Ex Parte 6,815,321 et al

Board of Patent Appeals and Interferences

Jul 10, 2012

95000483 (B.P.A.I. Jul. 10, 2012)

UNITED STATES PATENT AND TRADEMARK OFFICE
UNITED STATES DEPARTMENT OF COMMERCE
United States Patent and Trademark Office
Address: COMMISSIONER FOR PATENTS
P.O. Box 1450
Alexandria, Virginia 22313-1450
www.uspto.gov
APPLICATION NO. FILING DATE FIRST NAMED INVENTOR ATTORNEY DOCKET NO. CONFIRMATION NO.
90/009,482 06/08/2009 6,815,321 051700-5005-US 9596
24341 7590 07/11/2012
MORGAN, LEWIS & BOCKIUS, LLP. (PA)
2 PALO ALTO SQUARE
3000 EL CAMINO REAL, SUITE 700
PALO ALTO, CA 94306
EXAMINER
KIELIN, ERIK J
ART UNIT PAPER NUMBER
3992
MAIL DATE DELIVERY MODE
07/11/2012 PAPER
Please find below and/or attached an Office communication concerning this application or proceeding.
The time period for reply, if any, is set in the attached communication.
PTOL-90A (Rev. 04/07)
UNITED STATES PATENT AND TRADEMARK OFFICE
UNITED STATES DEPARTMENT OF COMMERCE
United States Patent and Trademark Office
Address: COMMISSIONER FOR PATENTS
P.O. Box 1450
Alexandria, Virginia 22313-1450
www.uspto.gov
APPLICATION NO. FILING DATE FIRST NAMED INVENTOR ATTORNEY DOCKET NO. CONFIRMATION NO.
95/000,483 06/09/2009 6815321 051700-5006-US 9039
24341 7590 07/11/2012
MORGAN, LEWIS & BOCKIUS, LLP. (PA)
2 PALO ALTO SQUARE
3000 EL CAMINO REAL, SUITE 700
PALO ALTO, CA 94306
EXAMINER
KIELIN, ERIK J
ART UNIT PAPER NUMBER
3992
MAIL DATE DELIVERY MODE
07/11/2012 PAPER
Please find below and/or attached an Office communication concerning this application or proceeding.
The time period for reply, if any, is set in the attached communication.
PTOL-90A (Rev. 04/07)

UNITED STATES PATENT AND TRADEMARK OFFICE
____________
BEFORE THE BOARD OF PATENT APPEALS
AND INTERFERENCES
____________
CHIMEI INNOLUX CORP.
Requester and Respondent

v.

Patent of LG DISPLAY CO., LTD.
Patent Owner and Appellant
____________________

Appeal 2012-004203
Reexamination Control 95/000,483 & 90/009,482
Patent US 6,815,321 B2
Technology Center 3900

____________

Before HOWARD B. BLANKENSHIP, ALLEN R. MACDONALD, and
JOSIAH C. COCKS, Administrative Patent Judges.

BLANKENSHIP, Administrative Patent Judge

DECISION ON APPEAL

Appeal 2012-004203
Reexamination Control 95/000,483 & 90/009,482
Patent US 6,815,321 B2

2
STATEMENT OF THE CASE
The Patent Owner appeals under 35 U.S.C. § 134(b) from the final
decision of the Examiner to reject claims 7-10, 12, 14, 16-20, and 22.
Claims 1-6, 11, 13, 15, and 21 are not subject to reexamination. Claims 10
and 16 have been amended.
We have jurisdiction under 35 U.S.C. § 315(a). Oral hearing was
April 25, 2012.
We affirm.

Invention
Figure 4C of the '321 patent is reproduced below.

Figure 4C illustrates a process for fabricating a thin-film transistor.
'321 patent col. 5, ll. 56-58.
A metal such as aluminum (Al) is deposited on a substrate 41 to form
a first metal layer 43. A second metal layer 45 is formed from a metal such
as molybdenum (Mo) and deposited on the first metal layer 43. Id. at col. 5,
ll. 59-63.
The second metal layer 45 is preferably patterned to have the width
W2 which is narrower than the width W1 of the first metal layer 43 such that
Appeal 2012-004203
Reexamination Control 95/000,483 & 90/009,482
Patent US 6,815,321 B2

3
W1 - W2 is between about 1µm and 4 µm. Each side portion of the second
metal layer 45 preferably has a width larger than about 0.5 µm and less than
about 2 µm. Id. at col. 6, ll. 17-26.

