American Axle: The Latest Twist on Patent Eligibility

By Jonathan P. Osha

Last month during the AIPPI Congress in London I had the privilege of interviewing USPTO Director Andrei Iancu for a full hour during one of the lunch sessions.  As always, Director Iancu’s comments were thoughtful, candid, and delivered with entertaining flair.  We spent a good part of our time discussing the issues around patent eligibility in the United States under 35 U.S.C. § 101.  He discussed the USPTO’s most recent guidelines on patent eligibility (see our earlier article here) and the Director’s belief that if those guidelines are followed, an equitable balance can be reached between preventing patents on truly abstract concepts and allowing patents that protect and encourage innovation.  When asked by the audience if any court has followed the USPTO guidelines, he gave a simple answer: “not yet.”

As encouraged as I believe all of us in the room were by Director Iancu’s positive comments, the latest § 101 decision from the Federal Circuit not only fails to follow the logic of the USPTO guidelines,[1] it reaches what is, in the opinion of this author, a new low in patent eligibility jurisprudence.  Because, if a method of manufacturing a propeller shaft is not eligible subject matter, it is difficult to imagine where a future line might be drawn.  Beyond the unfortunate outcome for American Axle & Manufacturing, Inc. (“AAM”), the majority’s analysis “smushes”[2] the law of patent eligibility, enablement, claim definiteness, and substantive patentability together into an “eligibility goulash”[3] that serves only to foster further uncertainty in an area of law already fraught with unpredictability.  Indeed, well beyond this one appellee, industry as a whole will suffer from this latest attempt to cast technical (albeit perhaps not adequately described or new) inventions as abstract concepts.

The decision is American Axle & Manufacturing, Inc. v. Neapco Holdings LLC et al., No. 2018-1763 (Fed. Cir. October 3, 2019).  The majority decision was authored by Judge Dyk, joined by Judge Taranto.  A strongly-worded dissent was entered by Judge Moore (more on this later).  The technology at issue was not software, nor business methods, nor financial models, but rather propeller shafts such as drive shafts for vehicles (also known as “propshafts”).  Propeller shafts are mechanical devices (usually hollow metal tubes) that transmit rotational energy from one place (e.g., a transmission or transfer case) to another place (e.g., a differential).  Vibration of propeller shafts during rotation is undesirable due to noise and other factors.  Thus, various methods existed for reducing vibration, including placing of weights to balance the shaft (in the same manner a wheel is balanced) and placing specially designed liners inside the propeller shafts.  The patent at issue, the ‘911 Patent, explains that this vibration actually occurs in three modes:

Bending mode vibration is a phenomenon wherein energy is transmitted longitudinally along the shaft and causes the shaft to bend at one or more locations. Torsion mode vibration is a phenomenon wherein energy is transmitted tangentially through the shaft and causes the shaft to twist. Shell mode vibration is a phenomenon wherein a standing wave is transmitted circumferentially about the shaft and causes the cross-section of the shaft to deflect or bend along one or more axes.

‘911 Patent, col. 1, ll. 44-52.  The disclosure, in sum, was directed to how to manufacture a propeller shaft and liner combination where the liner – which is a hollow tube of cardboard or other fibrous material inserted inside the propeller shaft – is “tuned” to dampen the vibrations of the propeller shaft in the various modes.  As noted by the majority, the ‘911 Patent admitted that it was known in the prior art to alter the mass and stiffness of liners to alter their frequencies to produce dampening.  Id., col. 2, ll. 5-37; American Axle, p. 4.  The lower court had treated claims 1 and 22 as representative of the asserted claims (claims 1-6, 12, 13, 19-24, 26, 27, 31, and 34-36) and the Federal Circuit followed suit.  These claims read as follows:

  1. 1. A method for manufacturing a shaft assembly of a driveline system, the driveline system further including a first driveline component and a second driveline component, the shaft assembly being adapted to transmit torque between the first driveline component and the second driveline component, the method comprising:

providing a hollow shaft member;

tuning at least one liner to attenuate at least two types of vibration transmitted through the shaft member; and

positioning the at least one liner within the shaft member such that the at least one liner is configured to damp shell mode vibrations in the shaft member by an amount that is greater than or equal to about 2%, and the at least one liner is also configured to damp bending mode vibrations in the shaft member, the at least one liner being tuned to within about ±20% of a bending mode natural frequency of the shaft assembly as installed in the driveline system.

