Note: Descriptions are shown in the official language in which they were submitted.
116 ~
This invention rela-tes to pattern -transfer and
particularl~y to a method and apparatus for providing a pattern on a
body or receiver by irradiating with light, x-rays, or other
radiation frorn a master or oriqinal of the pattern. The invention is
applicahle especiall~y, but not exclusively, to copying cameras and
projection printers.
A Problem encountered in transferring patterns is that
blemishes on the master are usually reproduced on the body. The
problem is especially acute when high definition, complex patterns
are being transferred, for example in making photomasks,
semiconductor devices such as integrated circuits and solar cells,
printed circuit hoards and like electrical devices. Masters for such
devices often have hlemishes caused by opaque defects in tile material
from ~hich the master is made, or by dirt or other contamination,
such as, particles of fiber or hair. In some cases, for example
where it brid~es a space between adjacent lines of a reticular
pattern, the hlemishes may render the clevice unusable.
An object of the present invention is to provide method
and apparatus for transferring patterns from a master to a body in
2() such a way that hlemishes in or on the master ~enerally are not
transferred.
According to one aspect of the present invention a
method of transferring a pattern onto a body, for example a
semiconductor device, photomask or printed circuit board, includes
the steps of exposing the body to radiation throucJh the interrnediary
of a master having two identical sets each of at least one pattern,
to produce on the body a latent imaqe of one of said sets, and
7~3
exposing the body a second time to superpose upon the latent image of
said one of the sets an ima~e of the other of said sets.
Since the blemishes are usually randomly distributed,
it is extremely unlikely that blemishes will occur at corresponding
positions in the different pattern sets. Consequently a latent image
of a blemish transferred to the body during the first exposure will
be removed during the second exposure. Conversely, a blemish exposed
during the second exposure will not produce a latent image because
the corresponding area of the body will have been exposed during the
first exposure.
Conveniently the required superposition of the second
latent image upon the first latent image is achieved by translating
the body and master one relative to the other between the first and
second exposures. Generally, such translation will be in the
direction of, and by a distance corresponding to, the spacing between
corresponding points on the two sets of patterns.
In one preferred embodiment of the invention
particularly suited to producing an array of superposed latent
irnages, both pattern sets are exposed onto the body in each exposure
2n step, and the body and master are translated one relative to the
other between such steps to produce at least one row of superposed
latent images. The first and last latent images in the or each row
will be of only one exposure, since each will have been exposed by
only one or other of the exposure steps. If desired, the
corresponding part(s) of the body can be discarded, perhaps after
processing. Alternatively the first and/or last latent image or
images may be arranged to fall beyond the edge of the body.
The relative movement hetween the body and master
hetween exposure steps may take place boustrophedonically to produce
an array of superposed latent images.
The master may comprise an array of patterns, the
different sets of patterns comprising identical patterns in different
lines of the array. The hody is again irradiated by way of the
master twice, the second time with the image of the array offset so
that latent irnages of identical patterns in one of said different
lines are superposed upon latent images of patterns in another of
said different lines. Where the array is generally rectilinear, each
line may comprise a row or column.
In each exposure or irradiation step the master pattern
set or sets may be transferred as a whole. Alternatively different
parts of the pattern set may be transferred in succession, until the
whole has been transferred. Such successive transfer may be effected
by scanning the body, either continuously or step-by-step. The body
and master mi~ht be aligne(l, before the first exposure, relative to
an alignment mark or marks, and realigned hetween first and
successive exposures at a position displaced from the first position
2() hy the distance required to give superposition.
Embodiments of the invention may employ size reduction,
typically by a factor of 5 or 10, in transferring the pattern to the
body. However, transfer may be effected without significant change
in size, for example by contact exposure, by proximity exposure, or
1:1 pro~jection exposure, especially where the master pattern set
itelf comprises an array of identical patterns. Further, some
enlargement may be provided.
'17~3
According to a second aspect of the invention,
apparatus for transferring a pattern to a body comprises a support
for a mas-ter having two identical sets each of at least one pattern,
a support for the body, means for irradiating the body by way of the
master to produce on the body a latent image of at least one of said
sets, and means operative between successive irradiation steps to
configure the apparatus such that a latent image of one of said
identical sets formed durinq one exposure step is superposed upon a
latent image of the other of said sets produced upon the body during
a preceding exposure step.
