Arguing Obviousness: Teaching Away versus an Alternative Approach

by Dennis Crouch

Obviousness is the central doctrine of patent law.  It is both the most common reason for rejection and often the most complicated issue because of both factual and legal complexities.  The new Chemours Co. decision provides an important addition to obviousness doctrine in two areas: (1) teaching away; and (2) commercial success.  The majority sided with the patentee and reversed the IPR Board’s single-reference obviousness determination.  The dissent argued that the opinion gives too much power to the teaching-away doctrine.

Chemours Company v. Daikin Industries (Fed. Cir. 2021)

Chemours is a 2015 spin-off from Du Pont. Its patents at issues here relate to a polymer having a high melt flow to allow high-speed extrusion to cover electrical wires.  U.S. Patent Nos. 7,122,609 and 8,076,431.  The claims require a standardized melt flow rate range: 30±3 g/10 min along with other physical properties.

Daikin petitioned for inter partes review, and the PTAB agreed that the claims were invalid as obvious based upon a single prior art reference, U.S. Patent No. 6,541,588 (“Kaulbach”).  Judge Dyk’s dissenting opinion in this case explains that “[t]he only material difference between [Chemours] claim 1 and Kaulbach is that Kaulbach discloses (in Sample A11) a melt flow rate of 24 g/10 min, slightly lower than 27 g/10 min, the lower bound of the 30 ± 3 g/10 min rate claimed in claim 1 of the ’609 patent.”  Actually, Kaulbach goes even further and expreslly discloses a melt flow rate of “15 g/10 min or greater” and includes the example of 24 g/10 min MFR.

On its face, Kaulbach discloses a flow rate that encompasses Chemours’ claimed rate — pretty good prima facie that this element was disclosed, but the Board hung its hat on the particular example and concluded that it would have been obvious to modify Kaulbach’s 24 example so that it fit Chemours’ 27 requirement.

The problem: Kaulbach also focused on having a narrow distribution of molecule sizes within its polymer, and those skilled in the art apparently did not know how to increasing flow rate from 24 to 27 while maintaining the narrow distribution. The Board resolved this tension in favor of an obviousness finding, but you can see that the Kaulbach inventors push against broad molecular weight approach:

[The prior art] mixtures have a very broad molecular weight distribution which according to conventional wisdom, results in an improved extrudability. . . . [However, our invention] has a very narrow molecular-weight distribution. i.e., a ratio of Mw to Mn of less than about 2.

Kaulbach prior art patent.

Teaching Away: On appeal, the Federal Circuit reversed the obviousness determination and held that the Board should have taken more care in considering Kaulbach’s teaching-away from broader molecular weight distribution:

[T]he Board appears to have ignored the express disclosure in Kaulbach that teaches away from the claimed invention and relied on teachings from other references that were not concerned with the particular problems Kaulbach sought to solve. In other words, the Board did not adequately grapple with why a skilled artisan would find it obvious to increase Kaulbach’s melt flow rate to the claimed range while retaining its critical “very narrow molecular-weight distribution.”

Slip Op.  You’ll note that an inventor would not need to shift far from Kaulbach’s 24 g/10 min example flow rate to get within Chemours’ claimed rate that includes about 27 g/10 min.  But, to do so pushes against the fundamental inventive concept of Kaulbach of a narrow molecular distribution. The PTAB’s decision “does not explain why a POSA would be motivated to increase Kaulbach’s melt flow rate to the claimed range, when doing so would necessarily involve altering the inventive concept of a narrow molecular weight distribution polymer.” Slip Op.  An important caveat here is that Chemours’ patent does not actually claim any particular molecular weight distribution, but expert testimony apparently convinced the Board that a broad molecular distribution was required.

In its decision, the majority concludes that the teaching-away negates a motivation to modify the prior art reference.  One oddity of the decision is that the court also spends some time reciting precedent on pertinent prior art, but does not apply that doctrine in any meaningful way.

= = = =

The majority opinion was authored by Judge Reyna and joined by Judge Newman. Judge Dyk wrote in partial dissent.  According to Judge Dyk, the majority placed too much weight on the teaching-away doctrine.

In particular, Judge Dyk reads the the prior art patent as providing an alternative narrow molecular weight distribution while also acknowledging the conventional wisdom and feasibility of having a high distribution. “This is not a teaching away from the use of a higher molecular weight distribution polymer. . . . The majority’s approach impermissibly expands the teaching away doctrine such that it encompasses a reference’s mere preference for a particular alternative. ” Slip Op. Dissent.

