U.S. patent number 3,721,584 [Application Number 05/027,744] was granted by the patent office on 1973-03-20 for silicon coated substrates and objects fabricated therefrom.
Invention is credited to Albert R. Diem.
| United States Patent |
3,721,584 |
| Diem |
March 20, 1973 |
SILICON COATED SUBSTRATES AND OBJECTS FABRICATED THEREFROM
Abstract
Silicon coated substrates for such uses as optical filters,
photo masks, passive circuits and as an information storage
means.
|
Inventors: |
Diem; Albert R. (Salt Lake
City, UT) |
| Family
ID: |
21839540 |
| Appl.
No.: |
05/027,744 |
| Filed: |
April 13, 1970 |
| Current U.S.
Class: |
359/580;
257/E23.144; 257/E21.266; 359/360; 428/428; 359/888; 428/334;
428/446 |
| Current CPC
Class: |
H01C
17/003 (20130101); C23C 16/24 (20130101); H01L
27/00 (20130101); G02B 5/285 (20130101); H01L
21/00 (20130101); H01L 23/5222 (20130101); H01C
7/00 (20130101); G03F 1/54 (20130101); H01L
21/314 (20130101); G11C 13/048 (20130101); H01L
49/02 (20130101); Y10T 428/263 (20150115); H01L
2924/0002 (20130101); H01L 2924/09701 (20130101); H01L
2924/00 (20130101); H01L 2924/0002 (20130101) |
| Current International
Class: |
C23C
16/22 (20060101); C23C 16/24 (20060101); H01C
17/00 (20060101); H01L 21/00 (20060101); H01L
23/522 (20060101); H01L 23/52 (20060101); H01L
21/02 (20060101); G11C 13/04 (20060101); G02B
5/28 (20060101); H01L 27/00 (20060101); H01L
21/314 (20060101); G03F 1/14 (20060101); G03F
1/08 (20060101); H01C 7/00 (20060101); H01L
49/02 (20060101); B44d 001/18 () |
| Field of
Search: |
;117/212,5.5,38,124A,217,45,33.3 ;96/36.2 ;156/17
;350/164,165,166 |
References Cited
[Referenced By]
U.S. Patent Documents
Other References
Baumeister "Multilayer Filters" Institute of Optics, Univ. of
Rochester 1964 pp. 20-14 to 20-18..
|
Primary Examiner: Leavitt; Alfred L.
Assistant Examiner: Simmons; David A.
Claims
I claim:
1. A photomask comprised of a transparent substrate and a
pyrolytically deposited patterned silicon film on one surface of
said transparent substrate, said silicon film being of a thickness
between approximately 50 angstroms and 1,500 angstroms.
2. The photomask of claim 1 including a layer of photosensitive
material over said silicon film.
3. The photomask of claim 1 wherein said photomask is also
comprised of an additional film, said additional film being located
between said transparent substrate and said silicon film and
further being a one-quarter wave length film for a wave length of
light within the visible light spectrum.
4. The photomask of claim 3 wherein said photomask is also
comprised of one or more one-quarter wave length films, for a wave
length of light within the visible spectrum, on one or both
surfaces of said photomask.
5. The photomask of claim 4, wherein said one-quarter wave length
films are substantially one-quarter wave length non-reflecting
films.
6. A photomask comprised of a glass substrate with a patterned film
of pyrolytically deposited silicon on the surface thereof, said
silicon film being of a thickness which is substantially plum
colored when viewed under normal lighting conditions.
7. A light filter comprised of a transparent substrate, a silicon
film on one surface of said substrate, and a layer of metal on
selected areas of said silicon film partially diffused into said
silicon film, mounted to an enclosure having metal surfaces
disposed adjacent at least a part of said layer of metal, said
metal surfaces being attached to the adjacent said layer of metal
by an intermediate layer of solder.
8. The light filter of claim 7 wherein said solder between said
metal layer and said metal surfaces create a hermetic seal.
9. The light filter of claim 7 wherein said layer of metal is a
layer of electroless nickel.
10. The light filter of claim 9 wherein said silicon film is a
pyrolytically deposited silicon film.
11. The light filter of claim 10 wherein said electroless nickel
layer is partially diffused into said silicon film.
12. The light filter of claim 11 mounted to an enclosure having
metal surfaces disposed adjacent at least a part of said layer of
electroless nickel, said metal surfaces being attached to the
adjacent said electroless nickel layer by an intermediate layer of
solder.
13. An article for use in the fabrication of an improved photomask
comprising; a glass substrate having a a pyrolytically deposited
silicon film on one surface thereof, said silicon film being of a
thickness which is substantially plum colored when viewed under
normal lighting conditions.
Description
BACKGROUND OF THE INVENTION
1. Field of the Invention
This invention relates to the fields of silicon coatings and
integrated circuit manufacture.
2. Prior Art
In the prior art, masks used for semiconductor manufacture have
been primarily emulsion masks or chrome masks. The emulsion masks
consist of a photographic gelatin coating containing silver halide
particles on one side of a glass plate. Transparent and opaque
areas are formed by the conversion of the silver halide to metallic
silver. These developed emulsions are relatively soft and because
of this are subject to rapid wear and deterioration in use.
