By Jason Rantanen
Trustees of Boston University v. Everlight Electronics Co., Ltd. (Fed. Cir. 2018) Download Opinion
Panel: Prost (author), Moore, Reyna
This opinion provides an example of how Section 112 can function as a commensurability requirement. The court’s final lines say as much:
“Having obtained a claim construction that included a purely amorphous
layer within the scope of the claim, BU then needed to successfully defend against an enablement challenge as to the claim’s full scope…Put differently, if BU wanted to exclude others from what it regarded as its invention, its patent needed to teach the public how to make and use that invention. That is ‘part of the quid pro quo of the patent bargain.'”
Slip Op. at 14 (citations omitted).
I like the opinion because the court’s analysis fits neatly with my conceptual explanation of how enablement analyses are actually performed: it’s a two-step process, with the first step being the articulation of the relevant target and the second asking whether the patentee managed to hit that target. Here, the court defined the target as consisting of one of six possible permutations under the claim construction the patent owner had sought. Unfortunately, the evidence did not support the conclusion that a person having ordinary skill in the art could make that permutation without undue experimentation.
In the Federal Circuit’s words, Patent No. 5,686,738 “relates to the preparation of monocrystalline GaN films via molecular beam epitaxy.” Slip Op. at 4. These films are used in creating blue light LEDs. The plaintiff, Trustees of Boston University (BU), sued defendants for infringing the ‘738 patent. Following a jury trial, BU obtained a verdict of infringement and no invalidity. The district judge subsequently denied defendants’ renewed motion for judgment as a matter of law, in which defendants had argued that the asserted claim was not enabled. Defendants appealed.
Only claim 19 of the ‘738 patent was at issue. That claim states:
A semiconductor device comprising:
Two constructions were key here. The court construed “grown on” to mean “formed indirectly or directly above” and “a non-single crystalline buffer layer” to mean “a layer of material that is not monocrystalline, namely, [1] polycrystalline, [2] amorphous or [3] a mixture of polycrystalline and amorphous, located between the first substrate and the first growth layer.” (emphasis added). It also understood “growth layer” as including within its scope a monocrystalline growth layer. BU did not challenge any of these constructions on appeal.
In the Federal Circuit’s view, collectively these constructions raised six possible permutations for the relationship between growth layer and buffer layer:
- (1) monocrystalline growth layer formed indirectly on a polycrystalline buffer layer;
- (2) monocrystalline growth layer formed indirectly on a buffer layer that is a mixture of polycrystalline and amorphous;
- (3) monocrystalline growth layer formed indirectly on an amorphous buffer
layer; - (4) monocrystalline growth layer formed directly on a polycrystalline buffer layer; (5) monocrystalline growth layer formed directly on a buffer layer that is a mixture of polycrystalline and amorphous; and
- (6) monocrystalline growth layer formed directly on an amorphous buffer layer.
Slip Op. at 6. Of these six permutations, defendants focused their enablement argument on #6: whether a person of ordinary skill in the art could practice the invention with a monocrystalline growth layer formed directly on an amorphous buffer layer.
With the court’s attention focused on permutation #6, the enablement analysis was relatively straightforward. The evidence supporting enablement of #6 was weak: the only technique described in the patent, epitaxy, doesn’t work to grow a growth layer directly on an amorphous structure. BU tried to argue that the patent didn’t actually teach epitaxy, but the court was unpersuaded: “The ’738 patent’s specification is concise—just over four columns of text—and focuses on epitaxy. Indeed, it is saturated with the word ‘epitaxy’ or variants thereof.” Id. at 9. BU also argued that what the patent taught could not be epitaxy “because epitaxy invovles a crystalline layer on top of another crystalline layer,” and an amorphous layer is not “crystalline.” Id. at 10. But regardless of whether one called it epitaxy or not, the problem was that BU couldn’t identify anywhere in the specification that taught how to grow a monocrystalline layer directly on an amorphous layer. Nor was the expert testimony sufficient as it consisted of conclusory statements. Testimony that other people had successfully grown a monocrystalline layer on top of an amorphous buffer layer didn’t help either, as “[s]imply observing that it could be done–years after the patent’s effective filing date–bears little on the enablement inquiry.” Id. at 12.
