Full opinion text
FINDINGS OF FACT AND CONCLUSIONS OF LAW WILLIAM ALSUP, District Judge. Introduction After a bench trial, this order constitutes the findings of fact and conclusions of law. Both sides have submitted lengthy proposed findings and conclusions. Rather than address each and every proposal, this order will find its own way through the evidence and arguments. Any proposal that has been expressly agreed to by the opposing side, however, shall be deemed adopted even if not expressly stated herein. That a proposal has not been expressly covered herein does not necessarily mean it was rejected; it only means that the Court found it unnecessary to reach. The Procedural History of the Case Abbott Laboratories filed the first of these actions on May 28, 2004. Three subsequent actions were filed. All concerned four United States patents owned by Abbott and Therasense, Inc. U.S. Patent No. 5,820,551 is the subject of this order. These actions were originally assigned to The Honorable Martin J. Jenkins. Judge Jenkins issued a first claim construction order for certain other patents and a separate claim construction order for the '551 patent. He also issued three separate summary judgment orders. The final one, dated April 3, 2008, involved all parties and all patents in suit. While the final summary judgment order did eliminate several infringement counts, several other claims were still viable. Immediately after issuing the final summary judgment order, Judge Jenkins left the federal bench, and all four cases came to the undersigned. All four cases were subsequently consolidated and a trial date was set for May 27, 2008. All defendants were permitted to file one more round of summary judgment motions and each party was allowed motions in limine. A technology tutorial for the undersigned was also held. The motions for summary judgment and motions in limine were fully briefed. A first omnibus order ruled on the motions for which oral argument was not required. Argument was then heard for the remaining pending motions. In a second omnibus order, the final pending motions were decided. Defendant Roche Diagnostics Corporation subsequently settled on the eve of trial. During this time, the Court and counsel also addressed the shape of the trial. It was decided that a trial on the '551 patent would be held first with all defendants and would be broken up into three separate phases: (i) invalidity and unenforceability; (ii) infringement (if needed); and (iii) willfulness and damages (if needed). All parties then stipulated that phase one of the '551 trial would be tried to the bench. Defendants raised four issues for phase one: inequitable conduct, obviousness, prosecution laches, and non-compliance with the written-description requirement. When the trial on the '551 patent began, the remaining defendants were Bayer Healthcare, LLC, Becton Dickinson & Company, and Nova Biomedical Corporation (collectively “BD/Nova”). Before trial began, Abbott made a request to add Attorney Lawrence Pope as a live trial witness in its case-in-chief. During Attorney Pope’s deposition, Abbott’s counsel had on three separate occasions insisted to defense counsel that Attorney Pope would not appear in person at trial. This was said in aid of repeated instructions not to answer. Attorney Pope was, therefore, scheduled to appear only through video-deposition. Based on Abbott’s insistence at deposition that Attorney Pope would not appear live at trial, Abbott’s request to have Attorney Pope appear as a live trial witness was initially denied. This denial was on the ground that it would be unfair to defendants, who had relied on the deposition representations to their detriment in not bringing Rule 37 motions. During trial, however, Abbott renewed its motion to allow Attorney Pope to testify in person at trial. The Court then asked Abbott to submit a sworn proffer showing the proposed statement of Attorney Pope’s testimony. Abbott submitted a declaration signed by Attorney Pope detailing the facts he would cover in his testimony. Because of the seriousness of the accusation against Attorney Pope, the Court relented and allowed Abbott to call Attorney Pope in its case-in-chief on those topics raised in his declaration. After defendants closed their case-in-chief, Abbott moved for partial findings under Rule 52(c) that defendants had failed to meet their burden of proof with respect to their defense of prosecution laches. The motion was granted on the ground that defendants had failed to show any intent to delay prosecution of the '551 patent or that substantial prejudice resulted from any such delay. Abbott also moved for partial findings as to defendants’ remaining invalidity defenses. These motions were all denied. Abbott rested its case-in-chief on June 2 and closing arguments were heard on June 3. This order now follows. The United Kingdom Work United States Patent No. 4,545,382 (and its European counterpart) is a decisive item of prior art in this decision. Here is its story. In the late 1970’s and early 1980’s, two research groups at the University of Oxford and the University of Cran-field in the United Kingdom were working on electrochemical sensors to detect the concentration of specific components in solutions. In particular, they were interested in developing electrochemical sensors that