Ex Parte Middlebrooks et al

Patent Trial and Appeal Board

Mar 29, 2018

13667174 (P.T.A.B. Mar. 29, 2018)

UNITED STA TES p A TENT AND TRADEMARK OFFICE
APPLICATION NO. FILING DATE FIRST NAMED INVENTOR
13/667,174 11102/2012 Scott Anderson MIDDLEBROOKS
26111 7590 03/29/2018
STERNE, KESSLER, GOLDSTEIN & FOX P.L.L.C.
1100 NEW YORK A VENUE, N.W.
WASHINGTON, DC 20005
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
ATTORNEY DOCKET NO. CONFIRMATION NO.
2857.2730001 4302
EXAMINER
KREUTZER, COLIN WRIGHT
ART UNIT PAPER NUMBER
2882
MAILDATE DELIVERY MODE
03/29/2018 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 PATENT TRIAL AND APPEAL BOARD
Ex parte SCOTT ANDERSON MIDDLEBROOKS,
RENE ANDREAS MARIA PLUIJMS,
MARTYN JOHN COOGANS,
and MARC JOHANNES NOOT 1
Appeal2017-004515
Application 13/667,174
Technology Center 2800
Before ADRIENE LEPIANE HANLON, JEFFREY T. SMITH, and
MERRELL C. CASHION, JR., Administrative Patent Judges.
HANLON, Administrative Patent Judge.
DECISION ON APPEAL
A. STATEMENT OF THE CASE
The Appellants filed an appeal under 35 U.S.C. § 134(a) from an
Examiner's decision finally rejecting claims 1-3 and 5-14 under 35 U.S.C.
§ 103(a) as unpatentable over Krukar et al. 2 in view of Yoo et al. 3 We have
jurisdiction under 35 U.S.C. § 6(b).
1 The real party in interest, and the Applicant, is identified as ASML
NETHERLANDS B.V. Appeal Brief dated May 13, 2016 ("Br."), at 3.
2 Richard Krukar et al., "Reactive ion etching profile and depth
characterization using statistical and neural network analysis of light
scattering data," 74 J. Appl. Phys. 3698-3706 (1993).
3 US 2012/0078527 Al, published March 29, 2012 ("Yoo").
Appeal2017-004515
Application 13/667, 174
We AFFIRM.
The claimed subject matter is directed to an inspection method (claims
1-3, 5-12), a lithographic apparatus (claim 13), and a non-transitory
computer program product (claim 14) wherein diffraction pupil images of a
plurality of structures formed on a substrate during a lithographic process are
acquired and at least one discriminant function for the diffraction pupil
images is determined.
Representative claim 1 is reproduced below from the Claims
Appendix of the Appeal Brief. The limitation at issue is italicized.
1. An inspection method comprising:
acquiring diffraction pupil images of a plurality of
structures formed on a substrate during a lithographic process, a
process variable of the lithographic process having been varied
between formation of the structures, the variation of the process
variable resulting in a variation in the formed structures and
consequently the diffraction pupil images; and
determining at least one discriminant function for the
diffraction pupil images, the discriminant function being able to
classify the pupil images in terms of the process variable,
wherein the determining the discriminant function
includes at least one of logistic regression and support vector
machines.
B. DISCUSSION
The Examiner finds Krukar discloses an inspection method as recited
in claim 1 with the exception of the method used to determine the
discriminant function. The Examiner finds Krukar determines the
discriminant function using linear discriminant analysis rather than "at least
2
Appeal2017-004515
Application 13/667, 174
one of logistic regression and support vector machines" as recited in claim 1.
Final Act. 3. 4
The Examiner finds "Yoo teaches several common methods of
statistical classification analysis including linear discriminant analysis, as
well as logistic regression and support vector machines." Final Act. 3
(citing Yoo i-f 25).
The Examiner finds one of ordinary skill in the art would have
understood that logistic regression, support vector machines, and linear
discriminant analysis are functional equivalents. Final Act. 4, Ans. 7; 5 see
also Spec. i-f 5 6 (disclosing that "[ t ]ypical discriminant functions include
inter alia logistic regression, support vector machines and linear
discriminant analysis"). Therefore, the Examiner concludes that it would
have been obvious to one of ordinary skill in the art to modify Krukar "to
include at least one of logistic regression and support vector machines as an
alternative to linear discriminant analysis, as taught by Yoo."6
Final Act. 3--4.
The Appellants argue that Yoo is not analogous art and thus is not
properly combined with Krukar in the§ 103(a) rejection on appeal. Br. 8.
Quoting In re Bigio, 381F.3d1320, 1325 (Fed. Cir. 2004), the Appellants
contend that:
Two separate tests define the scope of analogous prior art: (1)
whether the art is from the same field of endeavor, regardless of
the problem addressed and, (2) if the reference is not within the
4 Final Office Action dated July 7, 2015.
5 Examiner's Answer dated November 3, 2016.
6 See In re Fout, 675 F.2d 297, 301 (CCPA 1982) ("Express suggestion to
