U.S. patent number 3,895,234 [Application Number 05/402,248] was granted by the patent office on 1975-07-15 for method and apparatus for electron beam alignment with a member.
This patent grant is currently assigned to Westinghouse Electric Corporation. Invention is credited to Paul R. Malmberg, Terence W. O'Keeffe.
United States Patent |
3,895,234 |
O'Keeffe , et al. |
July 15, 1975 |
Method and apparatus for electron beam alignment with a member
Abstract
A method and apparatus are provided for alignment of an electron
beam with precisely located areas of a major surface of a member.
Marks of predetermined shape are formed of cathodoluminescent
material and are positioned adjacent the major surface of the
member which is preferably substantially transparent to the
cathodoluminescense generated by the marks. An electron beam to be
aligned has at least one alignment beam portion of a predetermined
cross-sectional shape and preferably corresponds in size to the
alignment accuracy desired. The cathodoluminescence emissions
detected by a detector means are preferably positioned adjacent the
opposite surface of the member. The position of the electron beam
is moved relative to the member while continuing said detection
until the emissions detected indicate alignment of the alignment
beam portion with a corresponding mark. Preferably, said alignment
method is used in producing a very accurate component pattern in an
electroresist layer supported by a member utilizing either a
scanning electron beam or an electron image projection system for
the electron radiation beam, and most preferably in the making of a
novel photocathode source, which can itself be used in the
alignment of a patterned electron beam with a member.
Inventors: |
O'Keeffe; Terence W.
(Pittsburgh, PA), Malmberg; Paul R. (Pittsburgh, PA) |
Assignee: |
Westinghouse Electric
Corporation (Pittsburgh, PA)
|
Family
ID: |
27005005 |
Appl.
No.: |
05/402,248 |
Filed: |
October 1, 1973 |
Related U.S. Patent Documents
|
|
|
|
|
|
|
Application
Number |
Filing Date |
Patent Number |
Issue Date |
|
|
370558 |
Jun 15, 1973 |
|
|
|
|
Current U.S.
Class: |
250/492.2;
219/121.29; 250/397; 315/10; 219/121.12; 219/121.3 |
Current CPC
Class: |
H01J
37/3045 (20130101); H01J 37/30 (20130101) |
Current International
Class: |
H01J
37/30 (20060101); H01J 37/304 (20060101); H01j
029/50 (); H01j 031/49 () |
Field of
Search: |
;250/549,561,458,459,396,397,398,400,458,459,491,492A
;219/121EB,121EM ;315/10 |
References Cited
[Referenced By]
U.S. Patent Documents
Primary Examiner: Stolwein; Walter
Attorney, Agent or Firm: Menzemer; C. L.
Government Interests
GOVERNMENT CONTRACT
This invention is made in the course of or under Government
Contract F 30602-69-C-0280.
Parent Case Text
RELATED APPLICATION
This application is a continuation-in-part of copending application
Ser. No. 370,558, filed June 15, 1973, now abandoned.
Claims
What is claimed is:
1. A method of precision aligning an electron beam with selected
areas of a major surface of a substrate member for electronic
components comprising the steps of:
A. forming adjacent a major surface of a member a plurality of
marks of predetermined shape capable of generating
cathodoluminescent radiation on irradiation by an electron beam
corresponding to the area of the mark irradiated;
B. irradiating at least one of the marks with a corresponding
alignment beam portion of an electron beam to be aligned, said
alignment beam portion having a predetermined cross-sectional
shape;
C. detecting cathodoluminescent radiation emissions from said
irradiated mark;
D. moving the electron beam relative to the member until the
cathodoluminescent radiation emissions detected in accord with step
C indicate alignment of the alignment beam portion with the
corresponding mark.
2. A method of precision aligning an electron beam with selected
areas of a major surface of a member as set forth in claim 1
wherein:
the predetermined cross-sectional shape of the alignment beam
portion of the electron beam corresponds in size to the accuracy of
alignment desired.
3. A method of precision aligning an electron beam with selected
areas of a major surface of a member as set forth in claim 2
wherein:
the predetermined shape of the marks correspond in size to the
accuracy of alignment desired.
4. A method of precision aligning an electron beam with selected
areas of a major surface of a member as set forth in claim 1
wherein:
the predetermined shape of each mark is substantially the same as
the predetermined cross-sectional shape of the corresponding
alignment beam portion.
5. A method of precision aligning an electron beam with selected
areas of a major surface of a member as set forth in claim 1
wherein:
forming the marks includes recessing the marks into the major
surface to provide a substantially planar major surface with the
marks position.
6. A method of precision aligning an electron beam with selected
areas of a major surface of a member as set forth in claim 1
wherein:
forming each mark includes positioning an opaque layer adjacent a
layer of cathodoluminescent material to circumscribe a portion of
said layer in the predetermined shape and cause a differential in
cathodoluminescent radiation projected by said layer at said
portion on irradiation.
7. A method of precision aligning an electron beam with selected
areas of a major surface of a member as set forth in claim 1
wherein:
the major surface of the member is divided into contiguous fields
with the marks positioned symmetrically along boundaries between
said contiguous fields.
8. A method of precision aligning an electron beam with selected
areas of a major surface of a member as set forth in claim 7
wherein:
the electron beam to be aligned is a scanning electron beam for
selectively irradiating an electroresist layer on the major surface
of the member.
9. A method of precision aligning a patterned electron beam
generated by a photocathode source with selected areas of a major
surface of a substrate member for electronic components comprising
the steps of:
A. forming adjacent a major surface of a member having at least two
widely spaced marks of predetermined shape capable of generating
cathodoluminescent radiation or irradiation by an electron beam
corresponding to the area of the mark irradiated;
B. irradiating the marks with corresponding alignment beam portions
of a patterned electron beam generated by a photocathode source,
each said alignment beam portion having a predetermined
cross-sectional shape;
C. detecting cathodoluminescent radiation emissions from said
irradiated mark; and
D. moving the patterned electron beam relative to the member until
the cathodoluminescent radiation emissions detected in accord with
step C indicate alignment of the alignment beam portions with the
corresponding marks.
10. A method of precision aligning a patterned electron beam
generated by a photocathode source with selected areas of a major
surface of a member as set forth in claim 9 wherein:
the predetermined cross-sectional shape of the alignment beam
portions of the electron beam correspond in size to the accuracy of
alignment desired.
11. A method of precision aligning a patterned electron beam
generated by a photocathode source with selected areas of a major
surface of a member as set forth in claim 10 wherein:
the predetermined shape of the marks correspond in size to the
accuracy of alignment desired.
12. A method of precision aligning a patterned electron beam
generated by a photocathode source with selected areas of a major
surface of a member as set forth in claim 9 wherein:
the predetermined shape of each mark is substantially the same as
predetermined cross-sectional shape of the corresponding alignment
beam portion.
13. A method of precision aligning a patterned electron beam
generated by a photocathode source with selected areas of a major
surface of a member as set forth in claim 9 wherein:
forming the marks includes recessing the marks into the major
surface to provide a substantially planar major surface with the
marks positioned.
14. A method of precision aligning a patterned electron beam
generated by a photocathode source with selected areas of a major
surface of a member as set forth in claim 9 wherein:
forming each mark includes positioning an opaque layer adjacent a
layer of cathodoluminescent material to circumscribe a portion of
said layer in the predetermined shape and cause a differential in
cathodoluminescent radiation projected by said layer at said
portion on irradiation.
15. A method of precision aligning a patterned electron beam
generated by a photocathode source with selected areas of a major
surface of a member as set forth in claim 9 wherein:
step D is automatically performed by electrically processing the
electrical signal output corresponding to the the detected
cathodoluminescent radiation on modulation of the movement of the
alignment beam portions over the corresponding marks.