In light of how the court understood the focus of the enablement issue, BU’s final argument–that the specification enabled 5 out of the 6 permutations and that was enough–didn’t work. The court cited the “full scope” language from precedent here, and the “gap-filling” approach didn’t work here since that concept “is merely supplemental; it cannot substitute for a basic enabling disclosure.” Ultimately, the court concluded that the patent needed to enable all six permutations–a result consistent, the court noted, with the claim constructions that BU itself had sought. In the end, the court saw this as being in the vein of Liebel-Flarsheim: if you get the claim scope you ask for through construction, you’re also going to need to show enablement of that scope.
For enablement-timing gurus, there’s some interesting language in the section about the post-application successes to ponder.
Okay, so when I was chair of the patent group at my firm, some of the older associates would often enjoy m@sturbating themselves in front of the younger associates during our bi-weekly meetings. I never tried to discourage them. Now some people are suggesting that this old business should disqualify me from being the managing partner of the New York Office. Never mind that this all happened in the past! It seems like just another liberal witch hunt if you ask me.
W
T
F
?
Just trying to maintain my sanity while watching the ever-imploding freak show that is the modern Repyuke Party in the US.
Find an appropriate forum for your emotional releases (to put it kindly).
That is NOT this forum, which is supposedly about patent law issues.
Your thinly veiling the item with “patent group” is even more insulting than your straight-up rant as it appears that you KNOW BETTER and took an extra step to dissemble and try to add something extra to try to “make it pass.”
Why this type of C R P continues to run rampant is beyond me.
Patents = politics
You want to escape criticism of the Rep u k k-k e Party? Then help good people working to destroy it, you sniveling pile of dog shirt.
…case in point.
Just like Trump: no filter, no control.
Actually, MM, patents politics. Your narrative is ridiculous. Germany has a strong patent system. German thrives off its innovation. Germany has a strong social welfare system.
Plus in the US we had Jimmy Carter advocate for strong patents.
ALL that Malcolm is attempting to do is his usual “one-bucketing.”
The fact of the matter is that his political rants have NOTHING to do with patent law and he knows it.
As I pointed out above, the fact that he know this is evident in his attempt to paint a thin veneer of “chair of the patent group as a lead in to his rant. This makes his rant all the more egregious.
These political rants have been expunged wholesale previously. It’s not like Malcolm hasn’t been warned. Yet, because Malcolm has appointed himself watcher over the fields of rye (for things political), and Malcolm wants to “one-bucket” anyone that disagrees with any of his views as being “them,” he ignores what has been expunged and even attempts (poorly) to hide his rant.
Why Malcolm (apparently) receives such largess from the editors here when he is the clear number one source of blight is what draws questions as to the non-objectivity of rules application for this blog.
Undoubtedly, it is one of the reasons why other blogs have bypassed this one as any type of “America’s leading patent law source.”
I believe that the saxist MM would allow this at his firm.
Speaking of enablement:
At the entrance to the company’s Dublin, Calif., center recently, an agitated Model X owner was trying to drop his SUV off for repair. A Tesla service agent said he couldn’t leave the car there because the facility was too busy. He could make an appointment to bring it in another time, maybe in a couple of weeks.
“But I don’t want to drive it!” said Kaushal Bhaskar, a software engineer from nearby San Ramon who complained he sometimes couldn’t get the passenger door to open, while other times the door would open up all by itself – including once on the Interstate at highway speeds. “This is a safety concern for me!”
Maybe in 2030 some s00per genius will invent “wherein said robot car comprises a module configured to keep passenger door closed on freeway.” This is really s00per d00per complicated stuff! If only the PTO would grant more patents on logic we’d see progress soooooooo much more quickly.
Right, maximalists? Isn’t that how it’s supposed to work? I’m sure Elan “Total Fraud” Musk spends most of his time pouring over the patent literature for new logic disclosures so he can improve his shirty cars. Because that’s how it works, right?
LOL
Above is what happens when rants and strawmen collide…
““But I don’t want to drive it!” said Kaushal Bhaskar, a software engineer from nearby San Ramon who complained he sometimes couldn’t get the passenger door to open, while other times the door would open up all by itself – including once on the Interstate at highway speeds. “This is a safety concern for me!””
I wonder if he was a leftist and if the car was just trying to protect itself from being sent to gulag.