Representative Claim
7. A method of forming a thin film transistor
comprising:

forming a first metal layer on a substrate,

forming a second metal layer on the first metal layer;

simultaneously patterning the first and second metal
layers to form a double-layered metal gate, so that a total width
of the first metal layer is greater than a total width of the second
metal layer by about 1 to 4 µm.

Prior Art
Kakuda US 5,162,933 November 10, 1992
Nakamura JP Hei 7-297185 November 10, 1995
Oikawa JP Hei 4-50822 February 19, 1992

S. Mayumi, et al., A Novel A1 Edge Cutting Process for
Multilevel Metallization, 1987 Proceedings Fourth
International, IEEE VLSI Multilevel Interconnection
Conference, pp. 78-84 (June 1987).

Appeal 2012-004203
Reexamination Control 95/000,483 & 90/009,482
Patent US 6,815,321 B2

4
Owner’s Contentions
Owner contends that the Examiner erred in rejecting claims 7-10, 12,
14, 16-20, and 22 over the prior art. The Examiner adopted four grounds of
rejection proposed by Requester.
I. Claims 7-10, 12, 14, 16, 20, and 22 stand rejected under
35 U.S.C. § 102(b) as being anticipated by Kakuda.
II. Claims 7-10, 12, 14, 16-20, and 22 stand rejected under
35 U.S.C. § 103(a) as being unpatentable over Nakamura and Kakuda.
III. Claims 7-10, 12, 14, 16, 18-20, and 22 stand rejected under
35 U.S.C. § 103(a) as being unpatentable over Oikawa.
IV. Claims 7-10, 12, 14, 16, 18-20, and 22 stand rejected under
35 U.S.C. § 103(a) as being unpatentable over Oikawa and Mayumi.

ANALYSIS
The Examiners Right of Appeal Notice (“RAN”; mailed
May 18, 2011) is hereby incorporated by reference. We adopt the
Examiner’s findings as our own except for those findings with which we
may express disagreement in the analysis that follows. We will discuss three
of the grounds of rejection in this opinion.
Initially, however, we note that two terms that are relevant in this
appeal -- “hillocks” and “step coverage” -- do not appear to be expressly
defined in this record. The precise definitions are not critical, because the
prior art of record reflects that “hillocks” are to be avoided and good “step
coverage” is to be encouraged. However, for a general understanding of the
subject matter we note the following definitions:
Appeal 2012-004203
Reexamination Control 95/000,483 & 90/009,482
Patent US 6,815,321 B2

5
Hillocks are “bumps that form in a thin film during heating and
cooling,” IC Knowledge - Glossary of Semiconductor Terminology 2012
1
;
and
Step Coverage is “the ratio of the thickness of a film over a step
edge to the thickness in a flat area. Most metallization processes result in a
thinner metal layer over a step than in a flat area.” Id.

Section 102(b) -- Kakuda
The Reference
Figure 8 of Kakuda is reproduced below.

Figure 8 illustrates in section a thin film transistor and its vicinity in
an active matrix structure. The gate line 13 has a structure in which a MoCrx
layer 13b overlies an aluminum layer 13a. Kakuda col. 10, ll. 30-34, 45-47.
The side wall of the pattern has an inclination of approximately
50 degrees such that a defect in a stepped portion on the side wall could be
avoided. Id. at ll. 51-54.

1
available at http://www.icknowledge.com (last visited June 27, 2012).
Appeal 2012-004203
Reexamination Control 95/000,483 & 90/009,482
Patent US 6,815,321 B2

6

The Rejection
The Examiner finds that Kakuda describes the aluminum layer 13a as
having a taper angle of 50 degrees. In view of the disclosed thicknesses of
the MoCr layer and the aluminum layer, the Examiner finds that instant
claim 1 is anticipated because the principles of trigonometry demonstrate
that the first metal layer 13a is wider than the second metal layer 13b by
1.68 µm, which falls within the claimed range of about 1 to 4 µm. RAN at
11-14.
2

Owner does not dispute the Examiner’s basic calculations. However,
Owner contends that the Examiner’s finding that Kakuda describes the
aluminum layer as having a taper angle of 50 degrees is in error. Owner
contests the finding by advancing three arguments. Owner asserts that the
Examiner erroneously assumes that: (1) when referring to a “pattern,”
Kakuda is not referring to the molybdenum-based alloy layer (13b in
Figure 8), but to the entire multi-layered structure (13a and 13b in Figure 8);
(2) the side-walls of the aluminum layer (13a) in this multi-layered structure
are co-planar, continuous, and non-stepped with respect to the molybdenum-
based alloy layer (13b); and (3) the side-walls of the aluminum layer (13a) in
this multi-layered structure are non-curved. App. Br. 12-13.

