Ex Parte Middlebrooks et alDownload PDFPatent Trial and Appeal BoardMar 29, 201813667174 (P.T.A.B. Mar. 29, 2018) Copy Citation 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 Copy with citationCopy as parenthetical citation