could be used to test glucose levels in human blood. Doctors Irving Higgins, Hugh Hill, and Elliot Plotkin were part of these research groups. In 1981, both groups teamed with a newly founded company, Genetics International, which was co-founded by James McCann. One goal was to create the first commercial electrochemical sensor for glucose. In 1981, the researchers filed their first patent application. This became the '382 patent in the United States and No. 0078,-636 B2 in the European Patent Office. The specification taught an improved electrochemical sensor for use in various liquid mixtures. The sensor was an electrode coated with specified chemicals that generated a tiny but detectable flow of electricity in the presence of glucose. The technology itself will be described below. In brief, the chemicals coated onto the electrode combined with glucose or whatever other “substrate” was being tested to generate small currents of electricity, which could then be measured by an ammeter. The higher the concentration of substrate, the higher the electrical current, and the higher the meter reading. The patent disclosed certain ferrocene chemistry that allowed for fasting testing. Although the United States '382 patent lived out its seventeen years without incident, its EPO counterpart (iethe '636) was eventually revoked based on a German prior-art reference that was cited by a third party in a European opposition proceeding. That was in the mid-1990’s. The decision to revoke the patent was appealed, however, and the patent was eventually reissued by a technical board of appeal in the European Patent Office. Certain submissions made along the way by Abbott’s predecessors, however, have turned out to be important in this proceeding by reason of their non-disclosure to the PTO during prosecution of the '551 patent in suit. The research group continued its work on sensors for testing glucose levels in blood. Dr. Hill and his colleagues filed several additional patent applications, which were later combined to form a single United States patent application. All parties herein agree that the resulting U.S. Patent No. 5,820,551 — the patent in suit— claims priority to May 1983. James McCann and Drs. Hill, Higgins and Graham Davis were listed as the inventors. Originally, the claimed invention of the '551 patent was the development of a disposable electrode strip whose electrodes could be covered by a single drop of solution. These one-use strips would be inserted into a convenient unit for digital readout of the level of a target compound (like glucose) in a test liquid mixture (like blood). After a strip was used to generate a readout, it could be thrown away. The '551 patent was in prosecution for over fourteen years. During this period, Genetics International changed its name to Medisense, Inc. Various claims were rejected twelve times by the PTO examiner, David Shay. Eleven out of the twelve rejections relied on the '382 patent or its European counterpart, the '636 patent. During this prolonged prosecution, Medi-sense amended the proposed claims several times to overcome rejections by Examiner Shay — all without success. At times, Medisense also submitted declarations from persons of ordinary skill in the art to distinguish its claims from the prior art. None of the proposed amendments ever included a limitation for a sensor without a filter or a membrane. In the meantime, several other companies, including defendants Bayer Healthcare, LLC, and Becton Dickinson & Company, had begun manufacturing and selling disposable electrochemical sensors for diabetic patients. In 1996 — while the '551 patent was still pending before the PTO — Medisense was purchased by Abbott Laboratories. After the acquisition, Abbott brought in one of its in-house patent attorneys to take over the prosecution of the '551. That attorney was Lawrence Pope. Attorney Pope worked in conjunction with several technical employees at Medisense, including Dr. Gordon Sanghera, to “brainstorm” various arguments regarding the patentability of the '551. Dr. Sanghera had worked at Medisense since 1990. As of 1997, he was its director of research and development in the United States. His responsibilities included running competitive analysis in conjunction with the marketing department and supervising Abbott’s patent portfolio. Dr. Sanghera had also previously worked for Dr. Hill at his laboratory at Oxford University. He had researched electrochemical sensors, but he had not been involved in the research that led to any patents involved herein. Dr. Sanghera had, however, attended the European opposition hearings and had been active in crafting the submissions made in that appeal. Dr. Sanghera and Attorney Pope struck upon a new point of possible novelty previously overlooked in the pending prosecution. The new point was that the specification disclosed a sensor for use in whole blood without any protective membrane. Trouble was, a passage in the earlier '382 patent already seemed to disclose mem-braneless sensors. That passage read (col.4:63 — 66): Optionally, but preferably when being used on live blood, a protective membrane surrounds both the enzyme and the mediator layers, permeable to water and glucose molecules. To address this problem, Abbott decided to assert, as a matter of extrinsic fact, that in 1983 