DISTRICT OF MASSACHUSETTS ___________________________________ ) OCEAN SEMICONDUCTOR LLC, ) ) Plaintiff, ) Civil Action ) No. 20-12310-PBS v. ) ) ANALOG DEVICES, INC., ) ) ) Defendant. ) ___________________________________)
MEMORANDUM AND ORDER
March 28, 2023
Saris, D.J.
INTRODUCTION Plaintiff Ocean Semiconductor LLC (“Ocean”) alleges that Defendant Analog Devices, Inc. (“ADI”) infringes “at least” claim 1 of U.S. Patent No. 6,836,691 (“the ’691 patent”) entitled “Method and Apparatus for Filtering Metrology Data Based on Collection Purpose,” both directly and indirectly in violation of 35 U.S.C. §§ 271(a), (b), & (g) (Count VI). ADI has moved to dismiss on the grounds that (1) Ocean failed to plausibly allege that ADI or its foundry partners, specifically Taiwan Semiconductor Manufacturing Company Limited (“TSMC”), infringed the ’691 patent; (2) the ’691 patent claims involve patent- ineligible subject matter under 35 U.S.C. § 101; and (3) the ’691 patent does not support a cognizable infringement claim under 35 U.S.C. § 271(g). After hearing, the Court ALLOWS the motion to dismiss Count VI on the ground that claim 1 is directed to non- eligible subject matter. The Court also holds that dependent claims 6 and 7 are patent ineligible. Although the Court holds that dependent claims 4 and 5 are patent eligible, there are no
claims of infringement as to those claims. FACTUAL BACKGROUND The following facts are drawn from the Amended Complaint (Dkt. 61) and the ’691 patent (Dkt. 61-6) and are taken as true. I. The Parties Ocean, a non-practicing entity, is the assignee and owner of the ’691 patent at issue. It is a limited liability company incorporated in Delaware. ADI is a Delaware corporation with a principal place of business in Norwood, Massachusetts. It is a semiconductor company that designs, develops, sells, offers to sell, and imports semiconductor products in the communications, internet of things, automotive, computer, and consumer electronics
industries (“Accused Products”). ADI operates semiconductor fabrication plants within the United States to produce the Accused Products. II. Industry Conduct The ’691 patent provides the following background: there is a “constant drive within the semiconductor industry to increase the quality, reliability and throughput of integrated circuit devices, e.g., microprocessors, memory devices[.]” ’691 patent, 1:14-17. One technique for improving the manufacture of semiconductor devices includes using a factory wide control system to automatically control the operation of various manufacturing tools –– such as etch, polishing, and implantation tools --
employed on a semiconductor wafer. These systems run as follows: Often, semiconductor devices are staged through multiple manufacturing tools . . . generating data relating to the quality of the processed semiconductor devices. Pre-processing and/or post-processing metrology data is collected on a regular basis, generally in accordance with a sampling plan, for process control purposes. The collected metrology data is used by the process controllers . . . . Operating recipe parameters are calculated by the process controllers based on [a] performance model and the metrology information to attempt to achieve post- processing results as close to a process target value as possible. Reducing variation in this manner leads to increased throughput, reduced cost, [and] higher device performance[.] Metrology data is also used for other purposes not related to process control. One such use is for fault detection and classification (FDC). Fault monitors apply FDC techniques to identify devices or tools with fault conditions. For example, if a particular device has a critical dimension outside a predetermined range, it is flagged as being defective. The wafer may be reworked, the die may be marked defective, or the wafer may be scrapped, depending on the magnitude and nature of the fault condition. Process tools may be monitored during their processing runs. If an anomaly is observed during the processing, the tool may be shut down for maintenance. The wafers processed by the tool may be flagged for subsequent metrology to determine if the tool anomaly caused a degradation of the devices formed thereon. Again, the suspect wafers may be reworked or scrapped. Id. at 1:46–2:9. In other words, to optimize semiconductor device functioning and production, metrology1 data are automatically collected and either employed by process controllers2 to control a manufacturing process or by fault monitors3 to identify tools or devices with faults; these data can be collected through regular
