CAFC Reaffirms that Doctrine of Equivalents Applies to Ranges

US Philips v. Iwasaki (Fed. Cir. 2007)

Philips sued Iwasaki for infringement of its patents covering mercury-tungsten halogen light. The district court awarded summary judgment of non-infringement to Iwasaki.

Vitiation and the Doctrine of Equivalents: The asserted patent claims a target concentration of 1.6 ± 0.4 × 10-4 µmol/mm3. Based on that claim language, the district court ruled that the claim was “intended to establish the demarcation of boundaries [with] the type of precision that is closely analogous to the metes and bounds of a deed of real property.”  Thus, according to the court, allowing the claim to cover any concentration outside of the clearly claimed limits would vitiate the limitation.

On appeal, the CAFC rejected the notion that an expanded numerical range would vitiate the claim language.

“A reasonable juror could make a finding that a quantity of halogen outside that range is insubstantially different from a quantity within that range without “ignore[ing] a material limitation” of the patent claim.”

This decision conforms to prior decisions that allow equivalents for numerical ranges, but not for other limits (such as a claimed “majority.” Moore).

Notice of Infringement: 35 U.S.C. § 287(a) provides that damages for patent infringement only begin to accrue once the infringer was “notified of the infringement.” (Marking constitutes constructive notice). Notice requires a charge of infringement directed to a specific product, device, or action. The notice must also normally include the patentee’s identity. Here, the infringement letter (sent by “Mr. Rolfes”) did not specifically indicate that Philips was the assignee, but the court found notice sufficient because (1) Philips was correctly listed as the assignee of the patent and (2) Philips had granted the sending party the “responsibility for licensing and enforcing” the patent. Thus, Philips may collect damages from the date of receipt of the letter.


  • Rounding in Claim Construction: “‘1.0’ may be said to have more significant digits than ‘1’ with no decimal point. Because [the claimed] ‘10-6’ and ‘10-4’ are simply the numbers 0.000001 and 0.0001 expressed as powers of ten, the claim language provides no basis for inferring any level of precision beyond the single digit ‘1.’ The way that power-of-ten quantities are used in the specification, discussed supra, confirms that the quantities of halogen described by the claims are not intended to be more precise.” — Why does this matter — Less precision means that the accused product may still literally infringe due to rounding.
  • Approximately: “[T]erms like ‘approximately’ serve only to expand the scope of literal infringement, not to enable application of the doctrine of equivalents.”

20 thoughts on “CAFC Reaffirms that Doctrine of Equivalents Applies to Ranges

  1. 20

    I guess I should add some examples. Mabuchi Motor v Johnson shows the English court (unfettered by any DoE) construing the claim to embrace the Johnson Controls product, while in Munich the court found no literal infringement and no infringement by equivalent. In February 2006, Pozzoli got a finding of infringement in Duesseldorf but the (first instance) German Federal Patents court needs from back then till next year, 2008, to utter something on validity. Meanwhile, in England, the Court of Appeal already this spring has confirmed the High Court’s 2006 Not Valid and Not Infringed. Finally, HP v Waters was a tough case, back in 2002. It must have been, because the English Court of Appeal’s 2002 Decision (end of the road) took a view different from the London High Court a year earlier, on the scope of the claim. It seems to have been just as tough in Germany. Only a month or two ago (2007!) the Supreme Court remanded the 2002 case to the Oberlandesgericht, to do the job of claim construction in line with the way the German Supreme Court thinks that claims should be construed.

    Final thought: if you are doing the judging, at least in Germany, and you know you have to deliver a scope of “Infringement by Equivalent” afterwards, you are inevitably going to write the “literal” scope unnaturally narrow, to give yourself enough wiggle room later. What useful purpose does this artificially narrow “literal” claim construction serve? None. At least, not in Germany. So, let’s get rid of DoE, shall we?

