 |
Description  |
|
|
BACKGROUND OF THE INVENTION
1. Field of the Invention
The present invention relates to automatic alignment, and more particularly
to a method and an apparatus for detecting the alignment of objects
wherein a laser beam scans the objects to be aligned and is reflected by
alignment marks to produce scattered light which is used for alignment,
during production of semiconductor integrated circuits.
2. Description of the Prior Art
In this type of device, the practice has been to dispose a semiconductor
wafer and a mask one above the other and scan their alignment marks to
detect the position of the marks. For a better understanding of such a
system, U.S. Pat. Nos. 4,167,677 and 4,199,219 and here referred to. The
mask has an alignment mark, as shown in FIG. 1, (a), having first and
second bar-like mark elements M1 and M2 which are spaced and parallel with
each other but inclined (.theta.) with respect to the scanning line A, and
also having third and fourth bar-like mark elements M3 and M4 which are
spaced and parallel with each other but inclined (.theta.) oppositely with
respect to the scanning line A. The wafer has an alignment mark, as shown
in FIG. 2, (b), having first and second bar-like mark elements W1 and W2
which are inclined with respect to the scanning line A at the same angle
.theta. but in different directions. Those marks of the mask and wafer are
superimposed as shown in FIG. 1 (c) to achieve alignment therebetween.
When the marks in the state of FIG. 1, (c) are scanned by a laser beam
spot along the scanning line A, the beam is scatteredly reflected by the
mark elements M1, M2, M3, M4, W1 and W2. The scattered light is received
by a photoreceptor which produces pulse signals in response to the
positions of the respective mark elements, as shown in FIG. 1, (d). The
pulse signals are reformed into rectangular pulses by cutting the tops of
the pulse signals at a suitable threshold V with the aid of a comparator.
The reformed pulses are shown in FIG. 1, (e). The positional relationship
between the wafer and the mask can be obtained from the time intervals
between the mark elements. The wafer or mask is displaced in response to
the positional relationship thus obtained so as to provide alignment
therebetween.
Generally, scanning is effected by a spot of the laser beam. With a spot
beam, erroneous or less accurate alignment can occur, when a piece of dust
or an aluminium particle is present on the scanning line. In the case of
semiconductor wafers in the manufacture of integrated circuits, alignment
is required to a degree within microns or even to a degree within
sub-microns. So in this field, the alignment is required to be of ultimate
accuracy.
After the positions of the alignment marks are detected, the alignment
drive is effected to the wafer and/or mask. It should be noted that it is
seldom that one drive results in a correct alignment, usually the
alignment operation is repeated several times. Also, in order to enhance
the accuracy of the detection, the alignment marks are scanned several
times to take an average result. Further, in the case of a step and repeat
type aligner wherein the alignment operation is carried out for each shot,
the alignment operation is executed tens of times for one wafer.
Therefore, reduction of the time required for the alignment is highly
desired in addition to the accuracy thereof.
SUMMARY OF THE INVENTION
It is the principal object of the present invention to detect the alignment
marks efficiently, more particularly to reduce the time required for the
detection.
It is another object of the present invention to increase the contact
between an edge of the alignment mark and the illumination area of the
scanning light to mitigate the deterioration of the signal which may be
caused by a partial destruction of the mark edge.
It is a further object of the present invention to minimize the possibility
of taking false signals, given by dust, evaporated particles or the like
for true signals.
The present invention will be described in detail with a preferred
embodiment which relates, as an example, to the relative alignment between
a mask and a wafer. However, the invention also is applicable to the
positional alignment for working an inspecting apparatus for a
semiconductor device, a wire bonder or dicing machine, wherein the
alignment is an absolute one with respect to a reference.
BRIEF DESCRIPTION OF THE DRAWINGS
FIG. 1 shows conventional alignment marks of mask and wafer, wherein (a)
shows the alignment mark of a mask; (b) shows the alignment mark of a
wafer; (c) shows the state wherein the marks are superimposed; (d) shows
the output signals produced as a result of scanning the marks of (c); and
(e) shows the pulses reformed from the signals of (c).