substitute one equivalent for another need not be present to render such
substitution obvious.").
3
Appeal2017-004515
Application 13/667, 174
field of the inventor's endeavor, whether the reference still is
reasonably pertinent to the particular problem with which the
inventor is involved.
Br. 10, n. 5. A reference is said to be reasonably pertinent if it "logically
would have commended itself to an inventor's attention in considering his
problem." In re Clay, 966 F.2d 656, 659 (Fed. Cir. 1992).
There is no dispute on this record that Yoo is not from the Appellants'
field of endeavor. See Ans. 6 (agreeing that the two references come from
different fields of endeavor). Rather, the Examiner finds Yoo is relevant to
the Appellants' problem, i.e., "classifying measured data through the choice
of a suitable machine learning algorithm." Ans. 7.
The Appellants, on the other hand, argue:
[T]he inventors were concerned with inspecting "exposed
[semiconductor] substrates to measure properties such as
overlay errors between subsequent layers, line thicknesses,
critical dimensions" (see, e.g., Spec., i-f 38); whereas Yoo is
concerned with predicting the sensitivity of a gastric cancer
patient to an anti-cancer agent using polynucleotides having
specific nucleotide sequences (Yoo, i-f 20).
Br. 13.
The Appellants argue that "it would only be through classic
impermissibly [sic] hindsight that the Examiner would conclude a POSA
[person of ordinary skill in the art] would tum to the medical field taught in
Yoo to solve a lithography problem of the claims."
Br. 13 (emphasis omitted).
The Appellants' argument is not persuasive of reversible error.
Although the non-analogous art test provides helpful insight on the
underlying question of what is "prior art" within the meaning of the statute,
we should not be blind to the reality of the circumstances of the case before
4
Appeal2017-004515
Application 13/667, 174
us. In re Wood, 599 F.2d 1032, 1036 (CCPA 1979). Nor should we adhere
to rigid and mandatory formulas that overly limit the inquiry. KSR Int 'l Co.
v. Teleflex Inc., 550 U.S. 398, 421 (2007) ("Rigid preventative rules that
deny factfinders recourse to common sense ... are neither necessary under
our case law nor consistent with it.").
In this case, there is no dispute that Krukar discloses the inspection
method recited in claim 1 with the exception of the particular method used to
determine the discriminant function. The Examiner relies on Yoo to show
known alternatives to the linear discriminant analysis disclosed in Krukar,
i.e., logistic regression and support vector machines as recited in claim 1.
See Ans. 7 ("Yoo was cited to teach that many machine learning algorithms
were known as suitable functional equivalents of one another for the purpose
of classifying data."); Yoo i-f 25 (disclosing that statistical classification
analyses include linear discriminant analysis, logistic regression analysis,
and support vector machine analysis). One of ordinary skill in the art would
have understood that the statistical classification analyses described in Yoo
are not limited to problems involving cancer treatment but rather are directed
to data manipulation, in general. Ans. 8.
The Examiner finds that the problem to be solved by the inventor in
the step of "determining at least one discriminant function for the diffraction
pupil images" recited in claim 1 7 is "classifying measured data through the
choice of a suitable machine learning algorithm." Ans. 7. The Appellants
do not address the Examiner's finding. The Appellants, however, disclose
that "Figure 5 illustrates an embodiment for classifying captured diffracted
7 Br. 17.
5
Appeal2017-004515
Application 13/667, 174
pupil images of periodic structures obtained using a scatterometer" and
disclose that the discriminant function can be used to classify measured
images. Spec. i-f 55. According to the Appellants, "[t]ypical discriminant
functions include inter alia logistic regression, support vector machines and
linear discriminant analysis." Spec. i-f 56.
Based on the foregoing, a preponderance of the evidence supports a
finding that the combination proposed by the Examiner in the§ 103(a)
rejection on appeal is nothing more than the combination of familiar
elements according to known methods yielding predictable results. KSR,
550 U.S. at 416. Therefore, the§ 103(a) rejection of claim 1 is sustained.
The Appellants do not present arguments in support of the separate
patentability of any of claims 2, 3, and 5-14. Therefore, the§ 103(a)
rejection of claims 2, 3, and 5-14 also is sustained.
C. DECISION
The Examiner's decision is affirmed.
No time period for taking any subsequent action in connection with
this appeal may be extended under 37 C.F.R. § 1.136(a)(l ).
AFFIRMED
6