The means for configurinq the apparatus may comprise
translation means for moving said body and master one relative to the
other to such an extent as to provide the required superposition.
Generally, such movement will be in the direction of, and by a
distance corresponding to, the spacing between corresponding points
of the two identical pattern sets.
Thè translation means may be arranged to step the body
relative to the master, or vice versa, repeatedly, each step equal to
said distance to provide at least one row of superposed latent images
2n on the body.
The or each set of patterns may be part of an array of
patterns, conveniently arranged in rows and columns. The translation
means may then serve to translate the body and master one relative to
the other, by one row or column, or multiple thereof, between
successivè exposures.
The irradiation means may be arranged to transfer the
or each entire pattern set simultaneously. Alternatively, and
7~3
especially where the set comprises an array of identical patterns,
the irradiation means may serve to irradiate different areas of the
body in succession to transfer a complete imaqe of the or each set
onto the body. Such irradiation means may conveniently comprise
scanning means, for example a shutter member, controlling scanning of
said different areas, stepwise or in a continuous sweep.
More specifically, in a preferred embodirnent, the
apparatus includes an imaging device, for example a lens or mirror
system, between the master and the body, and a shutter between the
body and radiation source. The body and master are movable together
across the optical axis and relative to the imaginq device and
shutter, or vice versa, to scan and expose successively the different
areas of the body. The holder for the body comprises two parts, the
translating means servinq to move one such part relative to the
other, between successive scans, by a distance equivalent to the
spacing between the sets of patterns to be superposed.
Such implementation of the invention will now be
describer1 by way of example only and with reference to the
accompanying drawings, in which:-
Figure 1 illustrates parts of apparatus for irradiating
a body comprising a semiconductor wafer or photomask;
Figure 2 illustrates the formation of latent images on
the wafer or photomask during a first exposure;
Figure 3 illustrates the formation of latent images on
the wafer, or photomask, during a subsequent exposure; and
Figure 4 illustrates pro~jection apparatus for
transferring an array of images to a semiconductor wafer or
photomask.
7`~3
Fiqure 1 illustrates, schematically, parts of a
"step-and-repeat" camera for exposinq semiconductor wafers or
photomasks. To facilitate better understanding of the invention, the
detailed construction of the camera is omitted. The reader is
referred to IJ.S. patent specification No. 4,172,656, as an exemplary
disclosure of cameras of this kind and their use.
Referring to Figure 1, the apparatus comprises a table
10 which supports a semiconductor wafer 12. The uppermost surface of
the wafer 12 is coated with photoresist to be exposed to radiation
heneath an optical column part 14. The optical column 14 includes a
holder or platen 15 for a reticle 16, comprising a transparency
carrying the master pattern to bè reproduced upon the wafer 12. The
reticle 16 extends in a plane perpendicular to the optical axis 17 of
the column 14. A lamp 18 is disposed above the reticle holder 15 so
as to irradiate the reticle 16. An imaging device in the form of a
lens system 20 below the reticle 16 is arranged to form an image 16'
of the rnaster pattern (usually reduced by a factor of 5 or 10) in the
same plane as the photoresist-coated surface of the wafer 12. The
semiconductor wafer 12 and the reticle 16 are secured to the table 10
and holder 15, respectively, by vacuum chucks (not shown), or other
suitable devices.
The reticle lfi is provided with two identical sets of
patterns. In this case the sets comprise single patterns I and II,
respectively, their longitudinal center lines 19 and 21 spaced apart
by a distance D For convenience each pattern is shown as a sirnple
arrow rather than the complex pattern actually used for production of
an integrated circuit device and also is shown without any irnage
rotation.
l3
The reticle 16 may be produced bV any of the usual
processes, for example so-called rubylith cut-and-strip, hut
preferably is produced wholly or in part by a pattern generator.
Such pattern generators offer accuracy and positional precision of
less than 1 micron with orthogonality of about 1 are second and so
are capable of positioning the two master patterns relative to each
other very precisely.
In addition the reticle 16 has two fiducial marks, 22
and 24, near respective opposite edges of the reticle 16, the optical
axis 17, passing through the mid point between marks 22 and 24.
Correspondinq fiducial marks 26 and 28 are provided on a reference
surface 30 of the reticle holder 15 within the optical column 14.