As an alternative, chrome masks are now being made, wherein
chromium is evaporated onto a glass substrate and then a
photo-resist is put over the chrome. Subsequent exposure and
development of the photo resist exposes areas of the chrome which
can be etched away by a suitable etchant. After etching and
removing the photo resist, a chrome image is left on the glass
plate. The chrome is considerably harder than a photographic
emulsion and, therefore, is subject to wear and deterioration at a
much slower rate than the emulsion masks. However, the process for
making such masks is much more involved than the simple
photographic process for making the emulsion masks, thereby
resulting in a much higher cost for chrome masks than for emulsion
masks. Therefore, at the present time, the choice is between a
relatively inexpensive mask which has a short useful life and a
relatively expensive mask which has a longer useful life.
The opaque areas of both types of masks are opaque to visible light
as well as ultraviolet light. Consequently, when the mask is in
place over a semiconductor wafer, the areas of the wafer beneath
the opaque portion of the mask are not visible to the naked eye. In
addition, the chrome masks have the undesirable characteristic of
being highly reflective. Such reflective qualities tend to degrade
the definition that may be reproduced by chrome masks.
Another art involving a transparent substrate, with opaque films
thereon, is the art of optical filters for use with neon or glow
discharge tube read-out devices (e.g. as a window for the enclosure
containing the read-out device). At the present time, a plastic
filter is used which consists of a layer of polarizing material
laminated with a layer of phase rotating material. In operation,
the ambient light (e.g. light originating externally to the
read-out device enclosure) is polarized as it passes through the
filter and then phase rotated 45.degree. by the phase rotating
material. On reflecting from the inner surfaces of the enclosure
back to the filter, it is rotated an additional 45.degree. in the
same direction as the first rotation to complete a full 90.degree.
rotation. Consequently, the polarizing filter blocks this reflected
light and the glow tubes operate in an essentially black
background. Though such filters perform satisfactorily, they are
subject to deterioration by abrasion and scratches and do not have
sufficient strength for hermetic sealing and/or pressurization.
Another art employing substrates coated with opaque materials
involves their use as a means of storing information. In the prior
art, reflective masks have generally not been used. Also, the
closest equivalents for the permanent storage of information are
such things as punched cards and preprogrammed diode matrices;
neither of which is sufficiently similar to the present invention
to draw extensive comparisons.
Heretofore, in the integrated circuit art, semiconductor resistors
have been created by doping a portion of an existing semiconductor
substrate. However, such constructions have only been employed for
very short conductor paths because of their relatively high
resistance. The interconnections for integrated circuits have been
previously made by vacuum deposition of various metals in a pattern
over the substrate. Consequently, conduction path crossovers could
not be created by the direct crossing of either two semiconductor
conductors or two deposited conductors, but had to be created by
the combination of a semiconductor conductor in the substrate, or a
deposited conductor on the substrate, an intermediate silicon oxide
or dielectric coating, and a subsequent metal deposition. By this
method, crossovers have been obtained. In addition, though most
semiconductor processing consists of various heating cycles for
doping and/or oxidizing the surface of the semiconductor substrate,
the vacuum deposition of a metal lead requires separate and
distinct equipment and is a relatively expensive and slow
process.
Similarly, capacitors have been previously created either by using
the characteristic capacitance of a back biased P-N junction,
thereby requiring the proper bias be applied to assure the junction
is back biased at all times, or by the vacuum deposition of a metal
to a non-conducting surface of the substrate, followed by the
application of a thin film of dielectric such as silicon oxide, and
a subsequent second vacuum deposition to create the second
capacitor plate. In the P-N junction capacitor, the P-N junction
must be back biased and the value of capacitance that can be
obtained is low and is sensitive to both temperature and voltage.
It is used, however, because it is easy to fabricate and requires
no processing steps other than those normally used for creating P-N
junctions in semiconductor devices.
At the present time, integrated networks of resistors and
capacitors are not widely used, and when they are used, are
constructed on a semiconductor substrate by using the methods
described above. Such circuits have the limitations described above
including the requirement of the vacuum deposition processing
step.
In many other applications, it is desired to make the electrical
connections to electronic devices which are deposited on, or
created within, the surface of a substrate. By way of example,
precision carbon resistors are commonly made by heating a
non-conducting base or substrate in the presence of methane, which
deposits a thin layer of carbon to the surface of the substrate.
Electrical connections are conventionally made to these carbon
resistors by crimped metal caps or by applying a conductive paste
to the appropriate areas which is subsequently fired in an inert
atmosphere. The crimped metal cap, relying only on mechanical
pressure to assure electrical contact, is relatively unreliable,
and neither the crimped metal cap nor the paste and firing method
lend themselves to extremely small dimensional control and
fabrication.
BRIEF SUMMARY OF THE INVENTION
An object comprised of a substrate with a thin film of patterned
silicon may be used for many new purposes and to achieve many new
results, depending on the nature of the substrate and the thickness
of the silicon film. The word substrate, as used herein, is used in
the general sense to indicate some form of base substance or
substances. By way of example, one embodiment of the present
invention makes use of a glass or transparent plate as a substrate.