  2. 19

    Posthumous makes a great point. Why apply DoE? I think the answer is that, once you have created the monster, it takes on a life of its own. Judges know it waits in its cave, ready to be woken instantly as soon as needed. They find it intellectually more attractive to 1) construe the strict literal wording of the claim then 2) invoke a DoE to get where they want to. The alternative, namely to look purposively at the claim, in context and through the eyes of the PHOSITA, then define the single line between infringement and non-infringement, is intellectually too demanding, at least for a non-specialist, non-technical district court judge, to get down authoritatively in writing in his Decision.

    For a more powerful demonstration of the mischief of a DoE, compare how Germany (DoE) and England (no DoE) go about construing the identical claim, in the identical European patent, litigated in parallel, in two sister national jurisdictions in Europe. For legal certainty, England wins hands down, every time.

  3. 17

    I’ve found the original cite for ‘purposive construction’ – Catnic Components v Smith-Hill [1982] R.P.C. 183 (H.L.). This decision is reproduced at pp 84-88 of Cornish “Patents and Confidential Information”, ESC, 1985. It doesn’t say it was Lord Diplock, but that’s what I remember, and the actual expression at issue was “extending vertically”. R.P.C. = Reports on Patent Cases, H.L. = House of Lords.

  4. 16

    As a Brit, albeit only registered to practice in the US, my recollection is that “purposive construction” was actually coined by Lord Diplock in Catnic, where it was applied to “substantially vertical”.

  5. 15

    Having read this case a couple times now I can only surmise that Newman was fast asleep at the wheel this time. This is one stinking opinion.

    The worse part is their backtracking on what entity must send the notice of infringement. Note that the statute says not a word about who sends a notice of infringement, only that the infringer be notified. CAFC have to come along now and clean up their own droppings — the 2001 the Lans case — in which they screwed an inventor every which way but missionary and totally contorted 287(a) in the process.

    Lans, who sent an infringement notice to a bunch computer giants was both the inventor and the sole owner of the assignee company. His sending the notice of infringement as the inventor was held not to be sufficient notice. Lans re-writes the statute to say that the owner must send the notice.

    But the owner didn’t send the notice in Philips, either. So the CAFC have to pull their foot out of the Lans mess by saying, essentially, that yeah it’s OK for anybody to send the notice as long as you’re big-time. They pull this stunt by mashing together two different issues: 1) who has to be identified in the notice as the owner; and 2) who has to send the notice. Philips was identified as the owner in the notice, but (as with Lans) the owner did not send the notice. Finally, they try to distinguish Philips from Lans on the basis that the Philips notice included a copy of the first page of the patent. So what? Once you give the infringer the patent number, what else do they need? There is no first page involved when product is marked.

    Lans/Philips creates a double standard, which the CAFC should be embarrassed by. It looks like they’re bending over backwards to help the big money prevail.

    As to the ranges issue, CAFC concludes that 1.0 x 10e-4 is distinct from 1 x 10e-4, in that the former is precise and the latter is fuzzy. That comports with what I was taught in grad school as to significant digits. Consequently, Philips’ range was permissibly fuzzy on both ends. So why apply DoE? That issue was mooted by the fuzziness finding. The soft 1 x10e-4 includes 1.6 x 10e-4 literally. Philips had its latitude built into the fuzzy range.

    I also agree w/ Babel, what does “scope of infringement” mean? Claims have scope which may or may not include an infringing device. Infringement doesn’t have scope. I think we know what they mean, but that loose language could sink a patent.

    Post Humous

  6. 14

    One should keep in mind that Lord Hoffman’s interpretation of claim scope (with which I personally agree)is at one end of a seemingly eternal discussion in Europe. The amendments to the European Patent Covention, which will enter into force one month from now, include changes pointing towards a more generous view of equivalents. Instead of being “determined by the terms of the claims” the scope will “be determined by the claims” (Art. 69 EPC). And Protocol on the Interpretation of Art. 69 (to which UK law refers) will now add a specific reference to equivalents: “For the purpose of determining the extent of protection conferred by a European patent, due account shall be taken of any element which is equivalent to an element specified in the claims.”