So can BU seek a re-issue and disavow the claim scope as to embodiment #6? Since it’s ‘impossible’ to do, can we stipulate that the infringing device is not doing the impossible? Isn’t this just another strawman class of invalidity arguments? Consider, ‘means for electromagnetic coupling’ and that ‘covers’ faster than light coupling!!! Ergo – claim is not enabled.
two instant thoughts (not bothering to check the details, which would be necessary for a complete answer):
1) re-issue has time constraints
2) eliminating a limitation (or reducing one) may impact an increase of scope
Chem is not my bag, but I would assume the reason it encompassed impossible embodiments is because the patentee argued for an overly broad construction in order to bring the defendant within the scope, and the defendant either lost claim construction or didn’t fight it, knowing that the result would be an unenabled claim.
Conversely, it is entirely possible, as you say, that the defendant was infringing an embodiment that was correctly disclosed.
I am pretty sure that the defendant was not infringing the “impossible to do” embodiment.
Of course not, but as said above, there are two important claim construction rulings which ultimately resulted in the scope encompassing the sixth embodiment. It could be that those claim constructions were not fought because they were known to result in invalid scopes.
You can’t just “disavow” a scope to further limit a claim when there’s no support for that limitation in the spec – that’s called new matter. All you can do is disavow one of several possible reasonable interpretations of a construction, but then you’re stuck with all the consequences of that new construction. In other words, no iwasthere, you can’t “disavow embodiment six” you can only disavow one of the two claim constructions made. If you could disavow embodiment six, every specification would be patentable under 103 because you could just keep disavowing art that read on your genus until you got to something that had no art.
For example, it could be that defendant performed embodiment three, but that “a non-single crystalline buffer layer” when properly construed does not include “amorphous” (which is why the scope included embodiment six), and as a result of a different claim construction defendant would be non-infringing of a scope that encompasses 1,2,4,5. Conversely, you could disavow “grown on” to no longer include “directly” which would cut out embodiment six, but that would cut out embodiments four and five as well.
“You can’t just “disavow” a scope to further limit a claim when there’s no support for that limitation in the spec – that’s called new matter.”
Good point. Of course, plenty of boilerplate is (and should be) contained in a normal application to account for such contingencies.
And you also have a good point that any “disavowal” has to be precise (based on the claim construction impacts you identify.
“Conversely, you could disavow “grown on” to no longer include “directly” which would cut out embodiment six, but that would cut out embodiments four and five as well.”
You could just more narrowly tailor your disclaimin.
But if you are reading a claim to cover the “impossible to do” – which in reality is the definitional lexicography claim language – amorphous v. growing crystal directly on – (which requires a crystal structure in which to “grow” ergo impossible by definition or if possible then it wasn’t completely amorphous ) – shouldn’t a proper Markman order – construe the claim to be valid – under the law of administrative correctness and presumption of validity – Sort of like my example of electromagnetic ‘coupling’ – faster than light!! Ergo invalid. Isn’t this akin to reading (importing) the specification as ‘definitional’ to claim terms – and then faulting the ‘construed’ claim for written description? or some other 112 requirement?
The only exception to this rule is Ethan Hunt, who is always subject to disavowal.
The embodiment is not impossible. BU actually offered post-filing date evidence of folks who had accomplished that embodiment. The experts merely agreed that the embodiment in question was impossible at the time of filing.
In any event, I believe that you are correct that BU could seek a re-issue. As the opinion notes, however, BU fought for a construction that read on this embodiment, which leads me to suppose that at least some of the accused devices must fall into this non-enabled category within the claim scope. In other words, a re-issue claim along the lines you suggest will give up at least part of their infringement theory.
It would likely require a broadening reissue to remove a limitation from the claims, even if the limitation was “impossible.” Anyhow, reissue is not available because the patent has expired.
[T]he patent has expired.
Ah, yes, well that does rather limit the possibility for reissue.
In view of Random’s valuable point in 5.2.1.1 above, I should refine my response to I Was There and say that BU can seek a re-issue if they have written description support for their narrower claim.
Narrower claim? I was not paying attention to the details. I concur, add a limitation to preclude the impossible embodiment, if supported. But as noted, the reissue ship has sailed.
“The experts merely agreed that the embodiment in question was impossible at the time of filing.”
Did the experts agree to that? I thought they would have to disagree or else this whole thing wouldn’t even have been an issue.
Yeah, the claim also covers “a buffer layer having negative mass”, which is obviously not enabled.
True that.
“Isn’t this just another strawman class of invalidity arguments? ”
No because sometimes such embodiments become enabled before the end of the patent’s lifetime. Like the people that testified have already noted. But to get a patent that covers those embodiments they should have had to enable it at time of filing (or else make the claim scope match what was enabled).
