may be arranged as a toric optical
member.
[0164] From the invention thus described, it will be obvious that
the invention may be varied in many ways. Such variations are not
to be regarded as a departure from the spirit and scope of the
invention, and all such modifications as would be obvious to one
skilled in the art are intended to be included within the scope of
the following claims.
[0165] The basic Japanese Application No. 5-323721 filed on Dec.
22, 1993 is hereby incorporated by reference.
* * * * *