The axis through rnarks 26 and 28 must he parallel to the transverse
movement of the table 10. To position the reticle 16 relative to the
optical axis 17 before use, the fiducial marks 22 and 24 are aligned
with the marks 26 and 28 using a micro-manipulator assembly with
screw ad,justers for lateral, longitudinal and rotational adjustments
while viewing the two sets of fiducial marks by way of a split field
microscope (not shown).
2n The table 10 is movable, perpendicularly to the optical
axis 17 of the optical column 14, between an exposure position,
wherein the semiconductor wafer 12 is beneath the optical column 14,
and an alignment position wherein the wafer 12 is displaced, still in
the same plane as image 16', to one side of the optical column 14.
In Figure 1, the table 10 is shown in the alignment
position. There the wafer 12 is positioned beneath the objective
tubes 30 and 32, respectively, of a binocular microscope mounted
7:~3
ad~jacent the optical column 14. Ali(~nment marks 34 and 3fi in the
binocular tubes 30 and 32, respectively, are on an axis which is
parallel to the transverse movement of the table 1~. The wafer 12
has alignment marks 38 and 40 adjacent its opposite edges, usually
provided by a first processing step, which are to be aligned with the
marks 34 and 36 prior to exposure. The wafer is first aligned
coarsely by locating its edges against pins or other reference
surfaces. If no alignment marks have previously been provided on the
wafer, such coarse alignment may be used alone. To assist the reader
in distinguishing marks used in positioning the reticle from those
used in positioning the wafer, the marks associated wi-th the reticle
16 and its holder 15 will be referred to as "fiducial" ~arks and
those associated with the wafer 12 and microscope tuhes 3n, 32 will
be referred to as "alignment" marks.
Alignment of the wafer 12 is effected by minute
ad~justment of the position of the table 10, which comprises three
super~jacent parts, each movable relative to the others. The two
lowerlnost parts provide movement laterally and longitudinally,
respectively, and the uppermost part provides for rotation of the
2n wafer, in the plane of imaqe 16'. Actual movement of the table 10 is
by way of laser interferometric control system 42, coupled to the
parts of the table 10 as indicated by the broken lines 44. During
alignment, the position of the table 10 may be controlled manually by
the operator. After the wafer 12 has been aligned, further movement
of the tahle 10 is controlled by a computer program,ned to move the
wafer according to a predetermined sequence. Firstly, the tahle is
moved in the direction of arrow A (Figure 1) to the exposure position
7~3
beneath the optical column 14, the computer taking account of the
spacing between the microscope alignment marks 34 and 36, to
which the wafer 12 is aligned, and the optical axis 17.
Usually an array of identical patterns, arranged in
rows and columns, are required on the wafer. Accordingly the wafer
12 is positioned so that the first latent image will be formed at -the
beginning of the first row. Thereafter the table 10 is stepped
boustrophedonically between successive exposures to produce the
required array of latent images.
The lateral distance moved by the wafer 12 in steps
within each row is equal to the distance d between corresponding
points on the two identical patterns at the image plane. It will be
appreciated that distance d will be proportional to spacing D between
the corresponding points (specifically the longitudial center lines)
of the patterns I and II on the reticle 16, taking into account the
reduction factor (if any) of the optical column 14. Typically the
reduction will be by a factor of 5 or 10, although the invention
comprehends transfer without size reduction, or with enlargment.
Figure 2 illustrates the state of the wafer 12
following the first exposure, Patterns I and II have been reproduced
as latent images I' and II ', in ad~iacent areas 54 and 56
respectively, of the wafer 12. For clarity the latent images are
shown full size and without rotation. In addition an opaque defect
III in the part of the reticle 16 carrying pattern II, has produced a
corresponding latent image III ' on the wafer 12.
After the first exposure the wafer 12 is stepped hy
distance d transversely to the longitudinal center lines 19 ard 21 of
the two patterns and a second exposure made. Figure 3 represents the
situation after the second exposure. ~afer area 5~ has been moved
outside the field of the optical column 14 and so has not been
exposed a second time. Therefore it carries only a latent image I'.
Area 56 now occupies the position vacated by area 54 and has been
exposed -through the part of the reticle lfi carrying master pattern I.
Therefore area 56 now has latent images II' and I' superposed.
The first half of the reticle 16 has no defect in the
same position as defect III, so the corresponding region of wafer
1~ area 56 has been exposed during the second exposure step.
Consequently the latent image III' has been removed, as indicated by
the broken line in area 56 in Figure 3. It should be noted that the
radiation intensity and duration are se7ected in accordance with the
sensitivity of the photoresist to ensure adequate exposure by only
one exposure.