Kakuda’s “Pattern”
Owner refers to Dr. James O’Connor’s declaration of
February 25, 2010 (“O’Connor Decl.”), contained in the Appeal Brief’s

2
The Examiner’s Answer incorporates by reference the RAN.
Appeal 2012-004203
Reexamination Control 95/000,483 & 90/009,482
Patent US 6,815,321 B2

7
Evidence Appendix, in support of the view that Kakuda fails to disclose that
the aluminum layer has a taper angle of 50 degrees.
An initial problem with the O’Connor Declaration is the hearsay
nature of the testimony. The Declaration contains multiple references to an
“Expert Report” by “Dr. Gary W. Rubloff” that has not been provided as
evidence in this appeal. See O’Connor Decl. ¶ 2.
In any event, Owner refers to and relies on (App. Br. 16) paragraph 22
of the O’Connor declaration for the statement that “[t]he focus of the
Kakuda reference, is the sidewall of molybdenum.” The declaration
continues:
In summary, Dr. Rubloff found “[t]hus, with the
aluminum and molybdenum structure of Experiment II, Kakuda
is clearly concerned with the side wall of the top molybdenum
layer, not the aluminum layer.” Rubloff Expert Report ¶ 173.
Hence, the section of Kakuda cited by the Office Action does
not disclose an aluminum layer with an incline angle of 50°.
The Office Action “does not cite any disclosure of Kakuda to
support [the] conclusions that the 50° angle refers to the
aluminum layer.” Rubloff Expert Report ¶ 173. As a result of
my independent review and analysis, I agree with this
conclusion.

O’Connor Decl. ¶ 23.
At paragraph 21 of the Declaration, Dr. O’Connor quotes and concurs
with a statement from the “Rubloff Expert Report ¶ 172: “I believe that a
close review of the Kakuda reference, as a whole, suggests that the 50° angle
identified refers to the angle of the top molybdenum layer, not the aluminum
layer.” The basis for the conclusion appears to be stated at paragraph 22,
where the “Rubloff Expert Report ¶ 173” alleges that the “focus” of the
Appeal 2012-004203
Reexamination Control 95/000,483 & 90/009,482
Patent US 6,815,321 B2

8
Kakuda reference is the side wall of the molybdenum, pointing to a different
embodiment (“Experiment III”) than that relied upon by the rejection
(“Experiment II”).
We do not have the “Rubloff Expert Report” and its 173 (or more)
paragraphs for evaluation of the quoted statements in context. We are not
persuaded that Kakuda’s disclosure of the 50° angle refers to the aluminum
layer but not the molybdenum layer. The factual basis for the opinion rests
on the allegation that the reference is concerned with the side wall of the
molybdenum but not the aluminum layer in an embodiment that is not that
relied upon for showing anticipation, which is a further detriment to the
credibility of the testimony.
We have also considered the attorney arguments submitted to show
that the 50 degree angle disclosed by Kakuda does not refer to the aluminum
layer 13a in Figure 8 -- i.e., that the “pattern” refers to molybdenum layer
13b alone. We can agree with Owner to the extent that portions of the
reference are not precise in the use of the term “pattern.” However, the
“pattern” in the context of the paragraph at column 10, lines 36 through 54
relates to the formation of a photoresist pattern (l. 40) for etching of the gate
electrodes and related structures. Thus, the “pattern” does not appear to
refer to the molybdenum layer 13b alone.
3
The “pattern” refers at least to

3
We observe that the subject '321 patent refers to multiple “patterned”
layers in the description of the related art (col. 2, ll. 6-28) and that in
preferred embodiments of the invention a single photoresist step is used to
“pattern” both the first and second metal layer simultaneously (col. 6, ll. 6-
9).
Appeal 2012-004203
Reexamination Control 95/000,483 & 90/009,482
Patent US 6,815,321 B2