Truthfully this is mostly just a drafting problem though. And drafting issue. Further it would have been nice if the examiner had found the issue and made sure it was addressed, as I’ve had to do before.
“Consider, ‘means for electromagnetic coupling’ and that ‘covers’ faster than light coupling!!! Ergo – claim is not enabled.”
We try to be reasonable with the requirement. This case itself is on the cusp of whether it was reasonable or not to require them enable that specific kind of “at the time impossible” embodiment. Truthfully what should have happened was an adjustment to the claim, but they wanted to go broad.
Wait what? A subsequent discovery by another after your claimed date enabled one of your claimed embodiments – that was not otherwise enabled at your priority by your spec or the POSITA date? Isn’t that the thrust of the Morse case? “means for electromagnetic communication” by disclosed hard wired telegraph cables – can’t read on FM radio? RDOE?
Right, which is why BU was rightly shot down on this claim. I agree with you that (had their patent not already expired) they should be allowed to pursue a reissued claim to resolve the enablement issue. There is, however, nothing “strawman” about the argument for invalidity here.
BU claimed a claim that read on subject matter that they had not enabled, and which by common consensus of both sides’ experts was not enabled by the prior art at the time of filing. BU tried to present evidence that other people, after BU’s filing date had been able to practice that particular embodiment, and thus to make a sort of no-harm-no-foul argument. When that failed, they tried to argue “well, the claim reads on six classes of embodiments, and we enabled five out of those six, so close enough for government work.”
BU lost on both of those arguments, and they deserved to lose. Just because someone else manages later to make possible that which you thought up but did not enable, should not entitle you to the claim. Likewise, if you enable some subject matter within your claim scope, but fully 15% remains unenabled, then you are entitled to claim the 85% that you have enabled, but not the 15% that you do not enable. It is not at all a strawman to say that BU was trying to do that which the patent code does not permit.
Well i confess to not being into the weeds of the claims – but wasn’t it a claim construction – Markman – that construed the claims? But yes, I agree that if the patent owner demanded this construction – and won – then they had the enablement argument coming as just deserts. Law of the case and estoppel. And if this construction was necessary for infringement – then it was a fatal flaw in the case.
Question:
The claim construction can also give rise to the permutation:
7) A defect-free monocrystalline growth layer formed directly on an amorphous buffer layer.
This is obviously not enabled, because a layer can never be free of defects (entropy). You can play this game for probably any materials patent that exists. You can always come up with some impossible configuration and say the claim encapsulates it, and therefore goes beyond the scope…. Is this reasonable?
Sorry Dennis (and Judge Prost), this decision is just wrong.
It’s always possible for an accused infringer to gen up a strawman nonsense inoperable embodiment that falls within the claim. For example, for every semiconductor case, let’s consider a chip doped with bismuth or roentgenium (formerly unununium). Nope, whattaya know, it’s impossible. Does that make nearly every single semiconductor claim invalid? On Dennis’ and Judge Prost’s reading, yes, it does. Even though no engineer would ever think about this, the only reason it ever comes to thought is the litigant.
You mean we now have to go through every imaginary and stupid thing anyone could ever conceive, and disclaim them all?
A claim is supposed to give fair notice for infringement. “Fair” means “fair,” not (in this case) atom-by-atom analysis of every conceivable nonsense case that might not work.
This decision isn’t wrong, it’s a leaving of senses.
A slight miss here David, as it was the patent owner who attempted the “gin up” because of the patent owner’s push for a certain claim construction.
In light of how the court understood the focus of the enablement issue, BU’s final argument–that the specification enabled 5 out of the 6 permutations and that was enough–didn’t work. The court cited the “full scope” language from precedent here, and the “gap-filling” approach didn’t work here since that concept “is merely supplemental; it cannot substitute for a basic enabling disclosure.”
While I don’t generally agree with MM, I do agree that most granted patent scopes in software are junk, and it’s largely because of the logic in cases like this. Compare a demonstrated 5 of 6 ways being insufficient here, to 1 or 0 of countless ways of achieving a functional result claimed in your standard computer patent.
As Ben says below at 1: It’s a shame that software doesn’t get held to this standard.
Not because the law is off, but because the Office refuses to apply the law. It is literally clear error to tell applicants that their specification cannot substitute for an enabling disclosure or that the specification itself must demonstrate possession of the full claim scope.