16. Apparatus for selectively irradiating precisely located areas
of a major surface of a substrate member for electronic components
comprising:
A. a photocathode source for generating a patterned beam of
electrons including at least one alignment beam portion of
predetermined cross-sectional shape;
B. at least one mark corresponding to each alignment beam portion
and capable of generating cathodoluminescent radiation on
irradiation with an electron beam corresponding to the area of the
mark irradiated with the electron beam, each said mark being
supported adjacent a major surface of a member to be aligned with
the patterned beam of electrons and having a predetermined
shape;
C. means for positioning the member in a spaced relation to the
photocathode source for the patterned beam;
D. means for applying a potential between the member and the
photocathode source whereby electrons from the photocathode source
are directed to and selectively irradiate selected areas of the
major surface of the member;
E. electromagnetic means for directing the patterned beam of
electrons from the photocathode source to irradiate selected
portions of the major surface of the member close to the selected
areas wherein each alignment beam portion irradiates selected
portions of the surface portions of the major surface of the member
close to a corresponding mark;
F. detector means for detecting the cathodoluminescent radiation
projected by the corresponding mark and producing an electrical
signal corresponding to the area thereof irradiated by the
alignment beam portion; and
G. electrical means for moving the patterned beam of electrons
relative to the member responsive to said electrical signal from
the detector means to cause the alignment beam portions to align
with corresponding marks, whereby the patterned beam of electrons
from the photocathode source is located relative to the member so
that precisely located areas of the major surface of the member can
be selectively irradiated with the patterned electron beam.
17. Apparatus for selectively irradiating precisely located areas
of a major surface of a member as set forth in claim 16
wherein:
the patterned beam of electrons generated by the photocathode
source includes at least two spaced apart alignment beam portions
of predetermined cross-sectional shape.
18. Apparatus for selectively irradiating precisely located areas
of a major surface of a susbstrate as set forth in claim 17
wherein:
the electrical means includes modulation means for oscillating the
movement of each alignment beam portion on a corresponding mark,
phase detection means for detecting along orthogonal axes the error
from coincidence of the alignment beam portions and the marks and
outputting the electrical signal corresponding thereto, and
actuating means for changing the electrical input to the
electromagnetic means responsive to the electrical signal from the
phase detector means to bring the alignment beam portions and the
marks into alignment.
Description
FIELD OF THE INVENTION
The invention relates to the making of integrated circuits and
other micro-miniature electronic components with submicron
accuracy.
BACKGROUND OF THE INVENTION
The present invention is an improvement on the electron beam
fabrication system described in U.S. Pat. No. 3,679,497, issued
July 25, 1972, and assigned to the assignee of the present
invention.
The electron beam fabrication system employs a scanning electron
microscope to produce a photocathode source or electromask. The
photocathode source is adapted to project, on irradiation,
typically by ultraviolet radiation, an electron beam in the desired
pattern which impinges on an electroresist layer on a major surface
of a member to implant in said resist layer a differential
solubility between the irradiated and the unirradiated areas.
Removing the more soluble portions of the electroresist layer after
irradiation, selectively exposes the underlying member or layer on
the member which can in turn be selectively altered through windows
in the electroresist of the desired component pattern typically by
etching, diffusion or deposition.
The electromask designates the pattern-bearing photocathode
assembly which is analogous to the photomask in the well known
photolithographic techniques. The electromask comprises a light
transmissive substrate such as quartz by which the photocathode
source is supported. The photocathode source is usually adapted to
generate the patterned electron beam by overlaying the substrate
with a layer such as titanium dioxide which is opaque to the
radiation to which the photocathode material is sensitive. The
negative of the desired component pattern is formed in the opaque
layer over which a contiguous layer of a photocathodic material
such as palladium is formed. The patterned electron beam thereafter
is generated by irradiating the photocathode layer through the
substrate an the opaque layer. See, e.g., U.S. Pat. Nos. 3,585,433,
3,588,570, 3,686,028 and 3,672,987.
The scanning electron microscope of such fabrication system
involves the use of a finely focused electron beam to generate a
planar component pattern having submicron accuracy in an
electrosresist layer or the like. The electron beam is
automatically moved through the pattern matrix on command from a
computer. The beam control information can be stored on a magnetic
tape which is fed into the computer which is used to command the
position and movement of the electron beam. While such as scanning
electron beam system can be used to directly develop a high
resolution pattern in an electroresist in making an integrated
circuit, the electron beam fabrication system involves the use of
such a scanning electron beam to make the photocathode source for
the electron image projection system.
The main problem in the use of such scanning electron beams is
maintaining the accuracy over the entire pattern field. Resolution
as such is not a problem in the use of the scanning electron beam.
Lines less than 0.5 micron in width can be reliably reproduced in
resist materials. However, the pattern field size of integrated
circuits are typically as great as 2000 .times. 2000 microns and
often as great as 4000 .times. 4000 microns. For electron optical
reasons it is impractical to deflect the electron beam through more
than a few degrees and still hold the high resolution of, for
example, 0.5 micron. It is possible to increase the field size by
increasing the distance between the cathode and the electroresist,
but this also correspondingly increases the diameter of the focused
electron beam and in turn sacrifices resolution. Thus, for a given
resolution, the field size for the scanning electron beam is
usually restricted to considerably less than the member size of the
integrated circuit. For example, for an electron beam of 0.2 micron
in diameter and a resolution of 0.5 micron, the throw of the
electron beam is limited to 2 inches and field size is limited to
about 2000 microns in diameter.
Another restriction on the size of the field is the accuracy with
which the electron beam can be deflected. This depends primarily on
the electronics used to control the beam position. At present, the
deflection accuracy of an electron beam of 0.2 micron diameter in a
2000 micron field is at best about 0.5 micron. Thus, a typically
2000 .times. 2000 micron field of an electron beam fabrication
system is divided into 4000 .times. 4000 array of points each 0.5
micron apart; with the 4000 addresses on each axis provided by a 12
bit digital-analog converter. While memories are available with far
greater numbers of addresses, larger fields and greater resolution
within the same size fields cannot be obtained because of beam
deflection limitations.
Ideally the problem can be remedied by sequentially developing the
electroresist in multiple patterns of small fields. However,
accurate registrations between contiguous fields is essential to
the operability of such a technique. The high resolution, e.g. 0.5
micron, of the electron beam fabrication system is lost unless the
same resolution can be maintained in alignment of the successive
patterns. Thus, the electron radiation for each field must be
aligned with respect to the adjoining field with a precision of 0.5
micron or less. Otherwise the precisions and economics of the
electron beam fabrication system will not be attained in the
integrated circuit device.
It has been proposed to simply accurately move the member on which
the integrated circuit is to be formed from field to field by
mechanical means, using a laser beam to maintain registration, see
e.g., U.S. Pat. Nos. 3,632,205 and 3,719,780. However, such a
system is not believed commercially practical. A change of only
0.01 % in the physical alignment of the member can lead to an 0.2
micron error in the registration of adjacent fields. On a 2 .times.
2 inch member such an alignment error results from an 0.002 inch
variation in dimension. Further, the deflection aberrations of the
electron beam are such that an exactly regular rectangular array of
fields is impossible to attain. Even with compensation, distortions
remain that add to the mismatch betwen adjacent fields. Thus, a
need exists for a multiple sequential alignment system which will
accurately and rapidly register adjacent fields of a desired
electroresist pattern.
The need is particularly acute in Large Scale Integration (LSI)
technology. LSI is the term applied to integrated circuits which
provide high complexity electronic circuits (e.g. more than a
thousand gates) in the same semiconductor wafer. Logically LSI
technology requires larger and larger wafers. And LSI wafers
measuring 2 and 3 inches on a side are now used. However, such
large patterns can be generated with a scanning electron microscope
only by combining several fields, where accurate registration
between fields is essential. Moreover, LSI technology requires a
greater density of electronic components and in turn better
generation and alignment resolution. The quantative yield of such
integration corresponds directly to the size of the wafer. The
probability of defects in the single crystal structure increases
directly with the volume of the wafer. Thus, the higher the
resolution, the greater the circuit density, the smaller the wafer,
the greater the quantative yield of the integration that
results.