sampling plans implemented in a facility or for other purposes. See id.; id. at 2:15-17. The ’691 patent presents the manufacturing problem which its inventions are “directed to”: Typically, when a process controller gathers metrology data to update its control model or generate a control action for subsequent processing, it retrieves metrology data related to wafers processed in the tool or tools under its control and employs that data to perform its control task. The data retrieved includes metrology data collected through the regular sampling plans implemented in the facility, and the metrology data collected for other purposes. Some of the metrology data does not accurately reflect the state of the process or the devices manufactured. For example, devices
1 In its technical tutorial, ADI defines “metrology” as the many “measurement steps” included in the manufacturing process. Dkt. 105-1 at 7. In this context, the usual and ordinary meaning of “metrology” is the “study of systems of measurement; the science of measurement; the branch of technology that deals with accurate measurement.” Metrology, Oxford English Dictionary, www.oed.com/view/Entry/117688. 2 In its technical tutorial, ADI defines “process control” as including “[r]eviewing measurements and considering whether and how to update the manufacturing process[.]” Dkt. 105-1 at 8. In its technical tutorial, Ocean defines “process control” as a “method executed in software having at least a sensor interface and a machine interface, allowing for the collection of tool state data and run-to-to-run control of a processing tool.” The usual and ordinary meaning of “process controller” is “a person who or device which regulates and controls an industrial process.” Process Controller, Oxford English Dictionary, www.oed.com/view/Entry/151794. 3 In its technical tutorial, ADI defines “fault” as “abnormal conditions or parameters” that can be detected and classified. Dkt. 105-1 at 8. In its technical tutorial, Ocean defines “fault detection” as a “method by which the data collected during manufacturing indicates that a parameter has exceeded the prescribed operating range.” processed by a tool that was malfunctioning may have characteristics that were affected by the malfunction (I.e., a special cause) rather than by normal process variation (i.e., common cause). Employing this data for use in process control routines may introduce a source of variation that cannot be addressed by the process controller and thus reduce the effectiveness of the process controller. The present invention is directed to overcoming, or at least reducing the effects of, one or more of the problems set forth above.
Id. at 2:10-29.
In other words, metrology data can at times not accurately reflect the state of a manufacturing process or device. Employing inaccurate data in process control routines introduces variation and reduces the effectiveness of process controllers and by extension tools and devices. The ’691 patent addresses this specific problem, among others. III. The Patent The ’691 patent is entitled “Method and Apparatus for Filtering Metrology Data Based on Collection Purpose.” ’691 patent, 1:1-3. In the background section, the “field of invention” is described as “relat[ing] generally to an industrial process, and more particularly, to a method and apparatus for filtering metrology data based on collection purpose in a semiconductor device manufacturing environment.” Id. at 1:8-11. According to the “summary of the invention,” “[c]ontext data for . . . metrology data is generated . . . includ[ing] collection purpose data.” Id. at 2:36-39. “Collection purpose data” is defined as data “indicat[ing] the initial purpose for the collection of the metrology data. For example, the purpose may be process control sampling, fault detection sampling, targeted fault detection, etc.” Id. at 6:18-21. Two overarching descriptions of the method patent are provided: “One aspect of the present invention is seen
in a method for filtering metrology data . . . . Another aspect of the present invention is seen in a system including at least one metrology tool, a computer, and a process controller.” Id. at 2:33- 43. The ’691 patent provides several embodiments, including the following one: In one illustrative embodiment of the present invention, the collection purpose data is used to filter the metrology data for subsequent uses. For example, a process controller would conventionally employ all metrology data for a particular tool and process- operation operation for updating the states of its control model and generating a control action for modifying an operating recipe parameter for the tool. By using the collection purpose data to filter the metrology data, metrology data collected for fault detection purposes, where the likelihood of a fault being present is higher, can be excluded. Filtering the metrology data in this manner may improve the performance of the process controller by removing outlier data that exhibits variation from a source other than normal process variation. If the process controller were to act on metrology data that included special causes of variation (e.g., tool faults), it would attempt to shift the process in a direction that might actually increase variation and reduce the stability of the process.