During the second exposure, an adjacent wafer area 58
is exposed through the second part of the reticle 16, produciny
another latent ima~e II' and defect latent imaqe III'. After the
second exposure, area 58 will be stepped to the position occupied by
area 56 and re-exposed during the third exposure. The process is
then repeated.
It will be appreciated that a defect in the reticle
part carrying master pattern I will not normally produce a latent
image on the wafer 12 because the corresponding area will already
have been irradiated through the unblemished part of the reticle lh
during the preceding exposure.
l~ti 1~3
The area 54, and a corresponding area at the opposite
end of the row of latent images, receive only a single exposure and
may, if desired, be discarded when the wafer is being cut up after
processing. Alternatively the apparatus can be arranged so that the
first and last latent images are formed outside the useful area of
the wafer 12.
The invention comprehends the use of a master having
more than one pattern in each of the identical sets. For example,
the master could have two identical sets, each of two or more
patterns, arranged in a square, or in a row, the stepping distance d
being adjusted as necessary or preferred. It should be appreciated,
of course, that each set might comprise a row and/or column in an
array of identical patterns.
Figure 4 illustrates, schematically, an arrangernent for
exposing a wafer from a master in the form of a mask comprising an
array of identical patterns. The apparatus exposes different areas
of the master mask in succession until the entire array has been
projected onto the wafer. Usually the projection ratio is 1:1 i.e.
the image is suhstantially the same size as the master.
The apparatus comprises a support plate 6n carrying on
one si~e a vacuum chuck 62 which holds a semiconductor wafer 64 so
that its photoresist-coated surface is substantially perpendicular to
the optical axis 68 of an imaging device 70 in the form of a lens
or mirror system. The device 70 is positioned between the wafer 64
and a mask 72, carrying an array 74 of identical patterns to be
reproduced upon the wafer 64. The mask 72 is held generally parallel
to the wafer 64 by a vacuum chuck 76, which itself is mounted on a
11
plate 78. The plate 78 is mounted on one side of and parallel to, a
second plate, 80. A shutter plate 82 is disposed at the other side
of the plate 80 and is generally parallel thereto. A lamp 84, behind
the shutter plate serves to irradiate the mask 72 by way of a slit 86
in the shutter plate 82 and aligned holes 88, 90 and 92 in the chuck
76, and plates 78 and 80 respectively.
The shutter slit 86 is arcuate and generally vertical
and serves to restrict light from lamp 84 so that only a narrow strip
of the array 74 is illuminated at any time.
The shutter plate 82, light source 84, and imaging
device 70 are fixed relative to each other. However the plates 78
and 80, and chuck 76 are movable together, in register with the wafer
support plate 60 and check 62, transversely to the optical axis 68
and the length of the slit 86, so that a strip of light scans across
the mask 72 and in doing so, projects the array 74 onto the wafer 64.
The scanning movement of the mask 72 and wafer 64 past the imaging
device 70 may be in discrete steps or as a continuous movement. It
will be appreciated that the mask and wafer could be fixed and the
shutter 82, lamp 84, and imaging device 70 be scanned across them.
2n The intermediate plate 78 is rnoveable relative to the
plate 80 by a drive mechanism 94 under the control of a control
system 96. The control system 96 is responsively coupled to a laser
interferometer device 98, which monitors displacment of the support
member 78 relative to the plate 80 by reflecting a laser beam 100 off
a reflective surface 102 on one side edge of the support mernber 78.
Interconnections between drive mechanisms 94, control system 96 and
device 98 are indicated at 104.
s
' 12
7~
In use, a wafer h4 is loaded into the chuck 62 and the
wafer fi4 and mask 72 moved to a predetermined alignment position at
which they are aligned relative to each other if necessary.
Typically alignment marks on the mask 72 will be aliqned with
correspondin~ alignment rnarks on the wafer 64, usually provided on
the wafer 64 by a previous processing operation. Any convenient
alignment system, for example split field microscope, may be
employed.
The wafer 64 and mask 72 are then scanned across the
axis of the optical system 70 to produce on the wafer 64 a latent
image of the array 74.
The wafer 64 and mask 72 are then returned to the
alignment position and the intermediate plate 7~ moved in the same
direction as the scanning movement, by means of the control system 96
relative to the rearmost plate ~0 by a distance equal to the spacing
hetween corresponding points in ad~jacent columns of the array 74.