skilled artisans would have believed that a membrane was essential even in the face of the '382 disclosure, ie., they would not have taken the quoted sentence literally. Attorney Pope then held an interview with Examiner Shay on November 4, 1997. Examiner Shay summarized the interview as follows (TX 469): Applicant indicated that he would like to submit claims specifically covering a compound specific electrode with the filtering membrane absent. The Higgins et al. (’382) disclosure was discussed especially] the paragraph spanning columns 4 & 5. It was determined that since Higgins et al. appear to require the membrane for use with whole blood (see example 8) an affidavit or other evidentiary showing that at the time of the invention such a membrane was considered essential would overcome this teaching. As arranged in the interview, Attorney Pope submitted a declaration by Dr. Sanghera on December 3, 1997, along with amendments to the claims. The declaration stated in relevant part (TX 443): THAT based on his historical knowledge he is confident that on the filing date of the earliest application leading to the present application on June 6, 1983 and for a considerable time thereafter one skilled in the art would have felt that an active electrode comprising an enzyme and a mediator would require a protective membrane if it were to be used with a whole blood sample. Therefore he is sure that one skilled in the art would not read lines 63 to 65 of column 4 of U.S. Patent No. 4,545,382 to teach that the use of a protective membrane with a whole blood sample is optionally [sic] or merely preferred. The entire submission was aimed at overcoming the “optionally, but preferably” sentence in the '382 patent. Attorney Pope submitted parallel remarks stating that those of ordinary skill in the art believed that the use of a protective membrane was “required” when testing whole blood and that they would have understood the sentence in question as mere patent phraseology, not a technical teaching. Based on Dr. Sanghera’s declaration and Attorney Pope’s remarks, Examiner Shay finally approved the proposed claims and the patent issued on October 13, 1998. The foregoing findings will be amplified with many details below. Invalidity In this action, the central axis of contention concerns membranes and, more particularly, their use as a permeable layer surrounding the chemistry coated onto the active electrode. Late in the fourteen-year prosecution, as stated, Abbott advanced the theory that the '551 specification revealed a sensor without a protective membrane. However, a key prior art reference — the inventors’ own '382 patent— had already stated that such membranes were optional and at most preferred in certain circumstances, as quoted above. This '382 sentence was raised by the examiner as having already taught that membranes were merely optional or preferred. In response, as stated, Abbott took the position (and still maintains) that the sentence would not have been understood in 1983 by those skilled in the art to have modified a supposed conventional wisdom that a membrane was necessary for testing in whole blood. Defendants disagree. They point out that the '382 sentence expressly stated that even for live blood, a membrane was merely “preferred” and that for all other cases it was “optional.” In no case was it said to be “required.” With this introduction of the central invalidity issue, this order will go back to square one. It will begin by setting forth the basic technology. It will then review the '382 patent, focusing on its entire disclosure, including the sentence in question, so as to place that sentence in full context, all from the point of view of one skilled in the art at the time of the alleged '551 invention (in 1983). * * * Although this summary of the technology is now stated in the present tense, this summary was all known in the prior art. The electrochemistry involves an electrode coated with an “enzyme” catalyst. The enzyme is particularly selected to react with glucose or whatever the test substance might be. Again, the substance being tested for is sometimes, as used by Medisense in the EPO proceedings, called a “substrate.” The enzyme-substrate chemical reaction generates electrons. The electrons are passed via yet another chemical called a “mediator,” also coated onto the electrode, to the active electrode itself. The electrons then flow as a tiny but measurable electrical current down the active electrode through an ammeter and back to the other uncoated electrode. The blood droplet or other solution under test provides an electrical path completing the circuit between the electrodes. The word “sensor” is sometimes used interchangeably with the active electrode, ie., the electrode painted with the active chemistry. An analogy is to a battery. Battery chemicals generate electrons and thus electricity, which can then be used to do work, such as to drive a meter. In the technology at hand, the active chemistry is the glucose, enzyme, and mediator. Together, they generate the electricity. Thus, when blood is placed between and across the electrodes, the chemicals coated onto the active electrode go to work, generating electricity or “signal.” The electricity passes through an ammeter, which detects the current. The current will ideally be in proportion to