According to the office you could, for example, posit estimating trip time by using GPS to receive coordinates and then dividing the distance by 30mph, label that an “example embodiment” and then summarily sweep google, waze, etc out of the field by just describing your algorithm in a vague manner (“an algorithm that estimates driving trip time”). Even that gives the office too much credit, because you disclose no algorithm at all, and simply say “the processor estimates driving trip time” and you’d still get a pass.
“While I don’t generally agree with MM,”
I cannot recall a single instance of such.
MM appears to categorically state that computer software is logic and logic is either not eligible or always obvious.
I believe that software can be eligible and non-obvious when the specification discloses and the claim claims an algorithm which was beyond the ordinary skill in the art to code. That’s a teaching that improves the art of coding.
I disagree with DDR on the merits, but I agree with the basic logic of that case – disclosure of a non-obvious means of solving a computer problem is eligible and patentable. I disagree with Enfish and McRO on claim construction, but given that claim construction those cases were correctly decided.
Those cases are distinguishable from 95% of the software cases, which a) rely upon prior skill in the art to enable what they claim is novel functionality, which is just legally self-contradictory and impossible or b) may have an invention in their specification but then ruin it by claiming broad functional results in their claim.
The fact that there is a patentable invention hiding in the spec doesn’t make any broad scope claimed patentable. The fact that “good” lawyers always claim those broad, unpatentable scopes leads to MM and me agreeing in result a lot of the time – because we evaluate claims, not specs – but doesn’t mean MM and I agree in general.
If MM took highschool math, he’d lose a lot of points for getting to the ultimately right answer by using a bunch of incorrect intermediate steps.
MM appears to categorically state that computer software is logic and logic is either not eligible or always obvious.
That’s pretty close. The obviousness issue arises when one discusses improvements to existing logic instructions. There are no “unexpected results” that flow from the addition of an “if-then” statement to existing software because the “modified” computer … follows the instructions (because … it’s a computer). Nor is there ever any teaching in the art saying “don’t add an additional ‘if-then’ step to the logic.” Data processing “problems” immediately suggest the “solutions”, at least at the incredibly abstract level that those “solutions” are claimed in patents (actually disclosing working code in different operating systems such that the “solution” would be realizable by an instruction-writer or an end user … well, that would be asking far, far too much of the Most Important People Ever; better to just create an effective tax on everyone and ridiculously inflate the value of functionally-claimed g@ rb@ge!).
It’s not accurate to say that every new logical operation is obvious. It is accurate to say that nearly all of them are incredibly obvious, and it is accurate to say that the PTO appears utterly incapable of coherently determining which one’s are and which one’s aren’t because the PTO confuses the context of the logic with the logic itself, and it is accurate to say that the PTO’s efforts to standardize/formalize the manner in which logic innovations are claimed and evaluated are undetectably weak.
I missed in your reply whether or not you agree with Malcolm that software IS logic.
If you do so believe, maybe you can help him with his copyright project, given that one cannot copyright logic.
I missed in your reply whether or not you agree with Malcolm that software IS logic.
I agree with In re Sherwood which said “In general, writing a computer program may be a task requiring the most sublime of the inventive faculty or it may require only the droning use of a clerical skill. The difference between the two extremes lies in the creation of mathematical methodology to bridge the gap between the information one starts with (the “input”) and the information that is desired (the “output”).”
Depending on the act to be performed, the claim to software performing the act may be little more than the logic of the act itself, or it may be significantly more than just logic. When the element is software which only requires “the droning use of clerical skill” it may be considered logic from the standpoint of eligibility and obviousness, because “clerical skill application” is not enough to impart eligibility or non-obviousness. When the element “requires the most sublime of inventive faculty” it cannot be considered mere logic for the purposes of 112 such that it is always described and enabled over the entire breadth of its scope.
It’s much like the Rosetta stone. Translating from one language to another adds absolutely nothing when the ability to translate is known, and adds absolutely everything when the ability to translate is unknown.
Cute – but does not answer my question.
(looks like a bunch of “caught” posts were release today)
This In re Sherwood?
link to en.wikipedia.org
Your past postings do NOT align with your statement here that you agree with the case.
To give an example I like – I think that if Google tried to claim the algorithm for selecting which route to use and determining how long it would take for their Google Maps driving directions functionality, I think that would be both patentable and eligible, despite the fact that it is iffy because a computer is just processing information and Alice is a threat.