Id. at 6:22-40. Claim 1 serves as the representative claim proposed by ADI and reads as follows: 1. A method, comprising: collecting metrology data related to the processing of workpieces in a plurality of tools; generating context data for the metrology data, the context data including collection purpose data; filtering the metrology data based on the collection purpose data; and conducting a process control activity related to one of the tools based on the filtered metrology data.
Id. at 8:19-27. Ocean highlights at least four other claims as representative: 4. The method of claim 1, further comprising: identifying a fault condition for a workpiece based on the metrology data; and changing the collection purpose data responsive to identifying the fault condition. 5. The method of claim 1, further comprising: identifying an absence of a fault condition for a workpiece changing the collection purpose data responsive to identifying the absence of the fault condition. 6. The method of claim 1, wherein conducting the process control activity further comprises updating a state of a control model employed by a process controller associated with one of the tools. 7. The method of claim 1, wherein conducting the process control activity further comprises determining at least one parameter of an operating recipe employed by one of the tools.
Id. at 8:39-57. DISCUSSION I. Legal Standard To survive a Rule 12(b)(6) motion to dismiss, the factual allegations in a complaint must “possess enough heft” to state a claim to relief that is plausible on its face. Bell Atl. Corp. v. Twombly, 550 U.S. 544, 557 (2007). “A claim has facial plausibility when the plaintiff pleads factual content that allows the court to draw the reasonable inference that the defendant is liable for the misconduct alleged.” Ashcroft v. Iqbal, 556 U.S. 662, 678 (2009) (citing Twombly, 550 U.S. at 556). “Plausible, of course, means something more than merely possible, and gauging a pleaded
situation’s plausibility is a context-specific job that compels [the court] to draw on [its] judicial experience and common sense.” Schatz v. Republican State Leadership Comm., 669 F.3d 50, 55 (1st Cir. 2012) (cleaned up). Certain documents, like the patent itself, are considered to “merge[] into the pleadings” where the “complaint's factual allegations are expressly linked to” and dependent upon a document, the authenticity of which is undisputed. Trans-Spec Truck Serv., Inc. v. Caterpillar Inc., 524 F.3d 315, 321 (1st Cir. 2008) (cleaned up). The parties dispute whether the Court should consider the expert report attached to the motion to dismiss. Plaintiff cites
dicta in a footnote from a recent Federal Circuit case which suggests “[w]hile supporting evidence, such as an expert declaration or research articles, would further support the existence of an inventive concept, such evidence is not always necessary to defeat a Rule 12 motion.” See Coop. Ent., Inc. v. Kollective Tech., Inc., 50 F.4th 127, 135 n.3 (Fed. Cir. 2022). Ocean urges the Court to consider the expert declaration first attached to its opposition to ADI’s initial motion to dismiss (Dkt. 25-1). Under the law in the First Circuit, the Court declines to consider the expert affidavit, which would transform this motion to dismiss into a motion for summary judgment. See Athena Diagnostics, Inc. v. Mayo Collaborative Servs., LLC, 915 F.3d 743,
755-56 (Fed. Cir. 2019) (holding that the district court did not abuse its discretion in declining to consider the expert report and convert the motion into one for summary judgment). II. Section 101 Patent Eligibility ADI argues that the action should be dismissed because the patent is not directed to patent-eligible subject matter under 35 U.S.C. § 101. See Alice Corp. Pty. Ltd. v. CLS Bank Int'l, 573 U.S. 208, 216 (2014). According to ADI, the ’691 patent is directed to the abstract idea of collecting, analyzing, and manipulating data by filtering, which are not eligible for patenting. See Elec. Power Grp., LLC v. Alstom S.A., 830 F.3d 1350, 1354 (Fed. Cir. 2016); Intell. Ventures I LLC v. Symantec Corp., 838 F.3d 1307,
1313 (Fed. Cir. 2016). Ocean asserts that the ’691 patent is not directed to an abstract idea, but rather a specific improvement to the way computers in semiconductor manufacturing operate. Enfish, LLC v. Microsoft Corp., 822 F.3d 1327, 1336-37 (Fed. Cir. 2016). In the Amended Complaint, Ocean identifies this concept as, “resolv[ing] technical problems related to a process controller collecting metrology data that does not accurately reflect the state of the fabrication process or the device(s) being manufactured . . . overcome[ing] problems specifically arising in the realm of computerized semiconductor manufacturing or fabrication technologies.” Dkt. 61 at 27. Moreover, Ocean argues that claim 1