In the electron image projection system, there is also a need for
precision registration and alignment of the patterned electron beam
and the member. In order to attain high quantative yields of
consistent and reliable electronic performance in integrated
circuits, it is necessary that the major surfaces of the member be
altered in precisely located positions and areas. This precision is
particularly important when a plurality of successive surface
treatments and alternations are to be made to one member. In this
connection, six or more successive treatments and alterations are
commonly needed to produce a typical integrated circuit for a
sophisticated electronic application. The high resolution of the
electron image projection system is lost in the juxtaposition of
planar component patterns unless the same high resolution of
projection, e.g. 0.5 micron and less, can be maintained in
alignment of the member with successive photocathode sources. Here
again, LSI technology puts particular stress on this
requirement.
SUMMARY OF THE INVENTION
A method and apparatus are provided for the alignment of an
electron beam with selected areas of a major surface of a member
with any desired degree of accuracy, example, 0.5 micron or less.
The optical distortion and deflection inaccuracies of a scanning
electron beam can all but be eliminated by permitting the scan to
be done in smaller fields, e.g. 200 .times. 200 microns square,
because of the precision alignment which the invention permits to
be maintained between fields. The invention also permits precision
alignment of a photocathode source or series of photocathode
sources with selected areas of a major surface of a member in the
electron image projection system as well as precision production of
photocathode sources without detector marks which detract from the
patterned electron beams projected by the sources.
Generally at least one and preferably at least two detector marks
capable of generating cathodoluminescent radiation corresponding to
the area of each mark irradiated by an electron beam are formed
adjacent a major surface of a member. The marks are of a
predetermined shape and preferably in accurate spatial relation
relative to each other. At least one cathodoluminescent mark is
irradiated with a corresponding alignment beam portion of the
electron beam to be aligned. The alignment beam portion is of a
predetermined cross-sectional shape and preferably corresponds in
size to the desired accuracy, e.g. 0.5 micron in diameter or width
for 0.5 micron accuracy. The light radiation emissions from the
irradiated cathodoluminescent mark or marks detected by a
photodetector means preferably through the member. The electron
beam is moved relative to the substrate until the
cathodoluminescent radiation emissions detected from the irradiated
marks indicates alignment of the alignment beam portion with a
corresponding detector mark.
The alignment beam portions and the detector marks may be of any
suitable relative size within practical limi-s provided the shapes
of both are predetermined. Preferably, however, each alignment beam
portion is of the same cross-sectional shape as the predetermined
shape of the corresponding detector mark so that alignment can be
determined simply by reading a maximum or a minimum in the
electrical signal from the detector means. Otherwise, electrical
processing of the electrical signals are needed, while the
alignment beam portions are oscillated over the corresponding
detector marks, to determine optimum alignment of the alignment
beam portions with the corresponding detector marks.
The detector marks of cathodoluminescent material may in certain
applications be directly formed on or in the major surface in the
predetermined shape. In other applications it may be more
appropriate that the cathodoluminescent marks be formed by forming
adjacent a cathodoluminescent layer a layer of material opaque to
light or electron radiation. The opaque layer causes a differential
in cathodoluminescent radiation projected by the cathodoluminescent
layer either by blocking the cathodoluminescence generated by the
cathodoluminescent layer of blocking the electron radiation
irradiating the cathodoluminescent layer. To illustrate, an opaque
layer capable of reducing the cathodoluminescent emissions of a
cathodoluminescent layer is formed adjacent the major surface of
the member, windows are then opened in the opaque layer
corresponding to the desired cathodoluminescent marks, and
contiguous layers of cathodoluminescent material are formed over
the major surface at least at the windows so that the opaque layer
circumscribes a portion of the cathodoluminescent layer in the
predetermined shape.
Where the cathodoluminescent radiation is detected on reflection by
a photodetector positioned on the same side of the member as the
electron beam, the steps would be the same and the sequence
altered; the relative positions of the opaque layer and the
cathodoluminescent layer with respect to the member are reversed.
In a similar embodiment, the opaque layer is positioned over the
cathodoluminescent layer to circumscribe an exposed portion of the
cathodoluminescent layer in the predetermined shape. The opaque
layer substantially reduces the electron radiation irradiating the
cathodoluminescent layer an in turn reduces the cathodoluminescent
radiation generated by the cathodoluminescent layer rather than
reducing the cathodoluminescent radiation projected.
Further, the negative of these various embodiments may be desired
in certain applications. That is, the opaque layer may be in the
predetermined shape either above or below the cathodoluminescent
layer. Or, the cathodoluminescent layer may simply circumscribe an
exposed portion of the major surface, the exposed portion being in
the predetermined shape for the cathodoluminescent marks.
Preferably, the alignment method is used in producing a very
accurate component pattern in an electroresist layer supported by a
member in making integrated circuits of high component density and
high resolution. A scanning electron beam for an electron image
projection system, such as are described in above-cited U.S. Pat.
No. 3,679,497, is preferably utilized as the electron radiation
beam. Where a scanning electron beam is used, the beam operates in
its entirety as the alignment beam portion in the system, and the
selective irradiation is preferably performed in contiguous fields,
e.g. 200 .times. 200 microns. The electron beam is thereby aligned
with the member at least between the irradiating of each field to
register the member and the electron beam from field to field
preferably by positioning the marks symmetrically along the
boundaries of the fields.
Where an electron image projection system is used, the selective
irradiation is usually performed simultaneously over the entire
surface of the member and preferably at least two widely spaced
apart cathodoluminescent marks of predetermined shape are
simultaneously irradiated in performing the alignment. The
patterned electron beam projected by the photocathode source
includes alignment beam portions of predetermined cross-sectional
shape which are preferably the same shape as corresponding
cathodoluminescent mark or marks. Further, with an electron image
projection system, the alignment method can be used in
juxtaposition alignment of a set of component patterns in making an
integrated circuit or other micro-miniature electronic component. A
set of photocathode sources are prepared having as part of their
patterned electron beam alignment beam portions of substantially
identical cross-sectional shape and spatial position. The alignment
steps can thus sequentially be repeated with the different
photocathode sources of the prepared set in making a precision
integrated circuit.
Preferably the present invention is utilized in making a novel
photocathode source for use in the electron image projection system
as described in above-cited U.S. Pat. No. 3,679,497. Such a
photocathode source is adapted to project an electron beam in a
predetermined component pattern on being irradiated with
sensitizing light radiation such as ultraviolet light. The
photocathode source of the present invention includes alignment
marks of predetermined shape capable of generating
cathodoluminescent radiation corresponding to the area of the mark
irradiated and preferably substantially transparent to the light
radiation to which the photocathode material is sensitive. Said
cathodoluminescent marks are positioned adjacent a major surface of
a substrate substantially transparent to the light radiation to
which the photocathode material is sensitive and by which the
photocathode material is supported. In certain instances, the
cathodoluminescent marks of said photocathod source may be formed
as above described with adjacent opaque layers which circumscribe
portions of cathodoluminescent layers in the predetermined shape
and cause a differential in cathodoluminescent radiation projected
by the cathodoluminescent layers at said portions to form the
desired cathodoluminescent marks.
Other details, objects and advantages of the invention will become
apparent as the following description of the present preferred
embodiments and present preferred methods of practicing the same
proceeds.
