COMMISSIONER'S DECISION
SUPPORT IN THE DISCLOSURE; OBVIOUSNESS - Semi Conductor Assembly
The inventions claimed are semi-conductor circuits, and apparatus to make them.
They were refused for obviousness, and under Rule 25 for inadequate disclosure.
It was concluded that the drawings did provide the necessary support, and since
the process deletes two steps from that required in the prior art (thereby making
it more suitable for automation), ingenuity is present.
Final Action: Reversed
********
Patent application 177,075 (Class 356-131), was filed on July 23, 1973
for an invent-ion entitled "Method And Apparatus For The Assembly Of
Semiconductor Devices." The inventor is Terry W. Noe, assignee to Texas
Instruments Incorporated. The Examiner ire charge of the application
took a Final Action on August 2, 1977 refusing to allow it to proceed
to patent. In reviewing the rejection, the Patent Appeal Board held a
Hearing on July 25, 1979 and at which the Applicant was represented by
Mr. J. Baker.
The subject matter of this application relates to an integrated circuit
package capable of being produced by automation. The drawings are as
follows:
(see formula I, II, III, IV, V)
Film 11 has interconnect patterns to which semiconductor chips 31 are
attached. When the film is in position a punch 33 is driven through a die 35
to cut the semiconductor and move it to a lead frame strip 21. At the same
time bonding tool 41 is elevated. to solder lead ends 22 with the bonding
areas 17.
In the Final Action the Examiner rejected claims 1 to 5 in view of cited
art, common knowledge of semiconductor manufacture, and for lack of support
under Rule 25. Claim 6 was similarly rejected except that Rule 25 was not
applied to it. The reference applied to reject the claims is a publication
entitled:
"IC's on Film Strips Lend Themselves to Automatic Handling"
S.E. Scrupski
Electronics, February 1, 1971, pages 44-43.
Two references of interest were also cited. They are:
Publication
"Applications of Polyimide Materials in
Electronic Circuitry"
J.R. Cannizaro
Solid State Technology, November 1969, pages 31-39.
Canadian Patent
837,800 Mar. 24, 1970 R.W. Helda
In the Final Action the Examiner rejected claim 6 using a detailed side-by-side
comparison with the Scrupski reference. Claims 1 to 5 were also refused and
the Examiner had this to say about them (in part):
...
Regarding process claims 1-5 the Applicant argues as follows in
his last letter:
X) Bonding of the chip is done by pulsing a current in the
Scrupski publication.
Y) Original claim 2 includes the phrase "said separation
being achieved by means of the motion of said punching
means as it approaches said bonding tool". This
passage supports the statement that the process step
of severing the thin film interconnect and the bonding
of the interconnect: to the lead frame is done "with
a single motion punching means." The applicant
holds that this method is inherent in the original
disclosure.
Z) Scrupski does not suggest applicant's reflow soldering
step.
With regard to part (X) the examiner is in complete agreement
with the applicant, as is also shown in the first sentence
of part (1) of true previous Office Action.
With regard to part (Y) the examiner maintains that there is
no suggestion nor any support whatsoever in the application
for a single motion punching means. Examination of page 4~
of Scrupski shows that when a chip is in position above a
lead frame, a temporary support moves in horizontally between
the two strips (interconnect film and lead frame) and a head
comes down to shear the leads out of the film. Thereafter,
the support moves back and the head holds the chip with
vacuum and then moves all the way down to the lead frame.
A quick tack of the leads then is made...
Clearly, Scrupski describes a dual motion punching means
which achieves separation of the thin film interconnect by
means of said punching means as it approaches the bonding tool.
However, only during the second downward motion of the
punching means is the interconnect bonded to the lead frame.
The last underlined passage describes exactly Scrupski's dual
motion method. Yet the applicant claims that the identical
passage in his original claim 2 supports a single motion punch-
ing means.
As the disclosure is of no further avail in deciding whether
a single notion punching means is disclosed, perhaps a look
at the drawings can help. Figures 3 to 5 show consecutive
punch positions. It is noted that the shearing die is completely
removed in Figure 4, just as Scrupski describes the method in
his article. Thus the present application clearly shows and
describes a dual motion punching means for severing the thin
film interconnect and bonding it to the lead frame where the
sheared unit is transferred by punch to a position in mated
contact with one unit of the lead frame strip.
The newly submitted matter on page 2 of this application is
refused for non-support in accordance with Rule 52 and must be
deleted.
With regard to item (Z) the difference between Scrupski's and
applicant's methods of bonding were pointed out in part (1)
of the previous Office Action. Scrupski prefers to use a separate
pressure and solder reflow operation to join the TF
interconnect to the lead frame. The applicant in his original
claim 1 and in a later amended version uses a combined pressure
and solder reflow operation. How exactly pressure is applied
during reflow soldering has never been explained. The word
"pressure" was merely dropped from the claim and substituted
with "temperature". Obviously the important parameter in
reflow soldering is temperature. Thus contrary to applicant's
statement that "Scrupski does not suggest applicant's reflow
soldering step", it appears that the applicant has benefitted
from Scrupski's reflow soldering process.