Other embodiments may use ceramics, polished metals or other such
materials. In its broadest sense, the word substrate may indicate a
composite base structure, such as an integrated circuit chip
comprised of various doped areas, platings and coatings. The term
"silicon" as employed herein refers to a silicon containing various
impurities such as oxygen, or silicon oxides up to 10 percent by
weight. In most instances these impurities are not intentionally
added but result from the nature or type of process employed.
Certain impurities may be added intentionally.
A transparent substrate with a thin film of silicon on one surface
may be used as a filter over such things as glow discharge data
readout devices. In such an application, the silicon film thickness
would be chosen to allow the light from the device (normally a
reddish light) to be easily seen, while greatly attenuating the two
way reflection of lower wave length ambient light, thereby allowing
the device to operate in a substantially dark background.
Similarly, a transparent substrate with a coating of silicon may be
used as a photomask in certain applications (e.g. manufacture of
integrated circuits) by etching away portions of the silicon
coating to form a pattern or image in the remaining silicon. In
such a mask the silicon, by thickness control, could be made
transparent to normal ambient light, but opaque to shorter wave
lengths light which are operative to perform a function such as
exposing a photoresist or emulsion. A glass substrate having a
pyrolytically deposited silicon film might also provide an article
for use in the fabrication of an improved photomask.
A reflective opaque substrate with a patterned coating of silicon
is useful as a reflective mask or as a means of permanently storing
information.
A substrate containing non-conducting areas, which are coated with
silicon, can be used as a basis for fabricating a multitude of
devices. By way of example, if the silicon coated areas are in a
pattern located on a silicon oxide coating over an integrated
circuit, such areas can be used as resistors and for circuit
interconnections, by either doping the silicon so as to create a
semiconductor resistor or by plating the silicon with a metal.
Circuit cross-overs may be created by a further coating of silicon
oxide over the plated metal, followed by another subsequent coating
of silicon and metal. Capacitive elements may be fabricated as
cross-overs, as previously described, having a substantial area,
thereby creating a significant capacitance at the area of
cross-over. Thus, using ordinary plating techniques and the
deposition of silicon by such processes as pyrolytic deposition,
additional circuits and circuit connections may be added to the
surface of an integrated transistor circuits without having to use
the vacuum processes required by the prior art. In addition,
resistors, capacitors and networks of these components may be
created on other substrates, such as a ceramic chip, to create
hybrid passive networks.
BRIEF DESCRIPTION OF THE DRAWINGS
FIG. 1 shows a transparent substrate with a thin coating of silicon
on one surface.
FIG. 2 shows a cross-section of a transparent substrate with a
patterned coating of silicon on one surface, which may be used as a
photo mask.
FIG. 3 shows a cross-section of a transparent substrate with a
coating of silicon and an intermediate coating to substantially
eliminate the reflection encountered at the interface of the two
transparent materials.
FIG. 4 shows the location of the mask, the semiconductor wafer and
the light sources prior to contact printing a semiconductor
device.
FIG. 5 shows a substrate with a resistor, a capacitor and a circuit
cross-over placed on a surface of the substrate.
FIG. 6 shows a cross-section of the resistor of the circuit shown
in FIG. 5 taken along line 6--6.
FIG. 7 is a cross-section of the capacitor of FIG. 5 taken along
line 7--7.
FIG. 8 is a cross-section of the circuit cross-over of FIG. 5 also
taken along line 7--7.
FIG. 9 shows a carbon film resistor with electrical contact with
the film made by a plated silicon film.
DETAILED DESCRIPTION OF THE INVENTION
FIG. 1 is a cross-section of a flat transparent substrate 10 with a
thin coating of silicon 12 on one surface. The substrate material
may be a relatively inexpensive material and, if pyrolytic
deposition techniques are used to deposit the silicon film, should
be capable of surviving, without detriment, processing temperatures
of 500.degree. to 550.degree.C. Many glasses are well suited to
such an application.
The thin silicon film 12 may be created by any of the common
techniques, such as pyrolytic deposition, vapor deposition or
chemical deposition. If either vapor deposition or chemical
deposition techniques are employed, the substrate need not have the
temperature capability previously mentioned. However, of particular
interest for some applications are silicon films which are
pyrolytically deposited to various substrates. Such films are
created by heating the substrate in either a silane environment or
an environment composed of a mixture of silane and an inert gas.
When so heated, the silane decomposes in the presence of the
substrate and silicon is deposited on the surface thereof. For a
pure silane atmosphere, the rate of deposit is entirely dependent
on the temperature of the substrate and time employed unless a
second order control, such as control of the amount of silane
available, is employed. If a mixture of silane and an inert gas is
employed, then by controlling the percentage of silane in the inert
gas, controlling the flow so that all surfaces for deposition are
exposed to the same amount of controlled mix, and by having the
surface temperature high enough to cause rapid deposition, the
amount of deposition will be at the same rate over all surfaces,
even though there are some variations in surface temperature. Thus,
controlling the mix and flow and also the temperature gives a very
fine overall control of the rate of silicon film buildup, resulting
in the deposit of a very uniform film of the desired thickness.