  1. A method for manufacturing a shaft assembly of a driveline system, the driveline system further including a first driveline component and a second driveline component, the shaft assembly being adapted to transmit torque between the first driveline component and the second driveline component, the method comprising:

providing a hollow shaft member;

tuning a mass and a stiffness of at least one liner, and

inserting the at least one liner into the shaft member;

wherein the at least one liner is a tuned resistive absorber for attenuating shell mode vibrations and wherein the at least one liner is a tuned reactive absorber for attenuating bending mode vibrations.

The U.S. District Court for the District of Delaware, on summary judgment, found both of these claims, and hence all of the remaining claims, invalid under § 101 as being directed to ineligible subject matter.  The Federal Circuit affirmed.

The majority’s opinion begins with an attempt to identify an “inventive concept” that quickly segues into a discussion of enablement:

According to  the  ’911  patent’s  specification,  prior  art liners, weights, and dampers that were designed to individually attenuate each of the three propshaft vibration modes—bending, shell, and torsion—already existed. ’911 patent, col. 1, l. 53–col. 2, l. 38. But these prior art damping methods were assertedly not suitable for attenuating two vibration modes simultaneously. See id. Thus, the patent identified “a need in the art for an improved method for damping various types of vibrations in a hollow shaft” that “facilitates the damping of shell mode vibration as well as the damping of bending mode vibration” simultaneously. Id. col. 2, ll. 39–43. AAM argues that the inventive concept to which these claims are directed is the tuning of a liner in order to produce frequencies that dampen both the shell mode and bending mode vibrations simultaneously.

AAM urges both that it “conceiv[ed] of the novel and unconventional concept of ‘tuning’ a liner,” and that it conceived of a tuned liner that “unlike previous dampers and absorbers . . . [can] dampen multiple types of vibration” simultaneously. AAM Op. Br. 13. AAM explains that “particular liners that are specifically tuned to match and damp multiple vibration modes and are utilized to manufacture improved propshafts . . . w[ere] entirely new and far from well-understood” at the time of the ’911 patent. AAM Op. Br. 27. Neither the claims nor the specification describes how to achieve such tuning. The specification also discloses a solitary example describing the structure of a tuned liner, but does not discuss the process by which that liner was tuned. ’911 patent, col. 8, ll. 4–23.

American Axle, pp. 6-7.  From this excerpt, we see the blurring of lines between the questions of what is new – novel and non-obvious – and what is adequately disclosed.  These are both valid fact questions that would have been appropriate to consider – albeit not on summary judgment.  But, in fact, these issues were not addressed as such, and instead were used to justify a foray into subject matter eligibility.

The district court had found, “the Asserted Claims as a whole are directed to laws of nature: Hooke’s law and friction damping.” J.A. 10.  Note that neither the term “Hooke’s law” (an equation that relates mass and stiffness to frequency of vibration) nor “friction damping” actually appears in the claims.  Having thus read Hooke’s law into the claims and grossly generalized the specific claimed structures and parameters as “friction damping,” the district court concluded the remaining elements of the claim “consist of well-understood, routine, conventional activity already engaged in by the scientific community . . . and those steps, when viewed as a whole, add nothing significant beyond the sum of their parts taken separately.”  J.A. 16.

On appeal, the Federal Circuit’s analysis begins properly with a recitation of the two-step Mayo/Alice test for patent eligibility:

Our analysis of § 101 follows the Supreme Court’s two- step test established in Mayo and Alice Corp. Pty. Ltd. v. CLS Bank International, 573 U.S. 208 (2014). At step one of the Mayo/Alice test, we ask whether the claims are directed to a law of nature, natural phenomenon, or abstract idea. Alice, 573 U.S. at 217 (citing Mayo, 566 U.S. at 77). If the claims are so directed, we then ask whether the claims embody some “inventive concept”—i.e., whether the claims contain “an element or combination of elements that is ‘sufficient to ensure that the patent in practice amounts to significantly more than a patent upon the ineligible concept itself.’” Id. at 217–18 (brackets omitted) (quoting Mayo, 566 U.S. at 72–73).

American Axle, p. 9.  However, from this point forward the analysis wanders astray.  First, the court characterizes what is disclosed, rather than what is claimed:

The ’911 patent discloses a method of manufacturing a driveline propshaft containing a liner designed such that its frequencies attenuate two modes of vibration simultaneously.

Id.  Then, the Federal Circuit uses this characterization to decide what the claims are “directed to.”  In so doing, the court disregards the various substantive limitations (see the “providing,” “inserting,” and “positioning” limitations of claims 1 and 22) and, citing to the district court’s summary judgment decision,  generalizes the “tuning” aspect of the claims down to a mere concept based upon its perception of what is “new” in the disclosure:

The claims are directed to tuning liners—i.e., “controlling a mass and stiffness of at least one liner to configure the liner to match the relevant frequency or frequencies.” J.A. 15.