The fact that *that algorithm* would be eligible and non-obvious does not mean a claim to the functionality (“determining a best route and determining trip time”) is eligible and non-obvious. A claim to the functionality is definitely not eligible under Alice, since people attempted to calculate a best route and trip time prior to computers being involved with it. A claim to the functionality is also obvious because the specification cannot possibly disclose all means of making those determinations, and given that the enablement preexisted its use to achieve its result is obvious (as is always the case in computer logic – functionality always achieves its useful function).
I can’t help it that “good” lawyers always submit the functional language to me, or that most cases involve the functional language. I recognize neither I nor anyone else at the office will ever examine a claim limited to what Google invented – their particular algorithm. I understand why Google doesn’t want to disclose their algorithm. And I understand why there is little commercial value in disclosing something that can be relatively easily designed around and the market will not punish someone for putting out a decent-yet-inferior product. That doesn’t change the law. The fact that a computer makes designing around easy doesn’t make an inventor working in that field into a super-inventor, because one of those design-arounds may be superior to Google’s algorithm (and if you replace Google with Average Guy, it becomes clear that’s the case).
In a more perfect world, there would be lower costs for a software patent to encourage the disclosure of less-commercially-valuable easily-designed-around algorithms. Just because there is not doesn’t make overbroad commercially valuable scopes valid.
Great points, particularly this one:
as is always the case in computer logic – functionality always achieves its useful function
which I made independently upthread before I read your comment. It’s true we don’t agree on everything. The fact that we agree on a lot makes the disagreements more interesting. Keep up the commenting.
Is it time (again) for another lesson in the abundance of “functional claiming” outside of the computing arts?
Calling David….
Or is it time to remind Random of the Vast Middle Ground that differentiates (and provides insight into allowed) “functional claiming” from his attempted “boogeyman”…
Is it time (again) for another lesson in the abundance of “functional claiming” outside of the computing arts?
Calling David….
Usually we save the big laughs for Friday but this week is turning out differently.
More of that Accuse Others (especially seeing as it was in an attempt by you to hold an actual discussion with David that you provided one of your biggest blunders ever by volunteering an admission against interests in stating that you knew and understood the controlling law of the exceptions to the judicial doctrine of printed matter and how software fit into that exception…
Schizo-what and your penchant for Accuse Others collides (yet again)
You also miss the physical impossibility of the particular 6th of 6th way… (which is the point I presented below to Ben, whose comment you affirm, but which affirmance also carries the misconception.
H 8ters are going to h 8te.
You also miss the physical impossibility of the particular 6th of 6th way… (which is the point I presented below to Ben, whose comment you affirm, but which affirmance also carries the misconception.
Sure, I ignore that, because I hope its not news to you that proving literal impossibility is sufficient to prove non-enablement of an embodiment. If impossibility were insufficient to prove non-enablement, there would be no non-enablement.
What may be news (especially to those like you who enjoy condoning functional claiming) is that in a situation where the claim covers six possible permutations, the specification enabling five of them is not enough. What is your response to a situation where the specification only enables zero or one permutations for a scope that is much larger than six permutations, i.e. the standard situation in functional computer claiming.
What’s the response with respect to Written Description, which applies the same “must be met over the whole scope” logic as enablement, and is much easier to show that the spec doesn’t have? You have to bring in an expert to say that the sixth embodiment was impossible in this case. You don’t need an expert to say that a specification fails to posit each means of achieving a broad class of results, you just need the ability to read.
Your news is a sub-point to my point (and is not only “not news” it is hardly revelational. Admitting that you do not meet all the (6) items specifically claimed, but somehow that you are “close enough” just is not the argument you want to have been made. Your error is in wanting that sub-point to be more than what it is.
Reply snagged in the “do not have a dialogue” count filter.
“The evidence supporting enablement of #6 was weak: the only technique described in the patent, epitaxy, doesn’t work to grow a growth layer directly on an amorphous structure. ”
Not technically correct, though if the “growth layer” is mono grown via epitaxy on amorph then it will have problems in the layer that results. Still, it could be done, thought the layer may be useless to make devices in (which might make it “not enabled” under the law, but that is a separate inquiry as to whether the layer could be made). Whether it could be done, or reasonably figured out in 1993 tho is a bigger question.
“BU tried to argue that the patent didn’t actually teach epitaxy, but the court was unpersuaded: “The ’738 patent’s specification is concise—just over four columns of text—and focuses on epitaxy.”
They should have argued that the claim wasn’t limited to epitaxy. Which it isn’t. And they’re correct.