A thin (e.g. in the order of 50 to 1,500 Angstroms) film of silicon
is relatively transparent to visible light of the longer wave
lengths, such as red, orange and yellow light, but is essentially
opaque to the shorter wave lengths radiation such as ultraviolet
light. However, the amount of light of any particular wave length
that will be transmitted through the silicon film depends on the
thickness of the film. By way of example, films of some thicknesses
will exhibit a maximum reflectivity of incident light and will
transmit a minimum amount of the incident light (of a particular
wave length) while films of other thicknesses will exhibit a
minimum reflectivity and will transmit a maximum amount of the
incident light. These films are referred to herein as one-quarter
wave length and one-half wave length films and are described in
some detail below.
When light passes from a medium of one index of refraction to a
medium of another index of refraction, a portion of the incident
light is reflected from the interface between the two mediums. For
instance, when light is directed through air perpendicular to the
surface of a transparent material, a fraction of the incident light
will be reflected in accordance with the following equation:
r = (n-1).sup.2 /(n+1).sup.2
where r equals the fraction of incident light that is reflected and
n is the index of refraction of the transparent material.
When substantially monochromatic light passes through a medium of
one index of refraction and is directed normal to the surface of a
thin film of transparent material of a second index of refraction
(such as a silicon film) on the surface of a medium of a third
index of refraction, the net reflected light depends not only on
the index of refraction of the mediums, but also on the thickness
of the thin film. If the film thickness is substantially equal to
one-quarter of a wave length for the light being considered
(one-quarter of a wave length for the light in the film, not in air
or free space) then the following will occur. Some of the light
passing through the first medium will be reflected upon first
striking the thin film (e.g. the second medium) and some of the
light will be transmitted through the thin film. Of the light
passing through the film, part will be reflected from the interface
between the film and the third medium and will again pass through
the film to be partially reflected by the interface between the
first medium and the film and partially to pass into the first
medium. Though an exact analysis of the total reflected light would
require a consideration of the endless reflections that occur at
the interfaces of the material (as well as the light absorption
characteristics of the materials) generally the major components of
reflected light in the first medium is that that was initially
reflected from the interface between the first medium and the
second medium and that that was initially reflected from the
interface between the second medium and the third medium. This
second component has traveled twice through the one-quarter wave
length film (once through and once in being reflected back) and
thus is 180.degree. out of phase (e.g. shifted one-half a wave
length) with the light that was initially reflected from the
interface between the first medium and the second medium.
Consequently, to the extent that the two reflected components are
equal, destructive interference results, thereby substantially
reducing the net amount of reflected light in the first medium. It
can be also shown that such a result is obtained if the film
thickness is any odd multiple of one-quarter of a wave length. Such
films are referred to herein as one-quarter wave length films and
are well known in the field of optics.
In addition, since the reflection from either interface depends on
the index of refraction of the mediums, the two reflected
components of light in the first medium may be made to
substantially equal to each other if the index of refraction of the
one-quarter wave length film is properly chosen. In such a case,
the net reflection is essentially reduced to zero. These films are
also well known in the field of optics and are referred to herein
as one-quarter wave length non-reflective films.
One further type of film is referred to herein and, for purposes of
clarity, is presently described. If the film thickness is made
substantially equal to one-half the wave length of the light used
(or odd multiples of one half the wave length) then the two
reflected waves in the first medium will be in phase and will
reinforce each other. Consequently, the net reflection from such a
film is maximized. Also, if the absorption of the film is low, then
the sum of the reflected light and the transmitted light must be
substantially equal to the incident light. Therefore, since the net
percentage of light that is reflected from such a film is
maximized, the percentage of light that is transmitted through the
film is minimized (e.g. minimized for that wave length light). Such
films are also well known in the field of optics and are referred
to herein as one-half wave length reflective films.
Returning now to FIG. 1, the first embodiment of the present
invention may be further described. It was previously stated that
the amount of light of any particular wave length that will be
transmitted through the silicon film depends on the thickness of
the film. For a silicon film thickness 12 of, say, a hundred
Angstroms, the transmission though the film is limited by the
reflections that occur at the interfaces and the absorption within
the film and substrate. Therefore, the transmission of longer wave
length light may be greatly enhanced by using a silicon film
thickness equal to one-quarter of the wave length of the light that
is desired to be transmitted (e.g. a one-quarter wave length film
as previously defined and described), since the amount of light of
approximately that wave length that is reflected from the surface
of such a silicon film is greatly reduced. Consequently, though
some of the light that would have been reflected will be absorbed
within the silicon film itself, much of it will be transmitted
through the film. For instance, a silicon film thickness
characterized by a green "interference color" (believed to be
approximately 1,250 A thick) will transmit nearly all of the red
light incident to the film while reflecting (and absorbing)
substantial portions of shorter wave length visible light.
A transparent substrate 10 with a thin silicon film 12, as
described above, is useful as a low pass light filter, that is, as
a filter which will pass light of the desired wave length but will
not pass substantial amounts of light of the shorter wave lengths
as described above. Such a filter might be used as a covering for
an enclosure containing neon glow discharge tubes or solid state
electroluminescent data readout devices which emit radiation of a
wave length approximately equal to that of red light. In such an
application, the data readout device will be easily visible through
the transparent substrate and silicon film. However, the majority
of ambient light, either natural or artificial, is normally of a
shorter wave length. Consequently, much of this light will not pass
through the silicon coated substrate (most of that which does pass
through will be absorbed by the walls of the enclosure and much of
that which is reflected from the walls will be re-reflected from
the coated substrate rather than to pass to the eyes of a person
viewing the readout device) and thus, the data readout device will
operate in an essentially black background.