9
the laminated molybdenum and aluminum films that are patterned by
etching. Kakuda col. 10, ll. 22-29.
Further, Kakuda states that the MoCrx thin film is easy to etch with the
aluminum etchant and that the “laminated” film patterned by etching “has
gently sloping side walls through the substrate” (col. 10, ll. 22-25). We find
it highly unlikely that the “laminated” film with sloping side walls refers to
the single molybdenum layer and does not include the aluminum layer --
especially when the reference, in the same paragraph, describes “laminating
the MoCrx thin film and the aluminum film.” Kakuda col. 10, ll. 25-29.
Further, the “defect” avoided in a step portion of the film (col. 10,
ll. 53-54) could refer, as Owner argues, to a defect in the molybdenum film
(col. 7, ll. 44-51), but that would not preclude the aluminum film from
having an inclination of approximately 50 degrees. Moreover, we find it
more likely that the avoided “defect” refers to improved step coverage as
opposed to a “defect” in the molybdenum film alone. See Dr. Richard B.
Fair Declaration (“Fair Decl.”), in Evidence Appendix of Requester’s
(Respondent) Brief, at ¶¶ 14-16.
Further, we agree that patent drawings generally do not define precise
proportions and may not be relied on to show particular sizes if the
specification is completely silent on the issue. App. Br. 20. See In re
Wright, 569 F.2d 1124, 1127 (CCPA 1977) (“Absent any written description
in the specification of quantitative values, arguments based on measurement
of a drawing are of little value.”); In re Wilson, 312 F.2d 449, 454 (CCPA
1963) (“Patent drawings are not working drawings [and arguments are not
persuasive when based on a] drawing obviously never intended to show the
Appeal 2012-004203
Reexamination Control 95/000,483 & 90/009,482
Patent US 6,815,321 B2

10
dimensions of anything.”). In this case, the Examiner is not taking any
quantitative values from Figure 8 of Kakuda but from the reference’s written
description. However, Figure 8 shows an aluminum layer 13a having
inclined walls. In view of Kakuda’s disclosure that the side wall of the
pattern has an inclination of approximately 50 degrees, a reasonable
inference is drawn that the aluminum layer walls are depicted as inclined to
be illustrative of the disclosed 50 degree inclination.
4

Kakuda’s Al Layer Sidewalls
Owner argues that the aluminum layer 13a sidewalls are curved rather
than linear as shown (Kakuda Fig. 8). However, Owner has failed to
establish (1) that the sidewalls are curved and (2) even if curved, why the
result of the relative width falling within the claimed range would be
different.
The factual basis for Owner’s arguments (App. Br. 17-24) relies on
three documents. The principal document is a “Y.I. Choi et al.” publication
that the Examiner refused to enter because it was a second submission after
the Action Closing Prosecution (Ans. 2). There is no indication in the record
that Owner petitioned the Examiner’s refusal to enter the evidence.
Accordingly, we have not considered the Choi publication. See 37 C.F.R.
§ 41.67(c)(1)(ix) (“Reference to unentered evidence is not permitted in the
brief.”).

4
The subject '321 patent indicates that the lateral surfaces of the second
metal layer 45 (Fig. 4B) are preferably etched to have a substantially
rectangular or inclined shape (col 6, ll. 29-31), and Figure 4B shows an
inclined shape.
Appeal 2012-004203
Reexamination Control 95/000,483 & 90/009,482
Patent US 6,815,321 B2

11
Owner also refers to Mayumi,
5
in particular Figure 7(a), as support for
the conclusion that an isotropic etch “typically” results in curved sidewalls.
App. Br. 18-19. However, the remarks fail to explain how one of ordinary
skill in the art would reach the conclusion based on the disclosure of
Mayumi. Further, Owner does not refer to any expert testimony in support
of the allegation. Moreover, we find no satisfactory response to the
Examiner’s finding (RAN 52- 54) that references of record show that the
relative etching rate for molybdenum can be controlled such that the side
edge portions of the adjacent aluminum film are inclined at particular angles.
Nor do we find a response to Requester’s observation (Respondent Br. 7)
that Seiki (US 5,811,835) discloses non-curved aluminum side walls having
an inclination angle.
The final document that Owner relies on is the Kakuda reference
itself. According to Owner, Kakuda at column 13, lines 29 through 44 tends
to show that it was known that the etch rate of each layer in a double-layer
structure must be identical to form a “non-stepped double-layer structure
with continuous and identical inclination angles across both layers.” App.
Br. 21. However, Owner’s response to the Examiner’s finding that the
relative etching rate for molybdenum can be controlled is based on evidence
that is not before us (“Choi et al.”).

Section 102(b) Summary/Conclusion
Upon consideration of the Examiner’s findings in the RAN and
supporting evidence, and Owner’s evidence and arguments in response, we

5
The reference is applied in present § 103(a) rejections.
Appeal 2012-004203
Reexamination Control 95/000,483 & 90/009,482
Patent US 6,815,321 B2

12
find that a preponderance of the evidence supports the Examiner’s finding
that Kakuda discloses that the width difference of the molybdenum and
aluminum layers falls within the claimed range of about 1 to 4 µm. We
therefore sustain the § 102(b) rejection of claim 7. Claims 8-10, 12, 14, 16,
20, and 22, not separately argued, fall with claim 7.