is not representative of other asserted dependent claims and all the dependent claims have distinctive significance that warrants separate analysis of patent eligibility. It alleges that claims 4, 5, 6, and 7 are representative claims. The post-Alice caselaw explaining patent eligibility is complex and often confusing, particularly in the area of computer software. To begin the Herculean task of digging through the Augean stables of this chore, I begin with the statute: “Whoever invents or discovers any new and useful process, machine, manufacture, or composition of matter, or any new and useful improvement thereof, may obtain a patent therefor, subject to the conditions and requirements of this title.” 35 U.S.C. § 101. “The Supreme Court,
however, has long interpreted § 101 and its statutory predecessors to contain an implicit exception: ‘laws of nature, natural phenomena, and abstract ideas’ are not patentable.” Content Extraction & Transmission LLC v. Wells Fargo Bank, Nat. Ass'n, 776 F.3d 1343, 1346 (Fed. Cir. 2014) (quoting Alice, 573 U.S. at 216). The Supreme Court outlined a two-step analysis for assessing eligibility under § 101. Alice, 573 U.S. at 217-18(involving a patent-ineligible computer-implemented scheme for mitigating “settlement risk”). The first step is to “determine whether the claims at issue are directed to one of those patent-ineligible concepts” such as an abstract idea. Id. at 218. If so, the court must then “consider the elements of each claim both individually
and as an ordered combination to determine whether the additional elements transform the nature of the claim into a patent-eligible application.” Id. at 217 (cleaned up). In step one, “[t]he ‘abstract ideas’ category embodies the longstanding rule that an idea of itself is not patentable.” Id. at 218 (cleaned up). “[I]n applying the § 101 exception, [courts] must distinguish between patents that claim the building blocks of human ingenuity and those that integrate the building blocks into something more, thereby transforming them into a patent-eligible invention.” Alice, 573 U.S. at 217 (cleaned up). If the court answers step one affirmatively, then it proceeds to step two where it “must examine the elements of the claim to determine whether it contains an
inventive concept sufficient to transform the claimed abstract idea into a patent-eligible application.” Id. at 221 (cleaned up). The analysis in Alice assists in understanding the two-step process required in this case. The Supreme Court held that the claims in Alice were directed to the “abstract idea of intermediated settlement,” a “fundamental economic practice long prevalent in our system of commerce.” Id. at 219 (cleaned up). As to inventiveness, the court held that the function performed by the computer was “purely conventional,” and amounted to electronic recordkeeping -- one of the most basic functions of a computer. Id. at 222-25. “The same is true with respect to the use of a computer to obtain data, adjust account balances, and issue
automated instructions . . . ‘conventional activities’ previously known to the industry.” Id. The method claims did not “purport to improve the functioning of the computer itself . . . [n]or [did] they effect an improvement in any other technology or technical field.” Id. Ultimately, such a generic “instruction” to apply the abstract idea of intermediated settlement using some unspecified computer was not “enough” to transform the abstract idea into a patent-eligible invention. Id. at 226 (citing Mayo Collaborative Servs. V. Prometheus Lab’ys, Inc., 566 U.S. 66, 77(2012)). The Federal Circuit has explained: “[A]fter Alice, there can remain no doubt: recitation of generic computer limitations does not make an otherwise ineligible claim patent-eligible.” DDR
Holdings, LLC v. Hotels.com, L.P., 773 F.3d 1245, 1256 (Fed. Cir. 2014). If “[n]othing in the asserted claims purports to improve the functioning of the computer itself or effect an improvement in any other technology or technical field,” then merely adding a computer to the claims does not provide an inventive concept sufficient to make the abstract idea patent eligible. Mortg. Grader, Inc. v. First Choice Loan Servs. Inc., 811 F.3d 1314, 1325 (Fed. Cir. 2016) (internal quotations omitted). Similarly, simply “implementing a mathematical principle on a physical machine, namely a computer, is not a patentable application of that principle. Alice, 573 U.S. at 222 (cleaned up). “[F]iltering content is an abstract idea because it is a longstanding, well-