BRIEF DESCRIPTION OF THE DRAWINGS
In the accompanying drawings, the present preferred embodiments of
the invention and present preferred methods of practicing the
invention are illustrated in which:
FIGS. 1, 2, 6 and 7 are fragmentary cross-sectional views of a
photocathode source of the present invention at various stages of
its manufacture;
FIGS. 3, 4 and 5 are alternative fragmentary top views of the
photocathode source of FIGS. 1, 2, 6 and 7 at a certain stage in
its manufacture,
FIGS. 8, 9 and 10 are fragmentary cross-sectional views of an
alternative photocathode source of the present invention at various
stages of its manufacture;
FIG. 11 is a schematic illustration of production of a highly
accurate component pattern in an electroresist layer on a member
utilizing a scanning electron beam in accordance with the present
invention;
FIG. 12 is a partial top view of the member of FIG. 11 without the
electroresist layer applied;
FIG. 13 is a flow diagram showing the inter-relationship of
functional component in utilizing the present invention to align
the scanning electron beam as shown in FIG. 11;
FIG. 14 is a cross-sectional view in elevation of an electron image
projection device employing the present invention;
FIG. 15 is a fragmentary cross-sectional view in elevation taken
along line XV--XV of FIG. 14;
FIG. 16 is a fragmentary cross-sectional view in perspective taken
along line XVI--XVI of FIG. 15; and
FIG. 17 is a block diagram of an electrical circuit for the
electron image projection device shown in FIG. 14 to automatically
align the electron beam pattern in accordance with the present
invention.
DESCRIPTION OF THE PREFERRED EMBODIMENTS
Referring to FIG. 1, substrate 10 having opposed major surfaces 11
and 12 is selected of a material transmissive to the radiation of
interest. For making a photocathode source as hereinafter
described, substrate 10 is typically quartz because it is
substantially transparent to ultraviolet light radiation and
relatively inexpensive. Wafers of sapphire or lithium fluoride may
also be used in lieu of quartz for the substrate.
A regular array of marks of cathodoluminescent material is formed
adjacent major surface by first applying to major surface 11 an
electrosresist layer 13. An electroresist is a material which is
sensitive to electron radiation to form in a layer thereof a
pattern of differential solubility. That is, the electron
irradiated portion of the layer is made either more or less soluble
to certain solvents than the non-irradiated portion of the layer.
Preferably the electroresist layer is light insensitive and
relatively stable and has a relatively long shelf life. Examples of
negative resists are polystyrene, polyacrylamide, polyvinyl
chloride and certain selected hydrocarbon silicones. Examples of
positive resists are polyisobutylene, polymethylmethacrylates, and
poly(alphamethylstyrene).
A good positive resist is polymethylmethacrylate of an average
molecular weight of over 100,000 containing a very low fraction of
polymer having molecular weight of 50,000 or less to avoid pin
holes during processing. It is rendered readily soluble in either
95 % ethanol (5 % water) or in a mixture of 30 % by volume of
methylethyl ketone and 70 % isopropanol when subjected to an
electron beam at 10 kilovolts to supply 5 .times. 10.sup.-.sup.5
coulombs per cm.sup.2. The portions so exposed are soluble in the
previously mentioned solvent, whereas the remainder of the resist
coating is not as soluble in such solvent.
Polyacrylamide is a good negative resist inasmuch as the electron
beam at 10 kilovolts applying 3 .times. 10.sup.-.sup.6 coulombs per
cm.sup.2 will render it slowly soluble in deionized water, while
the remainder of the coating will resist concentrated phosphoric
acid, which renders it useful as a coating for aluminum. This
electroresist is not removed by most organic solvents such as
methanol. It forms an excellent mask for a sputtering-etch
treatment of the substrate. The average molecular weight of a good
polyacrylamide resist that has given good results is 4.5 .times.
10.sup.7.
The electroresist may also be provided by any of the various
commercially available "photoresist" materials that are sensitive
to electron bombardment to become more soluble or insoluble in
specified solvents. Three such photoresists are AZ-1350 and
AZ-1350H made by Shipley and Microline PR-102 made by GAF.
The electroresist may also be one of various inorganic compounds as
well as organic compounds. For example, silicon dioxide and silicon
nitride (Si.sub.3 N.sub.4) coatings on a substrate when subjected
to an electron beam are rendered more soluble in an etchant.
Buffered hydrofluoric acid will dissolve more readily the electron
beam treated portions of a silicon dioxide layer as compared to the
portion of the layer not treated with the electron beam. This
characteristic is known as the "BEER" effect (i.e. Bombardment
Enhanced Etch Rate). Etch enhancement ratios of about 3 are
obtained, so that the electron beam bombarded portions will be
completely etched away while there will be only as little as a
third of the unbombarded layer that will be etched away.
For example, silicon substrates with an oxide layer of a thickness
of about 10,900 A has areas electron irradiated to a total of 0.5
coulombs per cm.sup.2 and then etched electrolytically with P etch
at 6 volts with the silicon being anodic to yield etch rates of 142
A per minute for the unirradiated areas and 416 A per minute for
the electron irradiated areas. When the same silicon substrate was
made cathodic at 6 volts, all other procedures remaining the same,
the etch rates were 134 A per minute for the unirradiated portions
and 391 A per minute for the electron irradiated portions. Again,
when the silicon substrate was made cathodic at 15 volts, the
relative etch rates were 153 A per minute for the unirradiated
portions and 460 A per minute for the electron irradiated portions.
It should be noted for this example that the etchant employed was a
buffered solution of hydrofluoric acid (pH 6.5) which is an aqueous
solution 10 M in NH.sub.4 F and 2.62 M in HF. It was also found
that the electron beam can be supplied as little as 0.25 to 1
coulomb per cm.sup.2 to secure enhanced etching in silicon dioxide
layers of 10,000 A in thickness; the unirradiated portions were
etched to 7000 to 8000 A in thickness while the 0.5 and 1.0 coulomb
per cm.sup.2 portions portion were completely etched through.
Irrespective of the composition, the thickness of electroresist
layer 13 is also important to the definition of the pattern formed
in it. The thickness of the resist layer 13 must be on the order of
the resolution desired in the component pattern. Typically, the
thickness will be between about 0.2 and 1.0 microns. If the desired
resolution is 0.1 micron, then the electroresist layer need be on
the order of 0.5 micron or less.
Still referring to FIG. 1, a predetermined pattern of windows or
marks 14 of predetermined shapes corresponding to the desired array
of cathodoluminescent marks are formed in electroresist layer 13.
The electroresist layer 13 is irradiated by a scanning electron
beam (not shown) of fine dimensions, e.g. 0.2 micron in diameter,
to implant the mark pattern in differential solubility in layer 13.
The position of the electron beam is sequentially moved on command
from a computer over the electroresist layer to selectively
irradiate and define the desired pattern in the electroresist, see
above-cited U.S. Pat. No. 3,679,497 for further description. After
irradiation, the electroresist layer 13 is developed to open
windows 14 and expose selected portions of major surface 11 of
substrate 10. The solvent used for this purpose will depend upon
the composition of the electroresist layer as noted above.
After forming the predetermined pattern of marks in layer 13, the
pattern corresponding to the desired array and predetermined shapes
of the cathodoluminescent marks is transferred to a pattern in the
substrate 10 by etching well pattern 15 in substrate 10 through
windows 14. The etchant suitable for this purpose will depend upon
the composition of the substrate 10. For quartz substrates a
solution of hydrofluoric acid is a suitable etchant for this
purpose. The depth of well pattern 15 is determined by the etching
time for the particular etchant composition used. Typically the
etching time is sufficient to form a well pattern of at least about
600 A in depth in the substrate 10.
Referring to FIG. 2, the remainder of electroresist layer 12 is
removed with a suitable solvent, and layer 16 of cathodoluminescent
material is formed over surface 11 by standard RF sputtering or
vapor deposition techniques, filling well pattern 15. The
particular cathodoluminescent material will depend on the
particular use of the invention. For use of the substrate 10 with
the regular array of closely spaced cathodoluminescent marks of
predetermined shapes, as shown, directly as a substrate for an
integrated circuit, for example, any cathodoluminescent material
can be used depending on the sensitivity of the electron beam and
the photodetector as hereinafter described. For use in making a
photocathode source, the cathodoluminescent material is preferably
a material, such as calcium fluoride activated with manganese (i.e.
CaF.sub.2 :Mn), which is substantially transparent to ultraviolet
light and the like. No difficulty is thus encountered in having the
cathodoluminescent marks reproduce as part of the patterned
electron beam projected by the photocathode source.