Such a filter has advantages over the presently used filters in
that the silicon coating is extremely hard and wear resistant. In
addition, these filters are relatively rigid, strong and suitable
for use where a hermetic seal is to be made between the filter and
the other parts of the enclosure. One method of creating such a
hermetic seal is to selectively metal plate the silicon surface
around the edges of the filter and then to create the hermetic seal
by such means as soft soldering or brazing (e.g. commonly referred
to as hard soldering) the metal plated surface to the metal
enclosure. The silicon surface may be plated by any of various
means, such as the electroless plating of nickel to the desired
surfaces of the silicon, or by doping the silicon in selected areas
by conventional semiconductor manufacturing techniques so as to
lower the resistance of the silicon coating and thereafter
electroplating with one or more metals to the doped silicon
surface. If pyrolytic deposition techniques are used, doping may be
accomplished during the deposition by using a silane atmosphere
containing the desired amounts of gaseous dopants or compounds of
such dopants. Peeling of the plating may be prevented by a
subsequent heating so as to diffuse some of the metal into the
silicon coating. It is to be understood that the partial diffusion
of the metal into the silicon film includes the diffusion of the
metal so completely into the silicon that the result has the
appearance of a homogenous material.
It is to be understood that the method of creating a hermetic seal
as described above is not limited to light filters. Such hermetic
seals might also be used for other purposes, such as to seal
electrical feed throughs or covers in the manufacture of precision
instruments. In such applications, the materials constituting the
substrate (the covers, feed throughs, etc.) need not be
transparent, but may be essentially any material of a reasonable
thermal expansion rate and to which the silicon coating will
adequately adhere, such as most metals, glasses and ceramics. The
advantage of such seals over conventional fused glass to metal
seals is that the ultimate heremetic seal may be created by any
metal joining process such as a relatively low temperature soft
soldering process rather than a high temperature glass fusing
process.
FIG. 2 is a cross-section of a transparent substrate 10 with a
patterned silicon coating 14 on one surface. The patterned coating
on the transparent substrate may be created from the silicon coated
substrate of FIG. 1 by such techniques as conventional photoetching
techniques. To obtain the pattern, the silicon coating would be
first coated with a photosensitive film and then exposed to a light
source through an appropriate photomask, so as to expose portions
of the photosensitive surface. Such exposure fixes the exposed
areas of the photosensitive film, and a subsequent immersion in an
appropriate developing solution will dissolve away the unexposed
areas of the photosensitive film. Immersion of the part (with the
developed photosensitive material thereon) into a suitable etchant
will remove the areas of silicon which are not coated with the
fixed photosensitive material. (One etchant that has been used for
this purpose is a solution comprised of 1 part HF (reagent), 6
parts nitric acid (reagent), 2 parts H.sub.2 O (deionized) and 4
parts of a solution comprised of 84.1 grams FeCl.sub.3 6H.sub.2 O
in 100 c.c. solution with H.sub.2 O (deionized). Finally, immersion
in appropriate solvent will remove the fixed photosensitive
material, leaving the substrate with a pattern of silicon on the
surface as shown in FIG. 2. The resulting coating is referred to
herein as a patterned silicon coating.
Such a transparent substrate and a patterned coating of silicon is
useful as a photomask in processes where the photosensitive
material is sensitive to short wave length radiation, such as
ultraviolet light, and where the silicon coating is essentially
opaque to such radiation. This type of mask has advantages over
both emulsion masks and chrome masks in that the silicon pattern is
exceptionally hard, durable and wear resistant, and is much cheaper
to fabricate than are chrome masks.
In addition, a silicon film mask as shown in FIG. 2 has a
characteristic unobtainable with either the emulsion masks or the
chrome masks. The silicon film, though being essentially opaque to
shorter wave length radiation such as ultraviolet radiation, is
transparent to much of the light in the visible range.
Consequently, a person viewing the photosensitive surface through
the silicon film mask may see the entire photosensitive surface,
including that part of the surface lying beneath the silicon.