Section 103(a) -- Nakamura and Kakuda
We refer to the Examiner’s findings in the RAN (at 16-22) in support
of the rejection of representative claim 7 over Nakamura and Kakuda.
Owner advances two arguments in response to the rejection (App. Br. 24-
25).
Owner’s first argument rests on the allegation that Kakuda fails to
teach a relative width of layers within the claimed range, which we find not
persuasive for the reasons noted in our review of the rejection of claim 7 as
being anticipated by Kakuda.
Second, Owner submits that one skilled in the art would not combine
the structures of Nakamura with the structures of Kakuda. Owner’s only
reference to evidence in support of the contention is to paragraph 27 of the
O’Connor Declaration.
We agree with the Examiner and Requester that an allegation that
Nakamura has solved the problems stated in the publication and thus
“Nakamura would not look to Kaduda” (O’Connor Decl. ¶ 27) fails to
demonstrate any defect in the Examiner’s prima facie case of obviousness.
The proper question is “whether the combination was obvious to a person
with ordinary skill in the art.” KSR Int’l Co. v. Teleflex, Inc., 550 U.S. 398,
Appeal 2012-004203
Reexamination Control 95/000,483 & 90/009,482
Patent US 6,815,321 B2

13
420 (2007). “Under the correct analysis, any need or problem known in the
field of endeavor at the time of invention and addressed by the patent can
provide a reason for combining the elements in the manner claimed.” Id.

Section 103(a) -- Oikawa and “Result Effective Variable”
Owner argues in response to the § 103(a) rejection over Oikawa that
the Examiner has failed to establish that the “width difference” of the metal
layers represents a “result-effective variable” that was recognized in the
prior art. App. Br. 25-27.
However, at the oral hearing Owner’s counsel acknowledged that it
was known to adjust the widths for step coverage but argued that the
disclosure of the '321 patent taught a different reason to adjust the widths.
[MR. BREGMAN:] So before we continue, what’s the
Examiner talking about here with this unexpected result? So
he’s saying that -- sorry, with the result-effective variable. He’s
saying that everyone in the art knew that you could just adjust
this variable width to address the hillock problem. Well, that's
simply not true. There’s not a single reference in the record
that discusses that you could adjust the width, the variable
width, to affect or to address the hillock problem. People adjust
the width for other reasons -- for step coverage, for various
other reasons -- but certainly no one appreciated that you would
adjust these variable widths to deal with hillocks.

JUDGE MACDONALD: Now what reason did the
Examiner says it was being adjusted? Did they pick a particular
one?

MR. BREGMAN: Well, he was saying that people
would have known to adjust the range because of step coverage,
which was known, but who’s to say which optimum range?
Appeal 2012-004203
Reexamination Control 95/000,483 & 90/009,482
Patent US 6,815,321 B2

14
That’s not the result-effective variable to avoid hillocks. That's
a result-effective variable to address some other issue.

JUDGE MACDONALD: But is that really relevant
given KSR?

MR. BREGMAN: In what way, Your Honor?

JUDGE MACDONALD: In that he would arrive at this
particular structure for other reasons.

MR. BREGMAN: You see, I don't think you would
arrive at the structure because the result-effective variable for
adjusting the variable widths for step coverage could be a
completely different range. I mean, it's not the range that's
being optimized to address the hillock problem. It could be
bigger than four micrometers. It could be smaller than one.

JUDGE MACDONALD: Yes, but you just said it’s not
the range.

MR. BREGMAN: Well, no one appreciated that if you
found this difference in widths between one and four that you
could avoid hillocks. No one appreciated that. They said you
could have a different step, but that’s for step coverage
purposes.

Hearing Trans. 21-23.
However, Owner’s contention that a range for optimal step coverage
is different than that for preventing hillock formation is contrary to the
'321 patent’s disclosure.
The present inventors have discovered that a relationship
between the width of the first metal layer and the width of the
second metal layer of a double metal layer gate electrode is
critical to preventing deterioration of step coverage of a later
formed gate oxide layer in such a structure having a double step
Appeal 2012-004203
Reexamination Control 95/000,483 & 90/009,482
Patent US 6,815,321 B2

15
difference between the substrate and the gate. More
specifically, the present inventors determined that a structure
wherein the first metal layer 43 is wider than the second metal
layer 45 by about 1 to 4 µm, for example, 1 µm