Following disposition of layer 16, layer 16 is polished or possibly
lap etched to remove the cathodoluminescent material of surface 11
while leaving the cathodoluminescent material in well pattern 15
forming cathodoluminescent marks 17 in a regular array and of
predetermined shape (e.g. square as shown). A planar surface is
thus provided with marks 17 in position. The substrate can in this
form be used in accordance with the present invention for making an
integrated circuit or other microminiature electronic components as
hereinafter described.
The pattern of cathodoluminescent marks 17 of predetermined shapes
may be any regular array as desired. Preferably the pattern is
symmetrical to facilitate alignment of a scanning electron beam
from field to field as hereinafter described. Referring to FIGS. 3
through 5, some typical patterns of marks 17 are shown. The marks
may form a single parallel-line pattern as shown in FIG. 3, a
line-grid as shown in FIG. 4, or spaced rows as shown in FIG. 5. In
any case, the shapes of marks 17 and spacing between marks 17 are
selected to provide the desired alignment accuracy, and preferably
the size, i.e. width or area, of marks 17 are selected to
correspond to the desired alignment accuracy. Typically, the marks
will have a width (or diameter) less than 0.5 micron and a spacing
center-to-center of between 200 and 2000 microns. The alignment
system resulting has a usual accuracy of 0.5 micron or less. In
this connection, it should be noted that equivalent
cathodoluminescent marks 17 can be obtained by forming the
cathodoluminescent material in the negative of the marks 17. That
is, cathodoluminescent layers are formed to circumscribe exposed
portions of the surface of the members which are in the
predetermined shape of the desired cathodoluminescent marks.
Referring to FIGS. 6 and 7, substrate 10 with cathodoluminescent
marks 17 adjacent surface 11 is made into a photocathode source. A
contiguous layer 18 opaque to ultraviolet light is deposited over
major surface 11 and marks 17. Layer 18 may be either reflective or
absorptive of ultraviolet light. For an absorptive layer, layer 18
can be a titanium ion containing material in the form of an oxide
including titanium dioxide, or other materials such as
Fe.sup.+.sup.3 containing materials. When an oxide of titanium is
used, layer 18 may be made by reactively sputtering titanium in an
oxidizing atmosphere or by RF sputtering of titanium dioxide
itself. Alternatively, a layer of titanium may be vapor deposited
and then wholly or partially oxidized in situ to provide opaque
layer 18. For a reflective opaque layer 18, a layer of aluminum or
chromium of thickness of about 800 A may be deposited on exposed
portions of surface 11 and marks 17 by standard sputtering
techniques. In any case, a protective layer (not shown) of, for
example, quartz or other glass having a thickness of about 1500 A
may be formed over layer 18 to prevent subsequent oxidation and/or
permit periodic refitting of the photocathode source with a new
photocathode material.
Overlayer 18 is then deposited a contiguous layer 19 of
electroresist, and the positive or negative of a desired component
pattern is implanted in electroresist layer 19 in differential
solubility by irradiating the layer with a scanning electron beam
or the electron image projection system, see U.S. Pat. No.
3,679,497. The desired component pattern is then formed in
electroresist layer 19 to expose selected surface portions of
opaque layer 18 by dissolving the more soluble portion of layer 19
in a suitable solvent. The desired component pattern is then
transferred to the opaque layer 18 by etching or ion milling opaque
layer 18 through window pattern 20 in electroresist 19. A suitable
etchant for this purpose is dilute hydrofluoric acid.
Referring to FIG. 7, remaining portions of electroresist layer 19
are then removed with a suitable solvent such as trichloroethylene
or acetone and a contiguous layer 22 of photocathode material is
then deposited over exposed surface portions of surface 11,
cathodoluminescent marks 17 and opaque layer 18. Preferably the
photocathode material is an air stable material such as palladium,
gold, platinum, aluminum, barium, copper or cesium iodide.
Typically such photocathode material is deposited by vacuum
vaporization or RF sputtering. The resulting combination of layers
18 and 19 thus provided a photocathode source adapted to project an
electron beam in the predetermined component pattern on irradiation
by sensitizing light radiation such as ultraviolet light.
The result is a finished photocathode source shown in FIG. 7 which
can be used in accordance with the invention as hereinafter
described. Because of the recess of the cathodoluminescent marks
into the substrate, a photocathode source embodying the present
invention can be made without optical distortions arising from
non-planar surfaces. In this connection, the importance of the
cathodoluminescent marks 17 being substantially transparent to
ultraviolet light is emphasized to avoid unnecessary difficulties
in the design and manufacture of the photocathode source. It shoule
also be noted that the resulting photocathode provides a negative
projection. That is, it is designed to generate a patterned
electron beam which is the negative of the desired component
pattern. It is designed for use in the electron image projection
system with negative electroresists.
Referring to FIGS. 8 through 10, an alternative procedure is shown
for making a photocathode source embodying the invention. Substrate
30 having opposed major surfaces 31 and 32 is selected of a
composition typically on the same considerations as substrate
10.
Over surface 31 is formed a contiguous layer 33 of material opaque
to light radiation emitted by the cathodoluminescent material used
in the photocathode source. Preferably opaque layer 33 is composed
of a metal such as titanium.
Over opaque layer 33 is formed a contiguous layer 34 of
electroresist. The desired pattern array of cathodoluminescent
marks or the negative thereof, depending on whether the
electroresist is positive or negative, is thereafter implanted in
the electroresist layer as a pattern of differential solubility by
use of a scanning electron beam or the electron image projection
system as above described and referenced. The electroresist is then
developed with a suitable solvent, depending on the resist
composition, to form window pattern 35 corresponding to the desired
array of cathodoluminescent marks to expose selected portions of
opaque layer 33.
Window pattern 36 is then opened in opaque layer 33 corresponding
to the desired array of cathodoluminescent marks by selectively
etching or ion milling through the window pattern 35 in
electroresist layer 34. Dilute hydrofluoric acid is a suitable
etchant for this purpose where titanium is used for the opaque
layer. The pattern corresponding to the desired array of
cathodoluminescent marks of predetermined shapes is thus formed in
opaque layer 33. The remainder of electroresist layer 34 is then
removed with a suitable solvent such as trichloroethylene or
acetone.
Over the exposed portions of surfaces 31 and opaque layer 33 is
thereafter applied a contiguous layer 37 of catholuminescent
material of interest to provide the cathodoluminescent mark of
predetermined shape. Preferably layer 37 is applied by standard RF
sputtering or vacuum evaporation deposition techniques. The
cathodoluminescent material is preferably of a material such as
CaF.sub.2 :Mn.
It should be noted at this point in connection with FIGS. 8 and 9
that equivalent cathodoluminescent marks of predetermined shapes
sofaras the alignment system is concerned can be formed by forming
the opaque layer with the desired openings for the marks over the
contiguous cathodoluminescent layer instead of under the
cathodoluminescent layer. The opaque layer in this embodiment may
either reduce the electron radiation reaching the
cathodoluminescent layer or reduce the cathodoluminescent radiation
projected by the cathodoluminescent layer to a detector means
positioned above the major surface. Alternatively, the opaque layer
may be in the predetermined shape above or below the
cathodoluminescent layer to reduce the electron radiation reaching
the cathodoluminescent layer or the cathodoluminescent radiation
projected by the cathodoluminescent layer. It should also be
observed that "opaque" does not mean that the layer totally absorbs
or reflects the cathodoluminescent or electron radiation. The
"opaque" layer need only absorb or reflect sufficient light or
electron radiation to provide a discernible differential in
cathodoluminescence produced. The opaque layer may even itself be
cathodoluminescent provided it generates radiation which is
discernibly different from the cathodoluminescent radiation
projected by the cathodoluminescent marks.
Irrespective of the embodiment, the substrate 30 in this form, with
layers 33 and 37 in place, can be used for the substrate directly
in the making of an integrated circuit or other micro-miniature
electronic components as hereinafter described. Preferably,
however, the substrate is further processed to form a photocathode
source by forming opaque layer 38 over layer 37. Opaque layer 38 is
either absorptive or reflective to ultraviolet or other radiation
to which the desired photocathode material is sensitive. Preferably
opaque layer 38 is made of the same material and deposited in the
same manner as opaque layer 18 above described.
