Canons and Canards: Enablement and Utility in CFMT v. YieldUp

by Dennis Crouch

In Patent Law class today, we started the chapter on “disclosure” that focuses on doctrines of enablement, written description, and best mode as codified in 35 U.S.C. 112(a).

(a) In General.— The specification shall contain a written description of the invention, and of the manner and process of making and using it, in such full, clear, concise, and exact terms as to enable any person skilled in the art to which it pertains, or with which it is most nearly connected, to make and use the same, and shall set forth the best mode contemplated by the inventor or joint inventor of carrying out the invention.

Id. Our casebook begins with enablement, which asks whether the disclosure is sufficient to enable a skilled artisan to make and use the invention to its full scope without undue experimentation.  The first case for discussion was a Judge Rader opinion reversing a district court’s summary judgment dismissal.  CFMT, Inc. v. YieldUp Int’l Corp., 349 F.3d 1339 (Fed. Cir. 2003).  CFMT’s two asserted patents claimed improved apparatus/method for cleaning semiconductor wafers using a specific closed-environment setup.   After filing for patent protection, the inventors were hired by Texas Instruments (TI) to implement a version of the system. Their first try removed visible dirt and grease, but failed to meet TI’s commercial standards.  In particular, laser scan revealed lots of contaminants Id. (“filthy”).  The inventors then took six months of repeated experiments to eventually create a version that worked up to TI’s standard.  The district court ruled that amount of intense experimentation was too much and constituted undue experimentation.  Further, by the time the inventors figured out how to make the system work, they had done enough work for an additional patent application.  I.e., making the system work required additional inventing.

On appeal, the Federal Circuit tacitly agreed that the inventors did a lot of experimentation, but concluded that the district court erred in its focus.  In particular, Judge Rader found that the “invention” to be enabled is limited to the “invention as claimed.” The claims did not require a commercially ready system or a system that removes all contaminants to a TI’s electron microscope requirements.  Rather, the claims require a method for “cleaning semiconductor wafers” (i.e., “removal of contaminants”) without stating any specific standards.  From that perspective, the claimed invention was clearly enabled because it did do some amount of cleaning/removal. The court concluded: “In sum, any meaningful ‘cleaning’ would satisfy the claimed goal of ‘cleaning of semiconductor wafers.'” Id.

The Federal Circuit sees enablement as ultimately a question of law; albeit based upon underlying factual determinations.  Here, the lower court applied the wrong standard for enablement and thus prompted de novo review on appeal.

One aspect CFMT is the link between utility and enablement.  In particular, the enablement requirement also requires that specification enable use of the invention.  If the specification only leads to inoperable machines, then the it has failed the enablement requirement.  However, the inoperability standard is quite law and “applies primarily to claims with impossible limitations.” Id.  Here, the invention actually worked — just not at the level desired by TI.

= = = =

10 thoughts on “Canons and Canards: Enablement and Utility in CFMT v. YieldUp

  1. 4

    Thank you Prof. Couch. Great refresher on the enablement requirement. The case of CFMT v. YieldUp came up right after I graduated and was never mentioned in the training I got at the PTO. I assume the In Re Wands factors still apply and was my favorite rejection in the business methods section at the PTO since so many applications came off paper napkins before getting filed as apps. This is an important requirement still in some arts.

  2. 3

    I wonder when the opportunity arises for a litigant to claim that “shall”, as it appears in sec. 112a , actually means…. “may”, as mentioned in note 9. at the end of Gutierrez de Martinez v. Lamagno 515 U.S. 417. Perhaps “shall” in this statute has meant “may” all along…. 🙂

      1. 3.1.1

        I can’t, schnookums. But maybe you can help me, as I recall there was a Sup. Ct. case in which it was stated that when a statute contains the word “shall” but constitutional concerns arise, then the word “shall” is to be construed as being “may”, and it solves the constitutionality question. Do you know what case that was ?


          I don’t off the top of my head recall that case, but when it comes to the Constitution, the Supreme Court can get — shall we say — creative.

          (And let’s not confuse creativity with correctness, now)


            Thanks. I’m looking at 112a and was pondering whether anyone ever challenged its constitutionality. Of course a begged question is, why would anyone do that ? Pppl have agreed enough on its meaning, for so long, there perhaps never was a reason to challenge it, and maybe its unchallengable. We’ve seen quite a bit of discourse on “written description” in the usual places, so, I can play the game too. Let me invent possibly a new requirement….. the “pointing out” requirement. Haven’t heard of that one before. I wonder whether a litigant might someday fault an asserted patent for failing to “particularly point out” (which seems to be different from distinctly claiming). Certainly, particularly pointing out is different from generally pointing out….. hence, can we concoct a “particularity test” ?

            How would I attack 112a ? IDK, I’d try maybe to find a reason that the word “shall”, should be construed as “may”.

            Fun with words…. 🙂


                haha, if we were having coffee 12 years ago and I quipped: “ya know anon, I see a lot of these patents as being ineligible under 101, for a bunch of reasons and I bet the PTO could issue guidance including dozens of examples, and there’d still be a good number of learned practitioners semi-baffled by it all”, and it would have been understandable that any reasonable patent lawyer would chuckle at such a notion, dismiss the statement as crackpottish. So, mind wandering, in retrospect, it is difficult to resist asking myself “what’s next ?”

                Judge: “I rule in favor of defendant, since plaintiff did not particularly point out the invention, even though claim 1 is broad enough to cover the infringing device, it appears to only cover a subset of something bigger, as defendant’s exhibit 1 illustrates graphically, the actual invention in this case is larger than claim 1, and the applicant never pointed that out.” ok, maybe that stretches it too far…. grasping.

                Riddle: Is it possible to claim an invention in a claim, without particularly pointing out what the invention is anywhere in the spec ?

                I haven’t seen a patent yet, where the inventor wrote: “Particularly, the invention is ……” and that’s probably b/c its profane at some level to tie it down like if/when such lang. has potential to be used against one. Usually, I see the claim itself as being THE particularly pointing out of what’s regarded as the invention, but claims can be uber-broad sometimes (recall my “war on breadth” assertion) so there is a point maybe where a claim becomes so broad, it no longer can be considered to point out with sufficient particularity.

                I don’t think I nailed anything in the above, skirting around something though, I “sense” might be there.

                Its slightly saddening, that the subject patent Professor Crouch cited had to make it to Judge Rader, to set them straight, on what many undoubtedly consider “ancient patent principles.” have a good weekend!

  3. 2

    Off-topic, but every morning of late Law360 brings new reports of Judge Albright transferring cases out of WDTX. It appears that we may have reached a new equilibrium.

  4. 1

    Thanks for posting this, Dennis. Not too long ago, I was trying to remember which case stated that commercial readiness is not required for enablement, and a quick search was not fruitful.

Comments are closed.