...
In response to the Final Action the applicant deleted claim 6 and presented
new claims 1 to 5. An affidavit from a Mr. Jones G. Harper was submitted and
the Applicant made the following arguments:
The G.E. Process and related technology was purchased from
General Electric by the present applicant, Texas Instruments,
and represents the forerunner of the technology described in
the present application. T.I., in attempting to incorporate
the G.E. Process into their operations found that it had certain
quite severe deficiencies. One of these was that the interconnect
patterns tended to rupture, thereby creating open circuits. This
rupturing was especially prevalent during thermal testing of the
completed product.
A second serious problem with the G.E. Process was that the
bonding tip system which they disclosed as being "pulsed" thereby
generating a cyclic temperature variation tended to be very
unreliable in terms of providing a secure bond at all of the
bonding positions. This created an exceedingly high ratio of
defective parts.
The T.I. Research group set about to solve the defects in the G.E.
Process and the present invention is the result of those efforts.
In the technology provided to T.I. by G.E. the interconnect
pattern was in fact being formed using copper foil formed by
an electrodeposition process. Whether or not this is stated in
the article referred to by the Examiner, it is in fact so as
can be attested to by a sworn statement if necessary.
T.I. found that the solution to the rupturing of the interconnect
pattern was to use copper foil made by a rolling technique rather
than copper foil made by an electrodeposition technique. Rolled
copper foil is a well-known material which has long been available
and is formed by rolling copper on a hot mill. It is well known
that copper can be hot rolled to thicknesses of as little as
0.001-0.002 inch. It was found that because the rolled copper
foil has a distinctly different pattern of atomic bonding, it
was much more capable than electrodeposited copper foil of
resisting rupture due to thermal cycling stresses. Thus, it solved
the problem of the rupturing of the copper interconnect pattern.
...
Thus, the G.E. Process which was acquired by T.I. clearly
did have deficiencies which made it impractical for
commercial use. Through the use of ingenuity which was clearly
beyond normally expected skill, T.I. by way of novel procedures
satisfactorily solved the deficiencies such that a procedure
was obtained which is commercially useful. Such procedure is
described in amended claim 1.
...
At the Hearing Mr. Baker presented a new claim 1 and an affidavit from
Mr. K. Wolford.
The question before the Board is whether or not the Applicant has made a
patentable advance in the art, and also whether there is sufficient support
in the disclosure for what is claimed.
Our first consideration is that of non-support.
On page 5 of the Final Action the Examiner sets forth his reasons for
refusal under Rule 25 of the Patent Rules. He indicates that Scrupski
"achieves separation of the thin film interconnected by means of said punching
means as it approaches the bonding tool," which is claimed in Applicant's
original claim 2, as supporting a single motion punching means. He goes on,
stating that the "disclosure is of no avail in deciding whether a single punch
means is disclosed". Figures 3 to 5 of the drawings were also discussed and it is
noted that the "shearing die is completely removed in Figure 4, just as
Scrupski describes the method in his article".
In arguing his case Mr. Baker stated that the Applicant's "process does operate
with punch 33 moving downwardly through the shearing die 35, and continuing
its downward path until it contacts the lead frame strip 21". To support this
position he quoted from page 3 lines 16 to 21 where a brief description
of the punching means is found and from page 5, line 16 to page 6, line 4 where
the more detailed punch description is located.
We have carefully studied the drawings, as well as the pages of disclosure
referred to by Mr. Baker. From figures 3, 4 and 5 of the application we
see the sequence of movement of the punch, the bonding tool and the
position of the shearing die. Since the punch is shown in phantom on its
downward movement in figure 3, we conclude that the shearing die is in a
stationary fixed position. Combining the drawings with the disclosure,
statements on page 3 lines 16 to 21 and page 5 line 16 to page 6 line 4
leads us to conclude that a single motion punching means is used, and
the rejection under Rule 25 should be withdrawn. Perhaps if both the lead
strip and the film strip were shown in each drawing, and the line of punch motion
would he on the same vertical line in all drawings, It would be clearer and
easier to understand figures 3, 4 and 5. Further, in figure 5 the chip 31
should not be shown in the withdrawing punch.
Another question that has arisen involves the alignment accuracy of the film
and lead frame. On page 3 of the disclosure at line 8 Applicant states:
"While supplying both the lead frame and interconnect pattern (with the semi-
conductor ship attached) in strip form from large reels, successive units
are indexed in exact alignment with each other by sprocket drive means."
On page 6 at line 3 the Applicant states the "two strips are indexed in
registration." Scrupski states at page 47 that "a metal strip or dual in-line
lead frames is aligned via the indexing holes." From these statements we
conclude that acceptable alignment accuracy is obtained by the use of the
sprocket drive holes.
In the Final Action the Scrupski publication and common knowledge of semi-
conductor manufacture were applied in the rejection of claims 1 to 5.