Over opaque layer 38 is then applied electroresist layer 39. The
desired component pattern or the negative thereof, depending on
whether the resist is positive or negative, is thereafter implanted
in the electroresist layer as a pattern in differential solubility
by irradiating the layer with wlectron radiation. Preferably a
scanning electron beam or the electron image projection system as
above described and referenced is used to provide the electron
radiation. The selectively irradiated electro-resist is then
developed with a suitable solvent, depending on the resist
composition, to form window pattern 40 in layer 39 corresponding to
the positive of the desired component pattern. Thereafter, the
desired electronic component pattern is transferred to opaque layer
38 by selectively etching or ion milling window pattern 41 in
layers 33, 37 and 38 through window pattern 40 to expose portions
of surface 31. A suitable etchant for this purpose is dilute
hydrofluoric acid.
Referring to FIG. 10, remaining portions of electroresist layer 39
are then removed with a suitable solvent such as trichloroethylene
and a contiguous layer 42 of photocathode material is deposited
over exposed portions of layers 38 and surface 31 to finish the
photocathode source. Preferably the same photocathode material
deposition techniques are used for photocathode layer 42 as above
described for photocathode layer 22. It should be noted that the
finished photocathode source provides a positive projection. That
is, the electron beam pattern generated by the source is the
positive of the desired electronic component. The source is thus
designed for use in the electron image projection system with
positive electroresists.
Referring to FIGS. 11, 12 and 13, a method is shown for producing a
highly accurate pattern in an electroresist on a substrate by
aligning an adaptable beam relative to a member. A scanning
electron microscope adatable to the present invention is shown in
above-cited U.S. Pat. No. 3,679,497, assigned to the same assignee
as the present invention.
The member used is preferably a substrate prepared as shown in
FIGS. 3, 4 or 5, or alternatively as shown in FIG. 9 without layers
38 and 39 applied. For convenience of explanation a member is
prepared as shown and described in FIGS. 3, 4 or 5 is shown in FIG.
11 in use with a scanning electron beam. Alternatively the member
may be a semiconductor wafer or a suitable insulator or
semiconductor substrate with an epitaxially grown layer
thereon.
Referring specifically to FIG. 11, substrate 10 is provided as
above described with a closely spaced regular array of
cathodoluminescent marks 17 of predetermined shape formed adjacent
major surface 11. A layer 50 of a desired material such as a metal,
an insulator or a semiconductor material in which a component
pattern is desired is formed over exposed portions of surface 11
and marks 17 by sputtering, vapor deposition, epitaxial growth or
any other suitable technique. Thereafter a contiguous electroresist
layer 51 is overlaid on layer 50 over major surface 11 and marks
17. A series of photodetector means 53 are positioned in holder 57
adjacent the opposite surface 12 of the member to circumscribe the
cathodoluminescent marks 17. That is, each photodetector 53 is
positioned adjacent a mark 17 to detect cathodoluminescence
generated by irradiation of said mark 17. This arrangement of
course supposes that the member is substantially transparent to the
cathodoluminescent emissions; if the member is not able to transmit
the cathodoluminescence, the photodetectors may be positioned on
the same side of the member as the electron beam and close to the
marks as physically possible to avoid loss of resolution by
radiation scattering.
Referring specifically to FIGS. 12 and 13, this arrangement can be
used to align the scanning electron beam with the major surface 11
of the member 10 field by contiguous field for selective
irradiation of precisely selected areas of the major surface of the
substrate. As shown in FIG. 12, the member 10 is divided into
contiguous fields preferably bounded symmetrically in quadrature by
the marks 17, e.g. one at the intersection of each field, or one at
the center along each side of each field. To align the beam 52
with, for example, field 54, the scanning electron beam 52 is
modulated to overlap two opposite marks, for example marks 17.sub.1
and 17.sub.2 sequentially. The output signals from the detectors
53, positioned adjacent marks 17.sub.1 and 17.sub.2, respectively,
are fed to a cathode-ray tube 55 which also has the intended
locations of the marks for precise alignment inputted from the
computer 56 controlling the scanning electron beam 52. The scanning
beam 52 is thus moved relative to the member 10 until the signals
from the two detectors 53.sub.1 and 53.sub.2 indicate optimum
alignment of beam and marks and also alignment of the superimposed
input of the intended locations of the marks 17.sub.1 and 17.sub.2.
The electron beam 52 is then modulated to overlap the other two
opposite marks 17.sub.3 and 17.sub.4 sequentially until the outputs
therefrom coincide on the CRT 55 with optimum alignment of the beam
and marks and also alignment of the superimposed theoretical input
for marks 17.sub.3 and 17.sub.4 from the computer 56. The electron
beam 52 is then aligned and ready to selectively irradiate the
field on command from the computer 56. At the end of the
irradiation of the field 54, the member 10 is physically moved so
that the scan field of the electron beam is concurrent with the
next field 54' on the member 10 to be selectively irradiated. The
aligning sequence is then repeated as described above. At each
field 54, alignment is made by manually or automatically actuating
means to move member 10 relative to beam 52 as described in
above-cited U.S. Pat. No. 3,679,497, or to move beam 52 relative to
member 10 by deflecting the electron beam electromagnetically. The
alignment is accomplished by moving the beam or member relative to
the other until the light radiation emissions detected by
photodetector 53 indicates a predetermined alignment value for said
radiation emissions.
It should be noted that said predetermined alignment value will
vary with the relative relation of the predetermined shape of the
cathodoluminescent marks 17 and the predetermined cross-sectional
shape of the electron beam 52. Where the predetermined shapes are
the same, it is simply a matter of reading the indication of
maximum or minimum cathodoluminescence. Where the predetermined
cross-sectional shape of the electron beam 52 is different from the
predetermined shapes of the corresponding marks 17, the reading to
determine alignment is somewhat different. Optimum alignment is no
longer indicated by the maximum or minimum in the signal readings
for cathodoluminescence. Rather, plateaus are reached in the signal
readings, and optimum alignment is achieved by either selecting a
certain point on each plateau taking into consideration any change
in the geometric shapes of the electron beam and the marks, or
selecting a certain point on the signal rise from the means as the
electron beam moves into or out of the area of the corresponding
mark. The latter alignment sequence permits alignment with the edge
of the mark.
The alignment system shown in FIG. 13 is what one skilled in the
art would connote a manual system because an operator makes the
adjustments to align in accord with the read-out on the CRT. This
is not a preferred system because of the relatively long length of
time required to complete the alignment sequence. Thus, it is that
the alignment sequence is preferably adapted to an automatic
alignment system with electrical signal processing apparatus such
as that hereinafter described.
After alignment, the electron beam 52 is moved at each field 54
through a predetermined matrix corresponding to a desired component
pattern to be formed in or through layer 50, or the negative
thereof. Whether beam 52 irradiates the positive or the negative of
the desired component pattern is dependent on the type of
electroresist used. If a positive resist is used, i.e. where the
irradiation makes the resist layer more soluble in a given solvent,
then the positive of the desired electronic component pattern to be
formed in layer 50 is irradiated in layer 51. Conversely, if a
negative resist is used, i.e. where irradiation makes the resist
less soluble to a given solvent, then the negative of the desired
component pattern to be formed in layer 50 is irradiated in layer
51.
In either instance, the accuracy of the system is primarily based
on the size of fields 54 in which layer 51 is irradiated nd the
sizing and spacing of the marks 17. For example, if the integrated
circuit to be formed on the member if 0.080 inch .times. 0.080 inch
and an accuracy of less than 1 micron is desired, the electron scan
must be performed in four fields of 0.040 inch .times. 0.040 inch
in size. Or, if an accuracy of less than 0.20 micron is desired,
the same member is preferably scanned in 400 fields of 200 .times.
200 microns.