  7. 13

    David French writes:

    The British law, which is purportedly followed in Canada, was conclusively laid down by Lord Hoffmann of the House of Lords in the decision relating to Kirin-Amgen v. Hoechst:

    “Purposive construction” means understanding language in accordance with what a reasonable person addressed would understand the author to be using the words to mean.
    “Purposive construction” as applied to patent claims does not mean extending or going beyond the definition of the technical matter for which the patentee seeks protection in the claims. The question is always what the person skilled in the art would have understood the patentee to be using the language of the claim to mean. There is no presumption about the width of the claims. A patent may, for one reason or another, claim less than it teaches or enables.

    By way of contrast, the effect of the Doctrine of Equivalents under United States law is to extend protection to something outside the claims which performs substantially the same function in substantially the same way to obtain the same result.

    There can be no patent protection under UK law which extends protection outside the scope of the claims, so interpreted.

    link to

    A corresponding United States decision which may represent the beginning of a similar tendency is Phillips v. AWH (CAFC 2005, en banc), 415 F.3d 1303 (Fed. Cir. 2005), cert. denied, 126 S. Ct. 1332 (2006).
    link to

    The principal question in that case was the extent to which a court should resort to the patent specification in seeking to ascertain the proper scope of its claims. The conclusion was that claims should definitely be read in the context of the specification.

    In passing, the court cited Bates v Coe 98 US 31, 38 (1878): “in case of doubt or ambiguity it is proper in all cases to refer back to the descriptive portions of the specification to aid in solving the doubt or in ascertaining that true intent and meaning of the language employed in the claims”.

    But otherwise the CAFC does not focus on the intent of a Patentee, as discernible from a patent specification.

  8. 12

    DENNIS, take another look at your thread-starter. I think the Philips claim is NOT directed to a “target” value. That’s the accused embodiment, isn’t it. My reading of the CAFC Decision reveals the integer in the claim as:

    ….at least one of the halogens Cl, Br or I is present in a quantity between ten to the power of minus six and ten to the power of minus four….

    which, to me, is unambiguous in its numbers and its true contextual scope. The way I see it now, scope of Literal infringement is some sort of imaginary inner ring fence inside the area protected by the claim’s logically true boundary fence. Readers might well think that Philips is either very lucky or very competent, to have found the DE reference before rather than after filing the app.

    Or maybe it was a Philips reference anyway.

  9. 11

    Part of the joy of this blog is the privilege to contemplate the law as it might be, not just as it is. The DoE is a judicially created monster, with no warrant in the Statute. As soon as you cut off one tentacle, ten more grow in its place. Here, the DE art reference was known to the Philips patent attorney writing the claims for filing. So there was no need for any narrowing prosecution amendment, so no loss of elasticity of claim scope. A less diligent Applicant, who failed to find DE before filing, would get only the literal scope. Given the chances of finding, prior to filing, ALL the close art DE references, that’s more or less arbitrary justice isn’t it, and not fair to Americans.

    For legal certainty, there should be just one line, between what is inside the protected area, and what is outside. UK has such a line (Lord Hoffmann, in the Kirin case). It can be instructive to overlay the fact matrix of a case in country A with the test for claim scope of country B, and see what happens.

    In UK, one asks: Through the eyes of the PHOSITA, what was the inventor using the language of the claim to mean? Not a hard question to answer. Consider the present case, for example. The trouble with ranges is that some have hard edges and others soft, but the context always reveals which it is. Example A: Bake the leg of lamb in the oven, allowing from 20 to 30 minutes per pound (what if it’s not lamb but lamb’s mam you’ve got in the oven?). OR Example B: The Rules of Baseball stipulate that the ball must weigh between X oz. and Y oz. Don’t you think that the PHOSITA would see the latter range as hard-ended, but the former as soft-ended?

    Now apply the Cooper test cited here by the CAFC. How would a reasonable juror construe the range? Is it “insubstantial” to step outside it? Given the passion shown for baseball, I imagine it is the very reverse of “insubstatial” to stray even a milligram outside the official range of ball weights.