FIG. 2 shows an optical system according to an embodiment of the present
invention.
FIG. 3 is a block diagram of the control system for the optical system of
FIG. 2.
FIG. 4 shows the operation of an embodiment of the present invention
wherein (a) shows an example of positions of alignment marks; (b), (c),
(e) and (f) show the pulses obtained by the scanning, and (d) shows a
period of time for a certain checking.
FIG. 5 is a perspective view of a part of a device according to another
embodiment of the present invention.
DESCRIPTION OF THE PREFERRED EMBODIMENTS
Referring now to FIG. 2, an embodiment of the present invention, that is,
an alignment and exposure system, wherein a mask 1 is held by a mask
holder not shown. A wafer 2 is carried on a wafer stage 3. The mask 1 has
alignment marks M at two different locations, and the wafer 2 has
corresponding alignment marks W also at two different locations. The marks
are as shown in FIG. 1, (a) and (b), which are superimposed as shown in
FIG. 1 (c). The alignment and exposure system includes a laser beam source
10, a beam deflector 11, such as an acousto-optic element, juxtaposed
cylindrical lenses 12 and 13 having respective power axes which are
orthogonal, prism 14 and a rotatable polygonal mirror 15. The beam
deflector 11, cylindrical lenses 12 and 13, prism 14 and polygonal mirror
15 are arranged in the order named along the channel of the laser beam L
generated by the laser beam source 10. Along the channel of the laser beam
L which has been scanningly reflected by the polygonal mirror 15, there
are arranged intermediate lenses 16 and 17 and a prism 18 which deflects
the laser beam in either of two orthogonal directions depending upon the
angle of the mirror 15 deflection. At each side of the prism 18, there is
provided a series of optical systems. These optical systems are
symmetrical and include half mirrors 19a and 19b for deflecting the laser
beam L and passing the beam reflected by the mask 1 and/or wafer 2 to a
photorecptor, intermediate lenses 20a and 20b, aperture stops 21a and 21b
and objective lenses 22a and 22b. On the optical path for the beam
reflected by the mask 1 and/or wafer 2, there are provided after the half
mirrors 19a and 19b, mirrors 23a and 23b, imaging lenses 24a and 24b,
partial light blocking plates 25a and 25b which have central light
blocking portions for filtering, condenser lenses 26a and 26b and
photo-electric transducers 27a and 27b, which are all arranged
symmetrically so as to form two symmetric two series of photoelectric
detection optical systems.
The part of the optical system from the laser source 10 to the polygonal
mirror 15 will be further explained. The laser beam L is selectively
deflected to one of the two optical paths by the deflector 11, so that the
laser beam L is incident on one of the cylindrical lenses 12 and 13, each
of which has a condensing power in only one direction. The direction of
one of the cylindrical lenses 12 and 13 is perpendicular to the direction
of the other. The beam L which has passed through one of the cylindrical
lens has a slit-like cross-section, since it has a power in only one
direction. The laser beam L having a slit-like cross-section is refracted
by a prism 14 to be incident on the polygonal mirror 15 surface at the
position B as shown in FIG. 2. This applies both to the beam passed
through the cylindrical lens 12 and the beam passed through the
cylindrical lens 13. The beam deflector 11 is actuated by a driver 28
which is controlled by a controller 29.
FIG. 3 is a block diagram of means for controlling the inclination of the
area, illuminated by the laser beam on the surface to be scanned, in
accordance with output signals of the photoelectric transducers 27a and
27b. The output of the photoelectric transducers 27a and 27b is conveyed
through a wave reformation circuit 31 and a counter circuit 32 to the
control circuit 29 described above. Also connected to the control circuit
29, is the output of a timer circuit 33. The control circuit 29 produces
output to the light deflector 11 through the driving circuit 28 as
described above. The pulse signals produced by the wave reformation
circuit 31 are conveyed to a processing circuit 34 wherein the degree of
the alignment is discriminated. In accordance with the output of the
processing circuit, a driving mechanism D (FIG. 2) moves and/or rotates
the wafer stage by the amount of misalignment obtained by the processing
circuit 34 to bring the mask 1 and wafer 2 into alignment.