known method of organizing human behavior, similar to concepts previously found to be abstract.” Bascom Glob. Internet Servs., Inc. v. AT&T Mobility LLC, 827 F.3d 1341, 1348 (Fed. Cir. 2016)(finding “abstract” a claim for a “content filtering system for filtering content retrieved from an Internet computer network”). Thus, patent claims that “merely collect, classify, or otherwise filter data” are ineligible under § 101. See Intell. Ventures I LLC v. Erie Indem. Co., 850 F.3d 1315, 1327 (Fed. Cir. 2017). Even when the basic thrust of a claim is ineligible under step one because it is directed to filtering, the court must proceed to step two of the Alice analysis. For example, in Bascom,
the Federal Circuit found filtering software eligible at step two where the invention involved filtering mechanisms to solve the problem of websites “contain[ing] information deemed unsuitable for some users.” 827 F.3d at 1343. It reasoned that even though “[f]iltering content on the Internet was already a known concept . . . the patent describes how its particular arrangement of elements is a technical improvement over prior art ways of filtering such content” Id. at 1350. The Federal Circuit held that the claims were eligible at step two because they recited a “specific, discrete implementation of the abstract idea of filtering content.” See id. at 1350-52 (“The claims carve out a specific location for the filtering system (a remote ISP server)
and require the filtering system to give users the ability to customize filtering for their individual network accounts.”). In step two, the claim itself –– as opposed to something described in the specification –– must contain the inventive concept. See Two-Way Media Ltd. v. Comcast Cable Commc’ns, LLC, 874 F.3d 1329, 1338-39 (Fed. Cir. 2017) (a claim is ineligible if it only uses generic language to achieve solutions to various technical problems). When a claim directed to an abstract idea contains no restriction on how the result is accomplished and the mechanism is not described, then the claim cannot be patent eligible at step two. Internet Pats. Corp. v. Active Network, Inc., 790 F.3d 1343, 1348-49 (Fed. Cir. 2015) (finding the claims to be
generic data collection steps). After Alice, courts have frequently decided patent eligibility on the pleadings without resorting to additional discovery beyond the complaint and attached documents. See OIP Techs., Inc. v. Amazon.com, Inc., 788 F.3d 1359, 1362 (Fed. Cir. 2015) (affirming district court’s grant of dismissal on the pleadings); Ultramercial, Inc. v. Hulu, LLC, 772 F.3d 709, 717 (Fed. Cir. 2014). The Federal Circuit has affirmed that “[c]ourts may treat a claim as representative in certain situations, such as if the patentee does not present any meaningful argument for the distinctive significance of any claim limitations not found in the representative claim[.]” Berkheimer v. HP Inc., 881 F.3d 1360,
1365 (Fed. Cir. 2018). The Court must consider each of the dependent claims asserted as representative and examine the limitations in each claim. See Gree, Inc. v. Supercell Oy, 834 F. App’x 583, 589 (Fed. Cir. 2020). III. Application of the Alice Two-Step Process to the ’691 Patent ADI argues that claim 1 flunks step one of Alice because it is directed to the “abstract” idea of a method of collecting metrology data, filtering it, and conducting “process control activity” at a “high level of generality”. Dkt. 68 at 20. Under step one of Alice, based on the post-Alice caselaw, claim 1 is directed to the abstract idea of filtering. At step two, claim 1
of the ’691 patent does not contain a specific method of “conducting a process control activity based on the ‘collection purpose data.’” The claim is too general and abstract to add an inventive concept solving a specific technical problem in semiconductor manufacturing. The only specific technical problem mentioned involves improving performance of the process controller by removing faulty outlier data. See ’691 patent, 6:36-40. However, none of the limitations in claim 1 address this problem. Claims 4 and 5 of the ’691 patent, however, survive the Alice guillotine. Anticipating this challenge, Ocean argues that the dependent claims describe more specifically a solution to the