American Axle, p. 10.  It appears not to matter that “controlling,”  “mass,” “stiffness,” “match,” and “relevant frequency or frequencies” do not appear in claim 1.  Having thus divined an abstract concept from the disclosure and having written a new “summarized claim” to this abstract idea, one might expect the Mayo/Alice Step 1 inquiry to end there.  But it does not.  The court next follows the district court’s lead down the path of Hooke’s law:

Hooke’s law is a natural law that mathematically relates the mass and/or stiffness of an object to the frequency with which that object oscillates (vibrates). Here, both parties’ witnesses agree that Hooke’s law undergirds the design of a liner so that it exhibits a desired damping frequency pursuant to the claimed invention. For example, Neapco’s expert, Dr. Becker, stated that the tuning limitations claim “nothing more than Hooke’s law . . . [and/or] the law of nature / natural phenomenon for friction damping.”

Id., p. 10.

The claiming of a natural law runs headlong into the very problem repeatedly identified by the Supreme Court in its cases shaping our eligibility analysis. See Mayo, 566 U.S. at 71–73; Parker v. Flook, 437 U.S. 584, 590–95 (1978);

Id., p. 15.  But it’s not in the claim!

Like the claims in Flook, the claims of the ’911 patent are directed to the utilization of a natural law (here, Hooke’s law and possibly other natural laws) in a particular context. As in Flook, where the patent did not disclose how variables were measured nor the means by which the alarm system functioned, the claims here do not disclose how target frequencies are determined or how, using that information, liners are tuned to attenuate two different vibration modes simultaneously. The claims here simply instruct the reader to tune the liner—a process that, as explained above, merely amounts to an application of a natural law (Hooke’s law) to a complex system without the benefit of instructions on how to do so.

Id., p. 19.  In a single paragraph the court characterizes the claims both as being directed to “Hooke’s law” and as being directed to “Hooke’s law and possibly other natural laws.”  Regardless, none of these terms appear in the claim.  At best, they have been read into the claims based on deposition testimony and then conjured as the “abstract concept” at the heart of the claims without regard to the actual wording of the claims themselves.  Having created this straw man, the Federal Circuit then additionally muddies the waters by introducing the issue of enablement – lack of sufficient instructions as to how to carry out the claimed invention.  While this might be a rational basis upon which to consider invalidity of these claims, that analysis would be under 35 U.S.C. § 112, not § 101.

Ironically, when AAM attempted to argue that the claimed tuning is more than simply Hooke’s law, the court responded by noting no other tuning methodology is claimed, citing the need to “examine the elements of the claim.”  This is curious indeed in the context of analysis of a mathematical law that does not appear in the claim at all:

The problem with AAM’s argument is that the solution to these desired results is not claimed in the patent. We have repeatedly held that features that are not claimed are irrelevant as to step 1 or step 2 of the Mayo/Alice analysis. Alice, 573 U.S. at 221 (“[W]e must examine the elements of the  claim  to  determine  whether  it  contains  an  ‘inventive concept.’” (emphasis in original).

Id., p. 13.  This statement is correct and had it been followed it would have precluded reading Hooke’s law into the claim in the first place.

There are two actual questions presented in this case: 1) is the claimed “tuning,” together with the rest of the claim limitations, too broad to distinguish over the prior art; and 2) does the disclosure (not the claims themselves) teach a skilled person how to make and use the claimed invention without undue experimentation?  These two questions are hopelessly intermingled in the court’s analysis and are inapposite in any event to an analysis under § 101.

As a further bizarre twist to this analysis, the court suggests that if the claims had recited more equations or software (i.e., if the “enablement” was recited in the claims) then the claims might have been patent eligible:

…neither the specifics of any novel computer modelling nor experimental modal analysis are disclosed in the patent, much less included in the claims themselves, and these unclaimed features cannot function to remove claims 1 and 22 from the realm of ineligible subject matter. See ChargePoint, 920 F.3d at 766. This case might well be significantly different, if, for example, specific FEA models were included in the claims.

Id., p. 15 (emphasis added).

The casual reader could be excused for being quite confused at this point.  Isn’t it the role of the entire disclosure, not just the claims, to provide an enabling disclosure? (Answer: yes.)  Isn’t it the role of novelty and non-obviousness to determine whether a claim is too broad relative to the prior art? (Answer: yes.) Doesn’t the court have to perform the § 101 analysis based upon what is actually claimed (Answer: yes).  The court pays lip service to these concepts but this is clearly not borne out in the analysis.