Since the embodiment of the present invention has the structures described
above, the laser beam L generated by the laser beam source 10 takes one of
the optical paths determined by the beams deflector 11 which is controlled
by the control circuit 28, and the beam L is incident on the first
cylindrical lens 12 to be reformed into a slit-like beam. The reformed
beam is then refracted by the prism 14 to be incident on the origin B of
deflection on the polygonal mirror 15, by which the beam is scanningly
deflected. The beam L, after being deflected by the polygonal mirror 15,
passes through the intermediate lenses 16 and 17 and reaches one of the
inclined surfaces 18a of the prism 18, by which the beam is reflected
leftwardly in FIG. 2, and then reflected at right angle downwardly by the
half mirror 19a. The beam L is then passed through the intermediate lens
20a, aperture stop 21a and the objective lens 22a to be imaged on the mask
1 and wafer 2. The orientation of the first cylindrical lens 12 is so
determined that the slit-like illumination area, on the mask 1 or wafer 2,
formed by the laser beam L incident thereon extends at an angle
45.degree.0 to the scanning line A as shown by reference character l on
FIG. 4, (a), that is, substantially parallel to the alignment marks M1, W1
and M2. Thus, the beam L illuminates the first group of the alignment mark
elements as a slit-like beam substantially parallel to those elements.
When the illumination slit l scans along the scanning line A, the beam is
scatteredly reflected by the mark elements M1, W1 and M2. A part of the
scattered light is directed back through the objective lens 22a, the
aperture stop 21a, the intermediate lens 20a, the half mirror 19a, the
imaging lens 24a, the partial light shield 25a and the condenser lens 26a,
and finally reaches the photoelectric transducer 27a, which produces
output signals S1, S2 and S3 as shown in FIG. 4, (b). It should be noted
here that the illumination slit l formed by the reformed laser beam is
substantially superimposed on the substantial length of each of the mark
elements M1, W1 and M2, so that the sensitivity of the mark detection is
much higher than conventional scanning by spot illumination, thus
increasing the accuracy of the mark detection. The beam unscatteredly
reflected by the flat or mirror surfaces of the mask 1 and wafer 2 is
imaged on the central position of the blocking plate 25a so that such a
beam does not reach the photoelectric transducer 27a.
The signals S1, S2 and S3 are inputted to the wave reformation circuit 31,
and reformed into rectangular pulses P1, P2 and P3 as shown in FIG. 4,
(c), with the pulse width equal to the width of signals S1, S2 and S3 cut
at a predetermined threshold level. The pulses P1, P2 and P3 are given to
the counter circuit 32 and processing circuit 34. When a predetermined
number, three in the shown embodiment, are counted, a signal is given to
the control circuit 29. To the control circuit 29, another signal T is
applied from the timer circuit 33. The signal T is the one applied at a
predetermined timing and with a predetermined width H, as shown in FIG. 4,
(d). When the third pulse P3 is within the time period of H, the control
circuit 29 generates a signal to actuate the driving circuit 28 to change
the beam L channel to the other optical path by the deflector 11. By this,
the laser beam L becomes incident on the other cylindrical lens, i.e., the
second cylindrical lens 13. As described above, the power axis of the
second cylindrical lens 13 is inclined at a predetermined angle with
respect to that of the first cylindrical lens 12. So, the direction of the
slit-like illumination area on the mask and wafer changes to the state as
shown by the reference character l' in FIG. 4, (a). That is, the
illumination area is at 135.degree. to the scanning axis and substantially
parallel with the mark elements M3, W2 and M4.