It has been found in practice that two factors detract from the
clarity of the photosensitive surface as viewed by a person under
light in the visible spectrum. In use as a photomask, the mask of
FIG. 2 would be placed over the photosensitive surface with the
silicon pattern 14 resting on the photosensitive surface. To view
the photosensitive surface, visible light would be directed to as
to pass from the surrounding air, through glass substrate 10,
through silicon pattern 14, through a thin film of air to reflect
off the photosensitive surface, thereupon returning through air,
the silicon pattern 14, the glass substrate 10 and through the air
to the viewer. However, when light passes from a medium of one
index refraction to a medium of a second index refraction, a
certain percentage of the light is reflected at the interface
between the two mediums. By way of example, approximately 4 percent
of the initial light impinging on the glass substrate 10 is
reflected at the air-glass interface. Similarly, approximately 15
percent of the light passing through the glass is reflected from
the glass-silicon interface and approximately 30 percent of the
light passing through the silicon is reflected from the subsequent
silicon-air interface. These reflections detract from the clarity
and contrast of the photosensitive surface as viewed under visible
light for two reasons. First, normally it is desirable to have a
silicon film thickness which is approximately equal to a quarter of
a wave length for orange or yellow light, since the photosensitive
surface will normally be viewed under such light. Such a film is
approximately 830 Angstroms thick and appears "plum" colored with
reflected lights. This one-quarter wave length thick silicon film
effectively performs as a non-reflecting film as described before
and will substantially reduce the net reflection of light in the
lower wave length region of the visible spectrum, such as red,
orange and yellow light. However, for shorter length visible light,
such as blue light, the components of reflected light from the two
film surfaces are not substantially out of phase, consequently
there is less cancellation by interference, with a substantial net
reflection of these wave lengths. Consequently, most of the
reflected light from such a film, under ordinary lighting
conditions, will be of the shorter wave lengths such as blue light
and will appear as a hazy overcast to the photosensitive surface.
(In addition, the absorption of the silicon for these shorter wave
lengths further detracts from the effectiveness of the film as a
nonreflective film for blue light since the component of reflection
that passes twice through the film is greatly attenuated and
emerges too small to accomplish substantial destructive
interference). Since this color is not within the range of wave
lengths needed to be used for viewing the wafer, it can easily be
removed by a filter that passes only the red-orange-yellow
spectrum, or by a multilayered interference filter that passes all
visible wave lengths longer than the blue region. This latter
filter is more effective than the colored filters as it allows more
useful light to be used for viewing. This filter could be
incorporated in the mask construction by depositing the appropriate
multiple films. However, it is more practical to use one filter per
alignment machine or viewer than to reproduce the filters on each
mask.
The reflection at the air-glass interface can be eliminated in the
red-yellow wave length range by coating the back of the glass
substrate 10 with a similar one-quarter wave length anti-reflective
coating for that wave length range. However, since this reflection
was only approximately 4 percent, it would not be a significant
enough improvement to warrant the expense in most cases. Also, in
some applications and/or with some light sources, it may be
desirable to use an anti-halation backing to prevent fogging due to
light reflection and dispersion within the silicon film. Such
backings are well known in the field of photography and will not be
described herein.
The 15 percent reflection at the interface of the glass-silicon
interface could be significant enough to be undesirable, and if so,
this reflection may be substantially eliminated by an intermediate
interface of either of two types. One type is to place a
one-quarter wave length film between the glass and silicon, this
film having a refractive index approximately equal to the mean
difference between refractive indexes for glass and silicon (e.g.
equal to the square root of the product of the refractive indexes
of glass and silicon). With such a film, essentially complete
destructive interference results between the reflection from the
two interfaces. The other type is an inhomogenous film which is
applied to the glass to which the silicon is subsequently
deposited. As this intermediate film is first being applied, it is
composed of materials of a refractive index closely matching glass.
As the intermediate film builds up, the refractive index is then
linearly, or in multiple discrete steps, changed so that on
completion of the film the refractive index of the last deposited
material closely matches that of silicon. This may be done by
compounding two materials in a changing mix ratio as the film is
deposited; one material closely matching the index of refraction of
glass and the other material closely matching the index of
refraction of silicon. By way of example, one method of creating
such a film is to first deposit silicon trioxide on the glass, then
silicon dioxide, then silicon monoxide and finally just silicon.
Such materials may be deposited by conventional vapor deposition
techniques.
In addition, the silicon-air interface can be made anti-reflective
by the same technique (that is by the application of a suitable
quarter wave length thick coating to the surface of the silicon),
but such a film would cause some loss of resolution in the printed
image due to the spacing it would introduce between the silicon
pattern and the photosensitive surface. Normally, this distance is
made as small as possible by employing a vacuum between the mating
surfaces of the mask and the photosensitive surface and/or the use
of pressure on the face of the mask. Therefore, the gain from such
a film would be offset somewhat, depending on image sizes, and the
extent of collimation of the light source by the reduction in image
resolution. Such anti-reflective coatings are well known in the
field of optics and may be applied to the various reflective
surfaces as the particular application allows or requires.
If the mask is intended for a projection type of use wherein the
pattern is projected onto a surface by a lens system, the
one-quarter wave length antireflective coating could be of
significant value. In such an application, the silicon mask has the
additional advantage of high resistance to the elevated mask
temperature caused by high intensity illumination systems, as
compared with conventional emulsion masks.
A mask with various anti-reflecting films is fabricated starting
with a suitable substrate coated with the appropriate films which
are subsequently etched in a pattern to create the mask. By way of
example, FIG. 3 shows a substrate 10 with a silicon coating 12 and
an intermediate anti-reflective coating 16. The subsequent photo
etching of the surface will remove portions of the silicon coating
12 and, if desired, corresponding portions of the anti-reflective
coating 16, thereby creating a patterned silicon coating, beneath
which lies an anti-reflective coating. In some instances a
different etchant may be required for each coating.
In the preceding explanation, the advantages of a silicon coating
that is transparent to visible light and particularly the methods
of achieving and utilizing this result were described in detail.