= = = =

Secondary Indicia of Nonobviousness: During the IPR, Chemours  presented evidence of commercial success as objective indicia of nonobviousness.  Generally, the commercial success of an invention can be used as circumstantial evidence to prove that the invention is non-obvious.  The basic idea here is that an obvious valuable invention would already be in the market place.  Since there are a number of reasons why any particular product might be successful, patent law generally requires a showing of nexus between the claimed invention and the success — that the innovation (rather than marketing or non-claimed design elements) is responsible for the commercial success.

Here, the Board found the commercial success evidence lacking and thus gave it no weight in the obviousness analysis.  On appeal, the Federal Circuit found three ways that the Board’s decision failed to follow the law.

Nexus and Commercial Success: First, the Board found no nexus because commercial success could not be tied directly to inventive contribution of the Chemours patents.  Here, the Board’s decision was odd.  The Board essentially found that the Chemours patent had no inventive contribution (i.e., had no points of novelty) and therefore could have no nexus.  On appeal, the Federal Circuit rejected that analysis as contrary to law.

The Board found no nexus between the claimed invention and the alleged commercial success because Kaulbach disclosed all features except for the claimed melt flow rate. The Board then found that other prior art of record disclosed melt flow rates of 50 g/10 min.

Contrary to the Board’s decision, the separate disclosure of individual limitations, where the invention is a unique combination of three interdependent properties, does not negate a nexus. Concluding otherwise would mean that nexus could never exist where the claimed invention is a unique combination of known elements from the prior art.

Slip op.

Market Share vs Sales Numbers: The Board also held that market share evidence (rather than merely sales numbers) is required for a showing of commercial success. The Federal Circuit disagreed. “[M]arket share data, though potentially useful, is not required to show commercial success.”

Blocking Patents: Still on Commercial Success, the Board also rejected the commercial success argument after finding that Chemours patents at issue blocked market competition — so that any commercial success would be due to the patent’s existence rather than the innovation’s marketability.   On appeal, the Federal Circuit rejected this conclusion also as contrary to law, finding that the Board totally misapplied the blocking patent doctrine.

The Board erred by misapplying the “blocking patents” doctrine to the challenged patents themselves. A blocking patent is one that is in place before the claimed invention because “such a blocking patent may deter non-owners and non-licensees from investing the resources needed to make, develop, and market such a later, ‘blocked’ invention.’” However, the challenged patent, which covers the claimed invention at issue, cannot act as a blocking patent.

Slip Op. On this point, while I agree that Board’s decision did not follow traditional blocking patent doctrine its approach is interesting.

26 thoughts on “Arguing Obviousness: Teaching Away versus an Alternative Approach

  1. 9

    The inconsistency of the Federal Circuit is staggering. This gets assigned to a different panel and you likely would get an entirely different result (most likely a rule 36 judgment). Most of the panels give carte blanche deference to the PTAB, regardless of how factually/technically errant a PTAB appeal decision or FWD for an IPR may be.

    The PTAB is much worse for nexus and commercial success. In a case I was involved with, the PTAB said the iPhone owed some of its commercial success to advertising and not breakthrough functionality such as mobile Safari – notably, the advertising included an ad that exclusively focused on mobile Safari!

    1. 9.1

      “The inconsistency of the Federal Circuit is staggering.”

      Whew! Thank goodness Congress created the CAFC . . . in order to . . . end . . . inconsistency.

      Thank goodness.

  2. 8

    “the Board also rejected the commercial success argument after finding that Chemours patents at issue blocked market competition — so that any commercial success would be due to the patent’s existence”

    As I recall, the fact that the rest of the industry respects the patent (i.e. allows itself to be blocked by the patent) is also one of the secondary considerations supporting a finding of non-obviousness.

  3. 7

    “On its face, Kaulbach discloses a flow rate that encompasses Chemours’ claimed rate — pretty good prima facie that this element was disclosed…”

    No, no, no, no, no. Go back to patent law 101 instead of teaching it. What Chemours recited is NOT disclosed in Kaulbach – if the PTO thought it was disclosed, the PTO would have made a lack-of-novelty rejection. Kaulbach does not disclose the range recited by Chemours.

    Presumably what you meant to say was that this is “pretty good evidence that the range recited by Chemours was obvious”. That statement isn’t necessarily correct, but at least it’s not clearly wrong.