The marks 17 may be dimensioned so that the detection is totally
qualitative. That is, the marks are sized, e.g. 0.5 .times. 0.5
microns or less (spaced, e.g. 200 .+-. 0.2 microns apart) so that
whenever any light is detected by photodetector 53 the position of
electron beam 52 relative to member 10 is known within a tolerance
of 0.2 micron. Alternatively, marks 17 may be larger and the light
emissions measured quantatively so that the position of electron
beam 52 relative to substrate 10 is accurately known. In any event,
the marks 17 are of predetermined shapes preferably the same as the
cross-sectional shape of the electron beam. The resolution errors
rising from electron beam deflections and distortions are reduced
to the desired accuracy with the alignment system. Scanning
electron beam 52 can thus be aligned with a member with greater
accuracy than in prior alignment systems, see above-cited U.S. Pat.
No. 3,679,497.
Referring specifically to FIGS. 14, 15 and 16, the present
invention is used to produce a highly accurate pattern in an
electroresist layer on a member in an electron image projection
system. The photocathode source 43 is as shown in FIG. 10 and
described in connection therewith, except that the
cathodoluminescent layer is localized and two cathodoluminescent
marks 61 and 61' of predetermined shapes are provided diametrically
opposite each other on the periphery of member 60. The patterned
electron beam produced by the photocathode source including a set
of alignment beam portions of predetermined cross-sectional shapes
and corresponding substantially in spacing to cathodoluminescent
marks 61 and 61' positioned on member 60. In this connection, it
should be noted that marks 61 and 61' are in reality openings in
layer 62 which are opaque to light emissions from contiguous
localized layer 63 of cathodoluminescent material overlaying and
adjoining layer 62. Thus, the pattern of light emissions from the
cathodoluminescent layer 63 corresponds to the openings in opaque
layer 62 and provide the equivalent of a closely regular array of
cathodoluminescent marks as described above.
Member 60 with layers 62 and 63 in place is prepared for the
electron image projection system by applying or epitaxially growing
a layer 64 of suitable material such as a metal or semiconductor
material in which or through which a portion of the integrated
circuit is to be formed in or to major surface 65 of member 60.
Thereafter, electroresist layer 66 in which the desired highly
accurate component pattern is to be formed is applied over layer
64.
And with electroresist layer 66 applied, member 60 is inserted into
the electron image projection system as shown and described in
FIGS. 14, 15 and 16 and more fully described in above-cited U.S.
Pat. No. 3,679,497 in substantially parallel relationship with
photocathode source 43 also positioned in the system.
Referring to FIG. 14, an electron image projection device is shown.
A hermetically sealed chamber 70 of nonmagnetic material has
removable end caps 71 and 72 to allow for disposition of apparatus
into and removal of apparatus from the chamber. A vacuum port 73 is
also provided in the sidewall of chamber 70 to enable a partial
vacuum to be established in the chamber after it is hermetically
sealed.
Disposed within chamber 70 is cylindrical photocathode source or
electromask 74 alignable with selected areas of the major surface
of member 60 in substantially parallel, spaced relation. Member 60
is supported in specimen holder 75 as more fully described
hereinafter. Photocathode 74 and holder 75 are in turn positioned
in parallel array by annular disk-shaped supports 76 and 77,
respectively. Photocathode source 74 and holder 75 are spaced apart
with precision by tubular spacer 78 which engages grooved flanged
79 and 80 via gaskets 81 and 82 around the periphery of supports 76
and 77. The entire assembly is supported from end cap 71 of chamber
70 at support 76 to allow for ease of disposition of the
photocathode source and the member within the chamber.
Photocathode 74 is made cathodic and member 60 is made anodic to
direct and accelerate electrons projected from the photocathode to
the member 60. To accomplish this, holder 75 and supports 76 and 77
are of highly conductive material and spacer 78 is of highly
insulating material. A potential source 78A of, for example, -10Kv,
is applied between supports 76 and 77. The difference in potential
is conducted to and impressed on photocathode 74 and member 60 via
supports 76 and 77 and holder 75.
Surrounding chamber 70 are three series of electromagnetic coils,
positioned perpendicular to each other, to direct and focus the
impingement of the electron beam onto member 60. Cylindrical
electromagnetic coils 83.sub.1, 83.sub.2 and 83.sub.3 are
positioned axially along the path of the electron beam from
photocathode 74 to member 60 to cause electrons to spiral and move
radially as they travel the distance from the photocathode to the
member. These coils permit control of the rotation (.theta.) and
the magnification (M) of a patterned electron beam emitted from the
photocathode source and in turn focusing of the patterned electron
beam. Rectangular electromagnetic coils 84.sub.1 and 84.sub.2, and
85.sub.1 and 85.sub.2 are symmetrically positioned in Helmholtz
pairs perpendicular to each other, and to coils 83.sub.1 -83.sub.3,
to cause electrons to transversely deflect as they travel the
distance from the photocathode to the member. These electromagnets
permit control of the direction (in X and Y coordinates) of a
patterned electron beam emitted from the photocathode.
In operation, light source 86 such as a mercury vapor lamp backed
by reflector 87 irradiates a photocathode layer 88 (e.g. gold or
palladium) in the photocathode source or electromask 74. The
photocathode layer is irradiated through a substantially
transparent substrate 89 such as quartz overlaid with a layer 90
containing the negative of a desired component pattern. The layer
90 is of material (e.g. titanium dioxide) which is opaque to the
light radiation. The photocathode material is thus made electron
emissive in a patterned electron beam corresponding to the desired
component pattern to be formed in layer 64 or the negative thereof
depending on whether the electroresist is positive or negative. A
part of the patterned electron beam emitted from the photocathode
source 74 is at least one and preferably two relatively small
alignment beam portions 91 and 92 of predetermined cross-sectional
shape (e.g. squares of 300 .times. 300 microns) which are spaced
away preferably oppositely positioned along the periphery of the
patterned beam from the photocathode.
Referring to FIG. 15, member 60 is precision mounted within
physically permissible limits in holder 75 and in turn with respect
to photocathode 74. Member 60 has a flat peripheral portion 93; and
holder 75 has depression 94 into which substrate 60 fits. Holder 75
has pins 95, 96, 97 and 98 positioned in respective quadrants
around the periphery of depression 94. Member 60 is positioned by
resting flat peripheral portion 93 of member 60 against pins 95 and
96 and curvilinear peripheral portion 99 of member 60 against pin
97. The member is thereby located with an accuracy of about 25
microns or less. Movable pin 98, which is fitted with a compression
spring 100, is positioned and pushed against the curvilinear
portion of member 60 to firmly retain member 60 and in turn,
maintain member 60 precisely located.
As shown by FIG. 14, cathodoluminescent marks 61 and 61' of
predetermined shapes are formed on member 60 as above described.
Preferably marks 61 and 61' are widely spaced apart along the
periphery of the member, and preferably have predetermined shapes
the same as the predetermined shape of alignment beam portions 91
and 92 (see FIG. 16), which form a part of the patterned electron
beam emitted by the irradiated photocathode source 74. Marks 61 and
61' of predetermined shapes can be reasonably precisely formed by
selectively etching or ion milling the member through window
patterns of the desired cross-section in a photo- or electroresist
layer (as above described in connection with FIGS. 8 through 10).
The predetermined shapes of the alignment beam portions and marks
are thus preferably about 10 .times. 10 mils in any suitable
geometric shape, such as a square, rectangle or circle.