The wafer 64 and rnask 72 are then scanned across the optical system
again to expose the array once rnore and provide a second latent image
upon the wafer 64. With the exception of possihly one pattern at
2n each end of each row, each column of the second latent image will be
superposed upon a different column of the latent image previously
exposed onto the wafer. Clearly the offset distance could be equal
to multiples of the inter-column spacing. It is also possible to
move plate 7~ relative to plate ~0, in a direction normal to that of
the scannin~ movement, between scanning movements.
It should be noted that the use of a laser monitoring
system to monitor the displacement of the mask between scans obviates
the need for the alignment procedure to be repeated; and also allows
the process to be used even where no alignment marks have previously
been provided on the wafer 64. However, it would also be possible to
dispense with the laser metering, and realign between scans using a
set of alignment marks spaced apart from the first marks by the
distance required to superpose one column upon the next. The
prealignment to one set of alignment marks would then be carried out
as before, followed by the first scan. The wafer 64 and mask 72
would then be returned to the alignment position, offset
1n approximately by the distance required for superposition and
realigned precisely using the second set of marks before carrying out
the second exposure scan.
It will be appreciated that the need to manually
realign the mask and wafer before the second exposure might limit the
alignment accuracy ohtainable. However, with present projection
apparatus an alignment accuracy within 0.25 microns, adequate for
many applications, should be possible.
The apparatus may be modified by mounting the wafer
rather than the mask upon a holder having two or more parts, one part
being displaced relative to the other between successive exposures.
Various other modifications are possible within the
scope of the invention. For example instead of U.V. light other
radiation could be used, for example visible light, ions, x-rays or
electrons, the radiation-sensitive surface of the wafer or other
irradiated hody bein~ chosen to suit.
Although in the described embodiments the required
superposition is ohtained by translating the master and body one
14
1713
relative to the other, it is also envisaged that the images might be
translated by altering the radiation path, for example by means of
the imaging device, in which case the master and hody might be fixed
relative to each other. More specifically, the master and body might
lie in the same plane and the imaging device include means for
directing the radiation to pass from the master to the body.
Translation of the images could then be achieved by moving the
directing means relative to the body and master.
The array of images on the multiple pattern master may
itself be produced by an embodirnent of the present invention, for
example the boustrophedonically stepped method described with
reference to Figure 1.
Although specifically described with reference to
exposing wafers for integrated circuits, the invention encompasses
other manufacturing processes, for example the manufacture of masks
for subsequent use in exposing semiconductor wafers, of solar cells,
printed circuit boards, lead frames, tape-bond tapes, and hybrid
ceramics. It is further envisaged that the images might be an
enlargement of the master pattern especially when making printed
circuit boards.
It should be noted that the invention yields
considerable savings where many pattern transfer steps, each with a
different master are involved, such as in making integrated circuits
and the like. As many as twelve or more different patterns might
then be required. To obtain, in each transfer step, a double
exposure using two separate masters, each needing to be aligned to
the wafer and to the apparatus before its exposure, would require
~ 3
considerahly greater overall production time. Moreover this double
alignment would increase overall alignment inaccuracy, especially
since such alignment is often done manually. The corresponding
sources of error in embodiments of the invention are in the
positioning of the two identical patterns relative to each
other on the same master and the displacement of the master relative
to the wafer or other body. Both can be controlled very accurately
using contemporary laser interferometric position con-trol systems, or
systems of comparable accuracy.
1n Another disadvantage of double-exposure using two
masters, at least in relation to semiconductor devices, arises from
the usual practice of printing the alignment marks onto a batch of
wafers during the first processing step. Until the wafers have been
processed the alignment marks are not visible. Therefore the second
of two separate masters could not be aligned to them. This
disadvantage is overcome hy embodiments of the present invention in
which the intermediate re-alignment is not required.
While, in the examples described, the images have been
produced hy irradiation throu~h a master, that is the master is a
2n transparency for the particular radiation used, irradiation can also
be by reflection, or by a radiation having as its source the master
itself, this radiation being caused by irradiation of the master. In
all instances, the irradiation impinges on the body, or wafer, by way
of the master and can therefore be considered as being "from" the
master. The term "from" is used in the claims to encompass radiation
which passes through the master, is reflected from the master or
ernanates from the master, howsoever caused.
16