At the same time with the change, the direction of the beam incident on
mirror 15 changes so that the direction of the beam L reflected by the
prism 14 also changes. Therefore, the scanning line is not continuous, at
the point of the optical path change. It is preferable to determine the
distance between the mark elements corresponding to the interval, or to
determine the inclination of the prism 14 surface to fly the slit-like
illumination area backwardly.
The mark elements M3, W2 and M4 are scanned by the slit-like illumination
area l' having the different inclination. Similarly to the above described
operation, the photoelectric transducer 27a produces the output signals
S4, S5 and S6 as shown in FIG. 4, (b), and the wave reformation circuit 31
produces pulses P4, P5 and P6 as shown in FIG. 4, (c). With this, the
detection of the first group of mark elements finishes. If desired, the
pulses are processed together with the pulses P1, P2 and P3 previously
obtained to determine the state of alignment.
If, on the other hand, the third pulse P3 is not generated during the time
period T, that is, the mark element M2 is not detected, the control
circuit actuates the driving circuit 28 to change the optical path, after
the point of time within which the third pulse should have been produced.
By this, the inclination of the slit-like illumination area is changes
even if the third pulse is not generated. In this case, the pulses
transmitted from the wave reforming circuit 31 to the processing circuit
34 are as shown in FIG. 4, (e), since the third pulse is missing.
It is possible, because of the existence of dust or other foreign matter,
that larger than a predetermined number of pulses is produced. An example
of this is shown in FIG. 4, (f). The pulse P7 is shown as having been
produced by something other than the marks M and W. On this occasion, it
is not proper to change the optical path as soon as the third pulse P7 is
detected. To prevent this, the time period H is provided, so that the
optical path change takes place only when the third pulse is detected
during the period H, or after the period has elapsed.
As will be apparent from the foregoing, the direction of the slit-like
illumination area changes during one scanning operation of the marks M
(M1, M2 and M3) and W (W1 and W2). That is, the change takes place during
the period the beam L is being deflected in one direction. This saves
time, since it is not necessary to effect scanning operation in the manner
that the beam is first deflected through its full span to scan the object
with a first slit-like illumination area, and then the beam is changes to
provide a second slit-like illumination area, whereafter the beam repeats
the full span deflection to scan the object with the second slit-like
illumination area.
The laser beam further continues to be deflected by the polygonal mirror
15. The beam L reaches the other inclined surface 18b of the prism 18.
Then, the beam L is reflected rightwardly to pass through the intermediate
lens 20b, the aperture stop 21b and the objective lens 22b and reach the
mask 1 and wafer 2 at the second group of alignment mark elements. And the
detecting operation is carried out in the similar manner as described
above.
In a projection type alignment and exposure system, a large projection lens
40 is provided between the mask 1 and wafer 2, as shown in FIG. 5. The
laser beam L emitted from the laser source 10 is imaged on the mask 1 by
objective lenses 22a and 22b, and on the wafer 2 by the projection lens.
Since the principal ray of the laser beam L passes through the center of
the projection lens 40, the scanning operation is first effected on the
second group of the mark elements and then on the first group of the mask
elements, as opposite to the case, described hereinbefore, wherein the
mask 1 and the wafer 2 are contacted.
Although the foregoing explanation has been made with respect to an
alignment and exposure apparatus, more particularly, a mask aligner, it is
understood that the present invention is applicable to other objects, such
as the master alignment in a printing field.
It is possible to use an electro-optic element in place of the
acousto-optic element.
According to the present invention, described in detail hereinbefore, the
following advantages are provided:
(1) The illumination beam has a slit-like cross-section having an
inclination corresponding to that of the mark element, so that
signal-to-noise ratio is improved, resulting in better detection
sensitivity. Even if some foreign matters is present around the mark, the
light scattered by them is not detected; and
(2) The inclination of and the slit-like illumination area can be changed
during one scanning, the number of scanning operations is minimized.
While the invention has been described with reference to the structures
disclosed herein, it is not confined to the details set forth and this
application is intended to cover such modifications or changes as may come
within the purposes of the improvements or the scope of the following
claims.
* * * * *
|
|
 |
|
 |
Description |
|