Having established the claims as being directed to Hooke’s law (sometimes with other laws depending on the paragraph of the opinion), and determining this to be an abstract concept or natural law, the court then proceeds to Step 2 of the Mayo/Alice test: determining whether the claims recite an inventive concept that is “significantly more than a patent on the ineligible concept itself.”  Here again, the court’s analysis is confused, at best.  First, as noted above, the court belies its adherence to the two-point test by beginning with a conclusion as to inventive concept.  Namely, it starts by deciding the “inventive concept” is the specific manner of tuning, and that this is not adequately disclosed or claimed.  Thus, from this argument, Step 2 dissolves into Step 1.  The court waves aside the remaining limitations of the claims as “no more than conventional pre- and post-solution activity.”  Id. p. 22.  The court similarly waves aside the limitations of the dependent claims (finding waiver on the part of AAM).  Thus, the decision of the district court is affirmed.

The dissent by Judge Moore makes for lively reading.  Rather than discuss it, I will let it speak for itself (selected excerpts):

The majority’s decision expands § 101 well beyond its statutory gate-keeping function and the role of this appellate court well beyond its authority. The majority opinion parrots the Alice/Mayo two-part test, but reduces it to a single inquiry: If the claims are directed to a law of nature (even if the court cannot articulate the precise law of nature) then the claims are ineligible and all evidence of non-conventionality will be disregarded or just plain ignored. The majority rejects the notion that claims which contain an “inventive concept” survive the gatekeeper. In the words of the majority, “it makes no difference to the section 101 analysis whether the use of liners to attenuate bending mode vibration was known in the prior art.” Maj. at 13. I am deeply troubled by the majority’s disregard for the second part of the Alice/Mayo test, its fact finding on appeal and its repeated misrepresentation of the record, in each instance to the patentee’s detriment; all when we are to be applying the summary judgment standard no less.

Dissent, p. 1

The majority’s concern with the claims at issue has nothing to do with a natural law and its preemption and everything to do with concern that the claims are not enabled. Respectfully, there is a clear and explicit statutory section for enablement, § 112. We cannot convert § 101 into a panacea for every concern we have over an invention’s patentability, especially where the patent statute expressly addresses the other conditions of patentability and where the defendant has not challenged them.

Dissent, p. 2

The district court held that the claims at issue are ineligible under § 101 because they are directed to a natural law, specifically, “applications of Hooke’s law with the result of friction damping.” J.A. 11. Even the majority does not agree with the district court that the claims are directed to Hooke’s Law. Instead the majority concludes that the claims are ineligible because they are “directed to the utilization of a natural law (here, Hooke’s law and possibly other natural laws) in a particular context.” Maj. at 19. …. Section 101 is monstrous enough, it cannot be that now you need not even identify the precise natural law which the claims are purportedly directed to.


The claims at issue are directed to methods of manufacturing shaft assemblies for driveline systems for automotive vehicles using liners to reduce specific types of vibration. … As the patentee argues, the dependent claims further narrow the physical characteristics of the liners to be used and their positioning within the drive shaft: “Several dependent claims, for example, recite particular liner materials (e.g., cardboard or paperboard) and structures (helically- wrapped resilient member). . . . claims 12, 13, 19, 26, 27,” Appellant’s Reply Br. at 27.1 I do not see how these claims are directed to a natural law. And even if part one of the Alice/Mayo test was satisfied here, there  is  a  part two. The claims will not be held ineligible (remember § 101 is meant to be a gatekeeper) if the claims contain an “inventive concept.” There are many here, articulated in the claims themselves, about which there exist at least questions of fact which should have precluded summary judgment.

Dissent, pp. 3-4 (emphasis in original).

The tuned liner element is the crux of what bothers the majority in this case. The majority’s true concern with these claims is not that they are directed to Hooke’s Law (because this is clearly a much more complex system not limited to varying mass and stiffness), but rather the patentee has not claimed precisely how to tune a liner to dampen both bending and shell mode vibrations.  As the following quotes from the majority demonstrate, their problem with these claims is not one of eligibility, but rather one of enablement:

  • “Most significantly, the claims do not instruct how the variables would need to be changed to produce the multiple frequencies required to achieve a dual-damping result.” Maj. at 14–15.
  • “[T]he claims’ general instruction to tune a liner amounts to no more than a directive to use one’s knowledge of Hooke’s law, and possibly other natural laws, to engage in an ad hoc trial-and- error process of changing the characteristics of a liner until a desired result is achieved.” Maj. at 15.
  • “The claims here simply instruct the reader to tune the liner . . . without the benefit of instructions on how to do so.” Maj. at 19–20.