Positioned behind marks 61 and 61' of predetermined shapes in
holder 75 are photodetector means 101 and 102, respectively, with
leads 103 and 104 and 103A and 104A, respectively, extending
through vacuum seals in chamber 70 at 105. Photodetectors 101 and
102 are adapted for detecting the cathodoluminescent radiation
produced by the marks 61 and 61', which is usually light radiation,
through member 60, and are substantially larger in size than and
circumscribe marks 61 and 61' so that they can detect even
dispersed and scattered cathodoluminescent radiation from member 60
close to or in the vicinity of the marks. For this reason, detector
means 101 and 102 are also positioned in as close a proximity to
member 60 as the geometry will permit. In this connection, it
should be noted that in some embodiments, e.g. where member 60 is
not transmissive of the cathodoluminescent radiation, it may be
appropriate to position the detector means 101 and 102 on the same
side of member 60 as the photocathode source so that
cathodoluminescent radiation is detected. However, it is preferred
that the detector means are positioned in as close a proximity to
marks 61 and 61' as the conditions will permit so that resolution
of the light radiation and in turn accuracy of alignment is not
lost between the cathodoluminescent marks and the detector means.
It is preferred therefore that the detector means be positioned
opposite member 60 from the photocathode source where possible so
that resolution and accuracy are not lost.
In operation, the alignment beam portions 91 and 92 of
predetermined cross-sectional shapes impinge on and overlap the
marks 61 and 61' of predetermined shapes, respectively. The
electron beams typically produce a cathodoluminescent radiation
which corresponds to the degree of overlap between the alignment
beam portions and the marks. Alignment can be accurately recorded
therefore simply by observing the signals at detector means 101 and
102. The alignment beam portions and cathodoluminescent marks can
thus be brought into precise alignment simply by detecting as the
predetermined alignment value where the detected
cathodoluminescence indicates the optimum alignment of the
alignment beam portions with the marks.
The present invention thus provides a method of alignment of the
patterned electron beam generated by the photocathode source 74
with precisely located areas of a major surface of the member 60.
Further, the member may be similarly aligned with successive
photocathode sources or electromasks by use of the same
cathodoluminescent marks and like alignment beam portions on the
successive photocathodes so that all of the patterned electron
beams will selectively impinge on the member with the desired
precision exactness, e.g. within a fraction of a micron. Error is
reduced to the precision with which the marks and the photocathode
sources emitting the corresonding alignment beam portion can be
shaped and spatially located, which presents no difficulty with the
scanning electron microscope and electron image projection
system.
The electric current flow from the detector means 101 and 102 may
be processed through suitable electronic amplifiers and
servomechanisms to automatically shift the entire patterned
electron beam relative to the member and precisely position the
alignment beam portions 91 and 92 with the marks 61 and 61',
respectively. For this purpose, suitable means such as a modulation
means is preferably used to oscillate the electrical input to the
electromagnetic means or coils and thereby cause the alignment beam
portions 91 and 92 to oscillate or move in typically a circle over
the marks 61 and 61' so that the electrical outputs from detectors
101 and 102 are modulated.
Referring to FIG. 17, there is illustrated in a block diagram the
electronics for adjusting the alignment beam portions 91 and 92
with respect to the marks 61 and 61' of the same predetermined
shapes and in turn precisely aligning selected areas of the major
surface of member 60 with respect to the entire electron beam
pattern from photocathode source 74. The modulated electrical
signal from detector 101 is conveyed via lead 104 to a preamplifier
106, which amplified signal is then conveyed via lead 107 to a
tuned amplifier 108. The output of amplifier 108 passes through
lead 109 to a phase adjustor 110 and then through lead 111 to a
dual phase detector 112. A gated oscillator 113 impresses reference
signals, which are 90.degree. out-of-phase, through conductors 114
and 115 and conductors 116 and 117, respectively, on the dual phase
detector 112. The outputs of phase detector 112 thus comprise
X-error signals via lead 118 and Y-error signals via lead 119,
which pass through gate 120 via leads 121 and 122 to integrators
123 and 124, respectively. The integrators 123 and 124 have
direct-current outputs to adders 125 and 126, respectively, where
the outputs are modulated with alternating current from the
oscillator 113 via leads 127 and 128, respectively. The added
modulated signals are then passed to and adjacent the controls in
power units (not shown) of the type customarily used to power the
electromagnetic coils, in this case the electromagnetic coils
84.sub.1, 84.sub.2, 85.sub.1 and 85.sub.2 in Helmholtz pairs.
Similarly, the modulated signal from the detector 102 is conducted
via lead 104A to preamplifier 129, and passed thereafter via lead
130 to the tuned amplifier 131 and then via lead 132 through phase
adjustor 133 and lead 134 to dual phase detector 135. Oscillator
113 suppresses the two 90.degree. out-of-phase reference signals,
above referred to, through leads 136 and 137 on dual phase detector
135. Two outputs from dual phase detector 135 are thus produced.
The one output signal via conductor 138, which corresponds to a
.theta. error signal, passes through gate 140 and lead 141 to
control a motor-driven precision potentiometer 142 to effect the
rotational control of the electron beam pattern by increasing or
decreasing the current to the electromagnetic coils 83.sub.1,
83.sub.2 and 83.sub.3. The other output signal via lead 139, gate
140 and lead 143 to a motor driven gang potentiometer 144 which
adjusts the main focus field controls the size of the patterned
electron beam.
The error signals in conductors 118, 119, 138 and 139 are cross-fed
electronically via leads 145, 146, 147 and 148, respectively, into
a four input delayed null detector 149 whose output is conveyed by
lead 150 to a set-reset flip-flop 151. The operation of the
flip-flop is initiated by actuation of a start sequence switch,
whereupon current begins to flow via leads 152 and 153 to energize
the utlraviolet source 86 to cause electron beams to be emitted
from photocathode source 74, including the two alignment beam
portions 91 and 92 of predetermined cross-sectional shapes.
Likewise, current from 152 passes through lead 154 to the gated
oscillator 113, which in turn feeds sinusoidal signals in
quadrature through lines 114-127 and 116-128 to the X and Y
controls 125 and 126, respectively. The entire electron beam
pattern, including the alignment portions 91 and 92, are thus
caused to oscillate typically in a circle of, for example, 6
microns diameter at a frequency of 45 Hertz.
Once the alignment portions 91 and 92 are substantially aligned
with marks 61 and 61', respectively, by operation of integrators
123 and 124, and potentiometers 142 and 144, the error signals
passing through leads 145, 146, 147 and 148 reach a zero value
which is detected by the null detector 149. The null detector
thereupon produces an electrical signal which passes through lead
150 to the flip-flop 151, which terminates the operation of the
gated oscillator 113 and closes gates 120 and 139 by signals
through leads 155 and 156, respectively. The time sequence of the
selective electron beam exposure of an electroresist layer on the
full area of member 60 is then begun and continued until the resist
is fully exposed.
The detection and alignment is accomplished in a time substantially
less than the time period required to significantly irradiate the
electroresist layer 66. A period of from 3 to 10 seconds is usually
adequate to produce a sufficient electron beam treatment of the
electron resist to cause it to be properly differentially soluble
in selected solvents. In turn, the desired electronic component
pattern of the electromask 43 is transferred to electroresist layer
66 without alignment distortions. Further, it should be noted that
if the cathodoluminescent marks 61 are closely and regularly spaced
the alignment can be accomplished with simply translational, i.e.,
x and y, corrections, without the need for rotational correction of
previous alignment systems, see U.S. Pat. No. 3,710,101 and U.S.
Pat. Application Ser. No. 264,699, filed June 20, 1972 and assigned
to the same assignee as the present application.
Further, it should be noted that the alignment can be made by
manually or automatically controlled by the physical movement of
the member 60 and/or electromagnetic deflection of the electron
beam pattern, see above-cited U.S. Pat. Nos. 3,679,497 and
3,710,101. In any case, the correction is continued until the light
emissions from marks 61 detected at photodector means 101 and 102
indicate a predetermined alignment value for said radiation
emissions.
While the present invention is particularly suited and has been
specifically described to align a scanning electron beam or an
electron image projection system with a member in the accurate
transfer of a desired electronic component pattern to an
electroresist layer supported by a member, it is distinctly
understood that the invention may be otherwise variously embodied
and used within the scope of the following claims. For example, the
invention may be used in the procedure for precision etching of
selected areas of metal sheets to obtain desired shapes and
patterns for various scientific and industrial applications.
* * * * *