    1. 7.1

      Take a look at range cases dealing with whether a range describes another range. From a purely math standpoint, I think a skilled person would understand that 15 or greater would have described a value such as 16, 17, etc. The question is, what is the upper limit of what can be considered narrow MWD. Improved extrudability relative to what?

  4. 6

    Here is another case where we have an EPO prosecution file to compare, EP-A-1759396. Its fifth exam report on the merits, issued in July 2015, and despite the USPTO issuing the patent in 2006, sets out that the claims are still so much lacking in clarity that “it is not at present practicable to carry out a full examination”. At that point, Applicant abandoned the application without even filing any “continuing” application ie a divisional.

    I look forward to further intensive scrutiny and debate in cases where the definitive, decisive issue is whether the prior art “teaches away”. I don’t have a feel for it at all.

  5. 5

    The units for MFI-value in Kaulbach are g/min – see image. The MFR units in Chemours are g/10 min. How are they accounting for the missing factor of 10?


          The patentee did not contest that issue. It wrote in the preliminary response: “As Daikin notes, Kaulbach incorrectly reports MFR in units of g/min rather than g/10 min. To avoid confusion, Chemours uses the correct units (g/10 min) in its discussion of Kaulbach.”


            yeah I figured it was just a mistake in the ref, the units would be way off if that order of mag was introduced. But, had the patentee wanted to dispute it that could have been a thing.


              “had the patentee wanted to dispute it that could have been a thing.”

              Had the same thought 6. Thankfully the patentee prevailed anyway, but what if they hadn’t?

              No way I would have conceded that point.

    1. 5.2

      The Kaulbach reference references an ASTM standard for determining the “MFI value”, and that standard produces results with units of g/10 min. So it seems like the omission of the “10” in the Kaulbach discussions MFI-value were drafting errors.

    2. 5.3

      I am sure it’s a typo. Take a look at ASTM 1238 testing standards for polyolefins. I believe they are all /10 min. The melt flow index is the amount measured in grams that exits the die in 10 minutes.

  6. 4

    With Reyna as author, one might suppose that he was listening more to Newman than Dyk (based on subject matter).

  7. 3

    Perhaps Random will weigh in on the (inherent) lack of enablement of the asserted art in that a one-sided range (lower limit given, no upper limit) is de facto non-enabled for the ‘infinity’ side…

    1. 3.1

      Perhaps Random will weigh in on the (inherent) lack of enablement of the asserted art in that a one-sided range (lower limit given, no upper limit) is de facto non-enabled for the ‘infinity’ side…

      Isn’t the claim term “having a melt flow rate of within the range of about 30±3 g/10 min”? Isn’t that bounded?

      But yes as a general principal, if you can’t do something at the margin, you haven’t enabled the whole scope. If there’s no upper limit on a speed, you ought to be able to do it instantaneously to enable it.

      1. 3.1.1

        Thanks – it was the reference rather than the contested claim that supposedly had an unbounded upper limit.


          An aspect here (from your perspective as an examiner): IF the reference
          for enablement due to the unrestrained (no) upper limit, does it then
          as a proper reference?


            anon, I don’t understand. What do you mean by “proper”? Do you mean perhaps that, if a reference is not “proper” for a novelty attack, it is not qualified to serve as the basis of a “proper” obviousness attack?

            I had supposed it to be trite law that a disclosure “X” in a reference “D1” cannot deprive any given claim “C” of novelty if D1 fails to enable that key disclosure X, and further, just as trite, that such a deemed failure by D1 to enable its disclosure X is not enough to disqualify D1 as a contribution to a “proper” obviousness attack on C.

            Or is it in this regard different in the USA from elsewhere in the world?


              What you may consider “trite” is an aspect that I seek to have our examiner contributors confirm.

  8. 2

    “Chemours’ patent does not actually claim any particular molecular weight distribution, but expert testimony apparently convinced the Board that a broad molecular distribution was required”

    Patent law is an embarrassment.

    1. 2.1

      Why? If it embarrasses you, then go find another profession, Mr. Examiner. Geez Louise!

  9. 1

    At first glance, this seems like a “fit for purpose” sanity check of the prior references that should be a part of the Graham Factor analysis.

    If I’m attempting to cobble “puzzle pieces” together, you “violate” the ‘picture’ on the puzzle piece, then the attempt to fit the pieces together should be rejected.

Comments are closed.