Scrupski describes the process of attaching chips to dual in-line lead frames
on page 47 in column 2 as follows:
A metal strip of dual in-line lead frames is aligned, via
the indexing holes, with the film strip holding the Multibonded
chips. When a chip is in positron above a lead frame, a
temporary support moves out horizontally between the two
strips and a head comes down to shear the leads out of the
film. The support then moves back; the head holds the chip
with vacuum and then moves all the way down to the lead
frame. A quick tack of the leads then is made and the lead
frame strip moves on to the next station for reflow soldering.
A comparison of Scrupski's process with that of this application does show
some essential differences. First, Scrupski has movement of the temporary
support horizontally between the two strips. Secondly, the "support
then moves back," and thirdly, there is the quick tack of the leads. Then
the. lead frame strip moves on to the next station for reflow soldering",
Each of these movements require a slight time pause in an automated process.
By contrast Applicant's process provides the finished product with the single
punch motion downward and sequential upward movement of the bonding tool for
reflow soldering. This, in our view, would be more desirable than what is
described in Scrupski.
Evidence of commercial success of the Applicant's process is outlined in
the affidavit from Mr. K. Wolford, which was submitted at the Hearing.
This affidavit states that assembly line production embodying the invention
has been commercially successful for a period of more than 5 years. It also
indicates that the S.E. Scrupski apparatus was not capable of commercial
operation at that time.
The other affidavit from Mr. G. Harper, which accompanied the response to
the Final Action, outlines the problems of the Scrupski process, and the
steps taken by Texas Instruments to overcome these problems. At the Hearing
the Examiner questioned whether Applicant was merely trying to patent the
proprietary process purchased from G.E. Mr. Harper declares in paragraph 2
of his affidavit that the technology purchased from G.E., what is described
in Electronics February 1, 1971 pages 44-48 was found to have certain de-
ficiencies in terms of commercial utility. Clearly, the process obtained
by T.I. from G.E., was that described in the Scrupski. publication, and it
did have the deficiencies which apparently have now been overcome by T.I.
At the Hearing Mr. Baker presented a new claim 1 to replace existing claim 1.
It reads as follows:
In a process for the fabrication of integrated circuit
assemblies, wherein a plurality of semiconductor chips are
attached to a strip of flexible dielectric film having
a corresponding plurality of conductive interconnect
patterns of rolled copper laminated thereon, the improved
method of attaching the interconnect patterns to an external
lead frame strip including a plurality of lead frame units
integrally joined comprising the steps of:
coating the appropriate portions of the lead frame
strip and the interconnect patterns with a
subtable low-melting metal or alloy; placing in
alignment (1) the to-be-bonded portions of one of
said interconnect patterns, (2) the corresponding
portions of a lead frame unit, (3) a heated bonding
tool maintained at an essentially constant temperature
well above the melting temperature of said low-melting
metal or alloy, (4) a punching means shaped to mate
with the bonding tool, and (5) a shearing means between
said punching means with said bonding tool, shaped
to mate with said punching means, advancing said
punching means toward said bonding tool, thereby severing
said one interconnect pattern from said strip of flexible
dielectric film by causing an engagement of said punching
means with said shearing means; and then continuing said
advancement of said punching means, with said interconnect
pattern in place thereon, and by bringing said punch
means and heated bonding tool in close proximity, sand-
wiching the interconnect pattern and the lead frame
therebetween at a temperature and for a time sufficient
to temporarily reflow and low-melting metal or alloy,
thereby completing the bond.
This claim specifies a process of fabrication of integrated circuit assemblies,
where chips are attached to a dielectric film having corresponding conductive
interconnect patterns of rolled copper laminated thereon. This includes the
steps of continuing the advancement of the punching means and bringing the punch
means and heated bonding tool in close proximity for a time sufficient to
temporarily reflow the low melting metal, thereby completing the bond. The
cited art dues not teach this specific combination, and in view of the
statements in the affidavit, we do not find it obvious therefrom
Accordingly claim 1, as submitted at the Hearing, is, we believe, allowable,
and we recommend that it be accepted.
Having recommended that the rejection under Rule 25 be withdrawn, we are
also satisfied that the Applicant has made a patentable advance in the
art. On the basis of the art cited there is, in our view, sufficient ingenuity
that the Commissioner ought not to refuse a patent (cf Crosley Radio
Corporation v Canadian General Electric (1936) S.C.R. 551 at 560). Consequently
we recommend that the objections made be withdrawn and that claim 1 and
dependent claims 2 to 5 inclusive be accepted.
G.A. Ashen S.D. Kot
Chairman Member
Patent Appeal Board, Canada
I have carefully reviewed the prosecution of this application and considered
the recommendation of the Patent Appeal Board. I concur with the reasoning
and findings of the Board. Accordingly, I withdraw the Final Action and
return the application to the Examiner for resumption of prosecution.
J.H.A. Gariepy
Commissioner of Patents
Dated at Hull, Quebec
this 19th. day of November, 1979
Agent for Applicant
Kirby, Shapiro, Eades & Cohan
Box 2705, Stn. D
Ottawa, Ont.
K1P 6H2