    Now ask what would happen if the District Court, having received evidence from the parties about the way the PHOSITA would understand the patent, had to decide what Philips was using the language of the claim to mean? Same result, no? Common sense, no?

    So, you could abolish the DoE by simply asking the District Court to apply Lord Hoffmann’s test, and lay down one clear bright line between what is inside the claim and what not.

    Only problem is: the court decides literal infringement and the jury equivalent infringement. To me, that’s about as barmy as (in Germany) having a twin track in which a court in Munich is deciding whether a claim is valid while, at the same time, away in Duesseldorf, another court is deciding whether exactly that claim is or is not infringed.

    KIS, I say. Like in England.

  10. 8

    So when I read this case I will learn that the claim at issue was never amended with respect to the concentration range, nor were any other claims that recited broader concentration ranges canceled or amended in the parent application or in any of its progeny … right?

  11. 7

    Haven’t read the case yet so, other readers, shout me down if I’ve got it wrong , but:

    What is claimed is a “target” value. Does it matter what actual concentration was in the accused embodiment that was being traded? What matters is the concentration the infringer was trying to arrive at, when he set up his manufacturing line. If he was a lamp manufacturer who sets his target value in the claimed range, but has poor quality control, he might infringe even if, in the lamps he makes, the concentration is a long way from the target concentration.

    Now i have to read the Decision, and see what it says about “target”.

  12. 6

    I think that certain flexibility in IPRS taht,s why PHILIPS can not win the case. The patented claim itself has some limitation that’s why PHILIPS ca not proof infringement. So, the patent prosecution has a great importance in patent examination procedure.

  13. 4

    Brilliant answer. Yes, I see what you’re saying as far as distinguishing literal infringement from DOE generally.

    But I’m still wondering, isn’t expanding the scope of literal infringement what is done by DOE? Granted, for the reasons you point out, patentee may not be able to raise DOE at all, in which case the question is moot from a practical point of view.

    Maybe I’m hung on the phrase “scope of infringement.” I thought it was the claim that has scope and the accused device either falls within the claim’s scope or it doesn’t. This phrase sounds like it’s saying the infringement has scope and the claim may or may not fall within that scope. What is the metes and bounds of an infringement?? So I’m not sure I have a picture of “infringement scope” in my brain, which is BTW headed for the pub because it’s Friday of a long weekend here in Canada.

  14. 3


    What I think Judge Linn is saying is that the term “approximately”, when used in the context of ranges, says that ends of the ranges (for example, x and y) are not precise endpoints but allow some “flexibility”. For example, “approximately 1” might include 0.8 or something like that. By contrast, as Warren Woessner has indicated, when you talk about DOE, the court had said there’s no literal infringement, so the court must then determine whether there’s anything in the prosecution history that prevents the literal claim scope from being “stretched” to encompass the infringing item (usually the alleged infringing item is the equivalent of the literal claim scope in achieving the benefit or benefits alleged in the patent). Even if there is no prosecution history that prevents such “stretch”, if the alleged infringing item is so different as to not achieve the benefit or benefits alleged in the patent, infringement under DOE will normally not be found. I hope this explanation makes sense.

  15. 2

    If a copyist literally infringes, that’s it. To assert infringement under the doctrine of equivalents in the first place, a patentee has to show that the prosecution history did not create any estoppel that would prevent him/her from going on to argue that the doe should apply to the claim. Some commentators and courts have gone so far as to say that prosecution history estoppel does not play a role in claim construction (which determines whether or not their is literal infringement or not in the first place). This is a fine point though, since the prosecution history can indeed be used to determine the literal scope of the claim. The same prosecution history is then reviewed to determine whether or not the patentee is estopped from arguing equivalents, if there is no literal infingement.

  16. 1

    Approximately: “[T]erms like ‘approximately’ serve only to expand the scope of literal infringement, not to enable application of the doctrine of equivalents.”

    This is going to sound like a dumb question, I know, but I’m going to chance it. . .

    What’s the difference between expanding the scope of literal infringement and enabling DOE?

Comments are closed.