Dissent, pp. 10-11.

“[T]o be enabling, the specification of a patent must teach those skilled in the art how to make and use the full scope of the claimed invention without ‘undue experimentation.’”  See Genentech, Inc. v. Novo Nordisk A/S, 108 F.3d 1361, 1365 (Fed. Cir. 1997). There is undue experimentation when “the trial and error required to practice the claimed invention could be unduly laborious.” Old Town Canoe Co. v. Confluence Holdings Corp., 448 F.3d 1309, 1320 (Fed. Cir. 2006). And whether undue experimentation is required is a question of fact. Transocean Offshore Deepwater Drilling, Inc. v. Maersk Contractors USA, Inc., 617 F.3d 1296, 1305 (Fed. Cir. 2010). The majority faults the ’911 patent because the claims themselves fail to describe “how to achieve such tuning.” Maj. at 7. The majority concludes, apparently de novo, that too much “trial and error” would be required to determine how to tune a particular liner to the frequencies associated with a given propshaft. The majority advises that if the claims had themselves mentioned using computer modeling to deter- mine how to tune the liner, it may have made all the difference. Yet, earlier the majority explains that such computer modeling and experimental modal analysis was already used in the prior art. How does adding a limitation to the claims to “use a computer program to figure out how to tune  the  liners”  alleviate  the  majority’s  concern  that these claims are directed to a natural law? Surely, this is the first time adding software to a claim would make it eligible.

Dissent, pp. 12-13 (emphasis added).

Today, the majority concludes that the ’911 patent claims are not eligible because they do not teach a skilled artisan how to tune a liner. The majority holds that they are directed to some unarticulated number of possible natural laws apparently smushed together and thus ineligible under § 101. The majority concludes that the inventive concepts “make no difference.” Section 101 simply should not be this sweeping and this manipulatable. It should not be used to invalidate claims under standards identical to those clearly articulated in other statutory sections, but not argued by the parties. It should not subsume § 112. It should not convert traditional questions of fact (like undue experimentation) into legal  ones. The majority’s validity goulash is troubling and inconsistent with the patent statute and precedent. The majority worries about result-oriented claiming; I am worried about result-oriented judicial action. I dissent.

Dissent, pp. 13-14 (emphasis added).

Thank you, Judge Moore.  But one last question:  what of the USPTO guidelines?  What if they had been applied to this case?

As explained in our prior article, the USPTO guidelines separate Step 1 of Alice/Mayo into two inquiries.  First, do the claims include a limitation that recites “as such” a mathematical concept, certain methods of organizing human activity, or mental processes?  If no, the claim should not be treated as reciting an abstract idea except in rare circumstances. Even if the claim does recite an abstract idea falling in one of these categories (or a law of nature or a natural phenomenon) the claim will not be considered “directed to a judicial exception” in the sense of Step 1 of the Alice/Mayo test “if the claim as a whole integrates the recited judicial exception into a practical application of that exception.”  See 2019 Revised Patent Subject Matter Eligibility Guidance.  Thus, under the guidelines, even if a judicial exception is recited, a “practical application” of that exception will cause the claim to pass Step 1, and Step 2 will not be reached.

The ‘911 Patent claims likely would have been found not to recite an ineligible concept such as a mathematical concept, certain methods of organizing human activity, or mental processes.  Even if they were so interpreted, a method of manufacturing a propeller shaft involving such steps as providing a hollow tube, tuning a liner, and then placing the liner in the hollow tube in a manner to dampen specific types of vibrations is certainly a “practical application.”  Accordingly, under the current state of practice at the USPTO the ‘911 Patent claims if examined today would be considered patent eligible and, if they were found to pass the other requirements for patentability (e.g., enablement, novelty and non-obviousness) they would issue as a patent.  However, under the approach of American Axle, these very same claims are invalid as being not patent-eligible, without even reaching the issues of enablement, novelty and non-obviousness.  Clearly there is dire need of a legislative solution here because, at the end of the day, if you can’t patent a method of manufacturing a drive shaft – what can you patent?

The opinions expressed herein are solely those of the author and do not necessarily reflect the opinions of Osha Liang or any other organization.

[1] The USPTO’s administrative guidelines do not have the force of law and thus are not binding on the federal courts.

[2] In accordance with the lexicon of the dissent, “smush” is a slang term that means “to mix together.”

[3] In accordance with the lexicon of the dissent, “goulash” is a mixed stew with many ingredients.