However, it is to be understood that the hardness, wear resistance,
and use of fabrication of silicon masks give these masks
substantial advantages over the prior art masks even without
obtaining (or making use of) the transparency of the silicon film
to visible light. By way of example, a silicon mask which does not
have any of the various antireflective coatings and its
transparency to visible light is not utilized is useful because of
its advantages of long life, low cost and low pin hole density in
the silicon coating. Similarly, a silicon mask in which the silicon
is substantially opaque because of such things as the presence of
impurities in the silicon will also have the above advantages over
the prior art masks. (In some applications such impurities might
intentionally be added so that the mask would have a visual
appearance similar to prior art masks).
FIG. 4 is a schematic diagram of a contact printing set up, such as
used in semiconductor device manufacture, using a silicon mask of
the present invention. The specific silicon mask shown has an
intermediate antireflective coating as was shown in FIG. 3 (details
of the pattern are omitted for clarity). The photosensitive surface
20 is a thin film of photo resist applied to the surface of a
semiconductor wafer 18. The photo resist 20 is sensitive to
ultraviolet light but is relatively insensitive to light in the
visible range. The photo mask is positioned over the photosensitive
surface 20 with the silicon image 12 in contact with the photo
resist. To assure a high quality printed image, normally a mask is
held firmly against the semiconductor wafer by a vacuum between the
wafer and the mask and/or by applying pressure to the back surface
of the mask. Because of these pressures, conventional emulsion
masks are frequently damaged due to the softness of the emulsion.
However, the silicon is well suited to such use since it is
extremely hard and wear resistant, and may be deposited and etched
by relatively inexpensive processes to yield exceedingly fine
patterns of material having a low pin hole density.
The mask and photo sensitive surface may be illuminated by either
visible light or by ultraviolet light. Light bulbs 19 are sources
of visible light and filters 21 are used to filter this light so as
to substantially eliminate the wave lengths that cause the hazy
overcast previously described. This light source allows an operator
to view the entire semiconductor wafer, for purposes of alignment
and inspection.
Light source 22 is a source of ultraviolet light and is then used
to expose the photosensitive surface 20 when the mask and wafer are
properly aligned. Such a silicon mask may be repeatedly used
without substantial wear and deterioration because of its hardness
and resistance to wear. In addition, such masks may be made to be
essentially free of pin holes since the pyrolytic deposition of
silicon by the decomposition of silane results in an extremely
uniform and pin hole free coating.
As an alternative embodiment of the present invention, the
patterned silicon film may be applied to a reflective substrate and
used as a reflective mask (for the wave lengths of light for which
the light absorption of the silicon is high such as ultraviolet
light.) In such an application, the reflection from the surface of
the silicon pattern can be substantially eliminated by applying a
one-quarter wave length non-reflective coating to the surface of
the silicon. If the pattern is created by the conventional
photoetching technique described earlier, the etchant and the
etching times should be selected so as to not unduly etch the
substrate and substantially detract from its reflective qualities.
Such a reflective mask could be used where optical enlargement or
reduction of the image was desired by passing the light reflected
from the mask through a suitable lens system.
In addition, objects of the construction of the "transparent" mask
or the reflective mask may be used for the permanent storage of
information or as a key or identification means. The storage means
may be accomplished by various schemes, one scheme being the
storage of data bit information as is done on punched cards in the
prior art.
As a key or identification means, the code or identification would
not be visible to the eye under normal lighting but could be
optically read by exposure to ultraviolet light and the viewing the
pattern by electro-optical means or by visually viewing the pattern
through an appropriate apparatus.
FIG. 5 shows a top view of a substrate 24 with a resistor 26, a
capacitor 28 and a circuit crossover 30 created by the deposition
(and processing) of various materials to the surface of the
substrate 24. Though the particular circuit shown is a very simple
circuit, it is to be understood that, by the techniques of the
present invention, extremely small and complex circuits containing
resistors, capacitors and/or circuit connections and crossovers may
be created on a small non-conducting substrate or on a
non-conducting surface of a composite substrate such as an
integrated semiconductor circuit.
FIG. 6 shows a cross-section of the resistor 26 of FIG. 5 taken
along line 6--6 of that figure. In this figure a dashed line 34
near the top surface of substrate 24 is to indicate that the
substrate may be of a composite structure, provided that at least
the surface of the substrate 24 is non-conducting (e.g., is an
insulator such as a silicon oxide). To the surface of substrate 24
is first applied a silicon coating 38. This coating may be applied
by any of the conventional techniques, such as vapor deposition.
However, of particular interest, for reasons which will
subsequently become apparent, is the pyrolytic deposition of
silicon to create the coating 38. This is achieved by the heating
of the substrate in the presence of silane, thereby decomposing the
silane and depositing the silicon from the silane so decomposed.
This process is presently commonly used in semiconductor device
manufacture. As employed herein, this process is used to deposit a
silicon film, generally polycrystalline in form, in the
non-conducting (insulative) surface of a substrate. Generally, this
deposit will not be a crystal for crystal extension of the
substrate lattice, and, in general, the substrate itself may be
polycrystalline in form, such as an ordinary ceramic. If the
substrate is a silicon semiconductor device, what is contemplated
is not a crystal for crystal extension of the substrate lattice but
rather the deposit of a silicon film upon an insulated coating,
such as a silicon oxide coating, over the semiconductor.