technical problem of overuse of faulty data at the heart of the patent. Dkt. 75 at 10. Claims 4 and 5, which incorporate all the limitations of claim 1, describe identifying the presence or absence of a workpiece fault and changing the collection purpose data in response. ’691 patent, 8:39-48. Construed in favor of Ocean, claims 4 and 5 are devoted to the abstract idea of “filtering” but are saved at step two because they carve out a specific innovation as to the use of fault detection data so that the “collection purpose data” is evaluated based on the presence or absence of evidence of faulty data. As Ocean argues, the ’691 patent resolved problems in manufacturing by conducting process control based on metrology data with faulty data excluded. Dkt. 75
at 7. The specification describes how, if the process controller were to “act on metrology data that included special causes of variation (e.g., tool faults), it would attempt to shift the process in a direction that might . . . reduce the stability of the process.” ’691 patent, 6:36-40. To avoid this potential instability, if metrology data is used in a fault detection analysis and the wafer is determined to be faulty, the “fault monitor may change the collection purpose such that the metrology data would be filtered out for subsequent process control activities.” Id. at 6:44-49. The process of doing this is elaborated as follows: Process control data is collected in accordance with a sampling plan . . . . The collection purpose code for this data is set at “01.” The fault monitor requests metrology data for random FDC [fault detection and classification] oversight. The collection purpose code for this data is set at “02.” The fault monitor may also use the “01” data for FDC oversight. In some cases the fault monitor may identify that a particular tool parameter was outside expected limits during a processing run or that health of a particular tool has degraded to a level indicating a need for maintenance of troubleshooting. The fault monitor may request additional metrology data be collected for wafers processed during the particular process run or by the degraded tool. This targeted fault detection data would have a collection purpose code of “03” [sic] analysis of the “01” and “02” indicates that a fault condition may exist, additional metrology data may be requested. Such additional metrology would have a collection purpose code of “03.” If the fault monitor identifies a faulty die or wafer, it may change the collection purpose code of the “01” or “02” or “03” data to “99.” If no fault is identified for the “02” or “03” data, the fault monitor may change the collection purpose code to “88.” . . . When the process controller gathers metrology data for process control purposes (e.g., state update or control action generation), it filters the metrology data, so that data that is less useful for process control purposes is ignored. For example, the process controller may gather “01,” “02,” and “88” data and exclude the metrology data where fault conditions are more likely to exist (I.e., the “03” and “99” data). Filtering the data in this manner may increase the efficacy of the process controller because non-process sources of variation;[sic] may be removed from the data set used for the process control purpose.
’691 patent, 7:16-62. The Court cannot conclude that, as a matter of law, the specific technical improvement to the manufacturing process in claims 4 and 5 is not inventive. The record, the complaint, the specification, and patent claims 4 and 5 together contain concrete allegations that a limitation of excluding false data was not routine or conventional. Claim 6 describes updating a state of control model employed by a process controller for one of the tools described in the specification; and claim 7 redefines a parameter employed by one of the tools, i.e., with the updated context data and state of control model. ’691 patent, 8:50-57. Ocean does not explain how
dependent claims 6 and 7 solve a technical problem with specificity. Accordingly, they are too general to survive step two. IV. Plausible Claims ADI argues that Ocean does not state plausible claims for direct infringement under 35 U.S.C. § 271, and that the claims of indirect infringement which relate to TSMC, are moot because TSMC entered into a retroactive license. The Court need not resolve these issues because the Amended Complaint only alleges infringement as to “at least” claim 1 which the Court has determined is patent ineligible. ORDER The Court ALLOWS the motion to dismiss the Amended Complaint (Count VI) without prejudice to repleading in 30 days if there is a factual basis for alleging that ADI infringed claims 4 and 5.
/s/ PATTI B. SARIS________________ Patti B. Saris United States District Judge