BU perhaps shoulda won. Specifically because you could make the upper mono layer not by direct epi on the amorph layer but by making the upper mono layer on a different substrate, and then transferring it over, or still yet other methods (some of which might even make a decent layer for making devices if you’re bombtastic at the art). Still, I guess looking at the filing date it isn’t exactly clear that it was enabled as of then.
“BU also argued that what the patent taught could not be epitaxy “because epitaxy invovles a crystalline layer on top of another crystalline layer,” and an amorphous layer is not “crystalline.””
Sounds like their lawyer probably made that argument and he’s a ta rd.
““The ’738 patent’s specification is concise—just over four columns of text—and focuses on epitaxy. Indeed, it is saturated with the word ‘epitaxy’ or variants thereof.” ”
That’s because epi is a primary embodiment, and which they would have used as an example in the spec. The rest of the methods would be more exotic, but still valid and enabling from other areas of the art. Still, at the date of 1993, whew might be hard to find hard evidence. Though I wouldn’t be surprised if it couldn’t be found.
“But regardless of whether one called it epitaxy or not, the problem was that BU couldn’t identify anywhere in the specification that taught how to grow a monocrystalline layer directly on an amorphous layer.”
They don’t need to if that is enabled by more exotic methods. Which it is. Though the question is about what was up as of the filing date. They should have been able to find evidence of such if it was enabled back at that filing date. Thus the problem here appears to be mostly related to the filing date, which is super far back like 1993. There isn’t a real question as to whether or not it is in fact possible to make, because that is answered in the affirmative easily (nowadays).
Aug. 30, 1993.
Shame really. Though it might be right result.
“In light of how the court understood the focus of the enablement issue, BU’s final argument–that the specification enabled 5 out of the 6 permutations and that was enough”
They need to start sanctioning people for claiming they can enable just part of the claim and get it. That stuff is confusing and could slip by the office’s personnel if the attorney makes such an argument and makes it confusing enough. They need to make clear that attorneys need to stop arguing such its directly agin the lawl.
“Ultimately, the court concluded that the patent needed to enable all six permutations–a result consistent, the court noted, with the claim constructions that BU itself had sought.”
It was the correct claim construction anyway.
“Not technically correct, though if the “growth layer” is mono grown via epitaxy on amorph then it will have problems in the layer that results. ”
You can’t perform epitaxy on amorphous materials. Epitaxy requires matching the crystal structure of the grown material to crystal structure of the underlying material; amorphous materials don’t have a crystal structure/the order is on the order of nanometers.
“Specifically because you could make the upper mono layer not by direct epi on the amorph layer but by making the upper mono layer on a different substrate, and then transferring it over, or still yet other methods.”
Emphasis on “grown on” in the claim. Epitaxial liftoff + layer transfer onto an amorphous layer wouldn’t be “grow[ing] on” the amorphous layer.
“Epitaxy requires matching the crystal structure of the grown material to crystal structure of the underlying material; ”
Thats old school bruh. Just not true in all cases now. See page 3-4 below, tho they use GaAs instead of GaN. People do GaN occasionally too tho I don’t have an immediate ref. on hand. They just do a little trick to get the growth layer started growing. Like put an amorph blob (can be small blob) on amorph sub, recrystallize the amorph blob, and use the recrystallized blob to start growth. That’s one trick, but there are many others.
link to link.springer.com
Recrystallization of films is not epitaxy, which is explicitly why in the article they refer to it as “artificial epitaxy,” as opposed to real epitaxy.
“Recrystallization of films is not epitaxy, which is explicitly why in the article they refer to it as “artificial epitaxy,” as opposed to real epitaxy.”
The epi comes after the recrystallizing. Artificial epi is in the genus epi because actual epi takes place (generally lateral and/or then in other directions). And what you just stated is explicit is is not explicitly in the article.
Further the layer just needs to be grown, and for #6 directly on the substrate, it doesn’t have to be direct crystal/crystal epi using the substrate surface as the seed.
Right, missed “indirectly on,” just saw “directly on.”
“Emphasis on “grown on” in the claim.”
Yeah that was my bad when typing the last comment, but there are still ways other than that as well. Many ways nowadays.
Funny thing is, semiconductors is the simplist of the electrical arts.
They should have argued that the claim wasn’t limited to epitaxy. Which it isn’t. And they’re correct.
How would that have helped BU? If the claim reads on an epitaxial growth embodiment, then they need to have enabled that epitaxial growth embodiment. It is beside the point to say that the claim reads on other things aside from epitaxial growth embodiments.