After the silicon coating has been deposited, a selected impurity
may then be diffused into the central portion of silicon film 38 so
as to form the semiconductor resistor 26. Such diffusion techniques
are well know to the art of semiconductor manufacture. In other
selected areas, metal is plated over the surface of the silicon
film 38, so as to form the circuit connections 40. This metal may
be plated to the silicon by electroless plating techniques or by
doping the silicon as was done to form resistor 26 and
electroplating the selected areas. It has been found that by
subsequent heating, a plating on the surface of the silicon film
may be partially diffused into the silicon thereby making it
virtually impossible to peel off. The net result is a semiconductor
resistor with electrical connections that has been created on an
insulator, rather than within a semiconductor substrate, and which
has been created without the use of a vacuum process. The precision
of the resistor is largely determined by the precision of the
etching process and the control of the diffusion process. The
largest single advantage of this new technique is that the
processes involved are substantially the same as those presently
used to manufacture semiconductor devices, and without additional
equipment, complex passive circuits may be deposited on the surface
of the insulative coating on a semiconductor device and connected
to and made part of the semiconductor circuit by the same
techniques.
FIG. 7 is a cross-section of a capacitor of FIG. 5 taken along line
7--7 of that figure. The construction and many of the construction
techniques are the same as those for the resistor of FIG. 6. The
first step is to deposit a patterned silicon layer 42 to the
nonconductive coating of the substrate 24. To this coating a metal
layer is plated which may, if desired, be partially diffused into
the silicon. Then an insulating layer 46 is deposited, such as
silicon oxide or silicon nitride. These insulating layers may also
be pyrolytically deposited as are the silicon films. Then a further
silicon film 48 is deposited and finally metal film 50, which forms
the top plate of the capacitor and the electrical connection to
this top plate. By keeping silicon film 48 relatively thin and/or
by diffusing part of the metal film 50 into the silicon film, the
thickness of dielectric is principally determined by the thickness
of the insulating film 46.
It is common practice in the semiconductor fabrication art to
provide a final layer on otherwise exposed silicon surfaces to
serve as an environmental protection layer. Layers of silicon
monoxide and silicon dioxide are commonly used for the purpose and
may be created by heating the device in an oxidizing atmosphere. It
is to be understood that any of various protective layers, well
known in the prior art, can be used in conjunction with the devices
of the present invention, but such coatings have not been otherwise
described or pictured in the drawings for clarity only.
FIG. 8 is a cross-section of a circuit cross-over of FIG. 5, also
taken along line 7--7 of that figure. The construction of the
cross-over may be essentially identical to the construction of the
capacitor shown in FIG. 7, with the exception that the area of the
cross-over is minimized for a circuit cross-over whereas this area
is maximized if a capacitor is desired. The films shown in FIG. 8
are a first silicon film 52, a second film of metal 54, a third
film of insulating material 56, a fourth film of silicon 58 and
finally a top film of metal 60 forming the cross-over conduction
path. Each of these films, of course, is formed in an appropriate
pattern to create the desired end result. Also, it is to be
understood that circuit crossovers consisting of a conduction path
crossing a resistor or resistor crossing a conductor may be readily
fabricated by the techniques previously described.
FIG. 9 is a cross-section of a carbon film resistor. Such resistors
usually consist of a thin film of carbon 70 pyrolytically deposited
to a non-conducting substrate 72 by the high temperature
decomposition of methane. A high reliability electrical contact can
be made to such deposited films by applying a silicon film 74 to
each end of the non-conducting substrate 72 (preferably by
pyrolytic decomposition of silane, thereby allowing the use of the
same deposition equipment as is used for the deposition of the
carbon film). To the silicon films 74 is applied one or more metal
platings 76 in the same manner as previously described (in the
preferred embodiment the first is partially diffused into the
silicon film as previously described.) Now the deposition of a
carbon film 70 on the center portion of the substrate 72, in a
manner so as to overlap the metalized silicon ends, forms a
positive electrical contact between the film 72 and the metalized
ends. This electrical contact may be established by the physical
contact of the carbon film 70 and the metalized silicon ends, or
may result from the partial diffusion of the silicon, metal and/or
carbon films. Permanent leads 78 may be readily attached to these
metalized ends using conventional thermal compression, ultrasonic
or soldered connections. An insulative coating 80, such as an epoxy
coating, provides the desired physical protection and electrical
insulation of the device. As with the prior devices fabricated by
the use of a deposited silicon coating, an advantage of the above
described method of making electrical contact, as opposed to the
prior art methods, is that a high strength and high reliability
connection can be made to the resistive film using substantially
the same processes as are used to create the resistive film itself,
thereby reducing the equipment and simplifying the production
processes for such devices. In alternative embodiments, the carbon
film can be first deposited and then the silicon and metal
deposited over the carbon, or the carbon film may be an
intermediate film between films of silicon and/or metal. Electrical
contact between the carbon film and the metal films may be achieved
by physical contact between the carbon film and the metal film, or
in the alternative may be achieved by the partial diffusion of the
metal through the silicon film to effectively make electrical
contact to the carbon film .
* * * * *