If they have enabled some of those other embodiments, then they need to limit their claims to the embodiments that they have enabled. One is not permitted to claim more than one has enabled.
“How would that have helped BU?”
It would have helped them at least get around the immediate argument and forced their opposition to narrow the argument (make #6 more narrowly defined as: “(6) monocrystalline EPITAXIAL growth layer formed directly on an amorphous buffer layer.”). After the argument is so narrowed it makes it easier to argue that it isn’t reasonable to require that level of enablement as a matter of drafting as there are other methods other than epitaxy that will suffice to enable the general structure actually claimed. Remember the actual claim isn’t actually all that narrow on that last limitation, it can be broadly enabled and at some point it is unreasonable to require enablement of everything that is outlandish (as has been pointed out above). Everytime someone claims a generic substrate and building a semiconductor device on it does that require the applicant to enable building a semiconductor device on a substrate made of a corn cob? A human ti t? Or at some point do we reasonably only require a generic enablement of genus language? The answer is yes we do.
“If the claim reads on an epitaxial growth embodiment, then they need to have enabled that epitaxial growth embodiment.”
Not necessarily. A lot of claims “read on” a lot of things that it isn’t reasonable to require enablement for. We generally give those a pass as noted above. He’s claiming the device/layers, not the method of making it. If it can be made by one method but not another it’s generally enabled, unless they’re getting specific in the claim to an extent that should require such. Yes it’s kinda bs that we do this but it’s apparently a tradition and it generally seems to work.
“It is beside the point to say that the claim reads on other things aside from epitaxial growth embodiments.”
Nah bruh.
“If they have enabled some of those other embodiments, then they need to limit their claims to the embodiments that they have enabled. One is not permitted to claim more than one has enabled.”
He’s claiming the device, not the method bruh.
“How would that have helped BU?”
It would have helped them at least get around the immediate argument and forced their opposition to narrow the argument to focus on 6 as more narrowly limited to epi growth. After the argument is so narrowed it makes it easier to argue that it isn’t reasonable to require that level of enablement as a matter of drafting as there are other methods other than epitaxy that will suffice to enable the general structure actually claimed. Remember the actual claim isn’t actually all that narrow on that last limitation, it can be broadly enabled and at some point it is unreasonable to require enablement of everything that is outlandish (as has been pointed out above). And further this is a device claim, epi growth is a method not claimed. In any event it makes the opposition have to start their argument again, this time with a weaker argument.
“If the claim reads on an epitaxial growth embodiment, then they need to have enabled that epitaxial growth embodiment.”
Not necessarily. A lot of claims “read on” a lot of things that it isn’t reasonable to require enablement for. We generally give those a pass, especially when it comes to devices that might can be made by one method but not others. He’s claiming the device/layers, not the method of making it.
“It is beside the point to say that the claim reads on other things aside from epitaxial growth embodiments.”
Nah bruh. Those other methods enable the device claim.
“If they have enabled some of those other embodiments, then they need to limit their claims to the embodiments that they have enabled. One is not permitted to claim more than one has enabled.”
He’s claiming the device, not the method bruh.
Judge Prost shoots down a patent? Nah, say it ain’t so…
And let’s not forget that while the judges used the spec to limit the patentee to have taught *only* cases of epitaxy only for growing the growth layer, if this patent is cited by an examiner as prior art, you can be sure the examiner will say it teaches *additional* forms of growing a crystalline layer on an amorphous layer. Where would patent law be without double-standards, right?
if this patent is cited by an examiner as prior art, you can be sure the examiner will say it teaches *additional* forms of growing a crystalline layer on an amorphous layer
At the very least, the patent suggests that other techniques exist. So what’s the problem with an Examiner making such a statement?
If the applicant disagrees, make the argument explaining why the Examiner is wrong. Is being put in a position of having to make that argument some kind of rank injustice?
Happens all the time, so there is no rank injustice at that point. The rank injustice happens afterwards when an examiner does not care that you have shown him as you suggest – and wants his first action with his mistake to count against you.
Are you being deliberately obtuse, or are you really that stup!d?
The CAFC just ruled that for purposes of 112 the spec only teaches using epitaxy. For an examiner to be allowed to say it teaches more than that for purposes of 103 constitutes a double standard. Whether or not an applicant has an opportunity to rebut isn’t germane to the discussion.
It’s a shame that software doesn’t get held to this standard.