the concentration of glucose. In this way, the meter can be calibrated to progressive concentrations of glucose. The user can then see when the glucose (or other substrate) is too low or too high. All of the foregoing was known in the prior art. * * * One of the contributions of the '382 patent — which was concededly prior art to the '551 patent — was a faster-acting ferrocene mediator coated onto an active electrode along with an enzyme. Faster acting meant faster response times and quicker test results. In the “Background of the Invention,” the inventors stated that the '382 invention would have particular value for “in vivo measuring or monitoring of components in body fluids” (col.l:16-17) and said “the determination of glucose in a diabetic human subject” was a primary application (col.l:20-21). The background stated further that the invention lent itself to temporary or permanent implantation. Although “the provision of an implantable glucose sensor [was] a major object of the invention” the inventors noted that “other and broader objects [were] not hereby excluded” (col.l:23-26). A few columns later, for example, the specification called out home-testing kits with disposable sensors. After acknowledging that in vivo glucose sensors had already been proposed by others, the inventors stated that they had recently carried out in vitro studies. Under “Summary of the Invention,” the '382 inventors stated that they had come to realize that mediator compounds could be associated with the sensor electrode structure itself to make such electrodes available for use by in vivo methods. The '382 invention was then described as a sensor electrode composed of a combination of enzyme and mediator (col.l:60 — 63). Preferably, the electrode was designed to determine glucose in vivo (col.l:65-66). A long passage then described various mediators and enzymes (col. 2:1 to col. 4:55). Again, a significant contribution was the ferrocene chemistry that was faster acting than in the prior art, thus reducing response time. At a few places in this passage, the inventors referenced membranes. For example, two paragraphs stated (col. 3:53 to col. 4:2): In that form of the invention using poly-viologens, as exemplified in the three modifications above, it is an objective to keep loss of active material (enzyme or mediator) to a very low level, i.e., by the surrounding membrane, co-immobilisation or covalent bonding. In a different form of the invention, however, still using glucose oxidase, a rather higher level of loss of active material is tolerated, giving a sensor electrode of reduced but still useful life, coupled with improve [sic] sensitivity and selectivity. In this form of the invention the electrode is composed of particulate carbon mixed with a low molecular weight mediator disseminated throughout the electrode and glucose oxidase. Chloranil and/or fluoranil are useful mediator substances. It is envisaged to construct from such an electrode a replaceable sensor tip to a needle-type probe for projecting only into the dermis so as to allow ready replacement. Put differently, after describing a membrane application, the “different form” of the invention dispensed with the membrane and thus “tolerated” a “rather higher level of loss of active material” (due to the absence of the immobilizing membrane). It was envisaged to have replaceable sensor tips for projecting into the dermis. Another version called out ferrocene-glucose oxidase as “particularly valuable” and stated “the enzyme layer is preferably immobilised at the surface of the underlying mediator, retained in a self-sustaining gel layer” or with “a retention layer there-over permeable to the glucose molecule” (col.4:13-16). “Immobilisation” was a reference to retaining the active chemicals on the electrode so that they would not fall away into the blood or other fluid. Then came the main sentence at the heart of this case (col.4:63-66): Optionally, but preferably when being used on live blood, a protective membrane surrounds both the enzyme and mediator layers, permeable to water and glucose molecules. This allowed water and glucose to pass through, kept in the chemicals, and kept out larger blood constituents like red blood corpuscles. The “Summary of the Invention” then turned to various applications and specifically called out implanted glucose sensors, digital readout diabetic home-testing kits, devices to take a blood sample from the finger, place it on the sensor, amplify the signal, and give a digital readout, and a watch-type device for monitoring glucose interstitial fluid in the skin with disposable-sensor cartridges in the back, which would plug into the electrodes. Next came a “Description of the Preferred Embodiments.” In total, the '382 patent contained thirteen working examples of preferred embodiments of the invention. Some of the examples described various procedures for producing the working chemistry of the sensor — ie., the enzyme and mediator. Other examples described possible configurations of electrodes and electrochemical sensors. Each was configured slightly differently depending on various test parameters, including the type of solution being tested. Some of the sensors included a membrane and others did not. Examples 1 and 2 described purification processes for producing quinoprotein glucose dehydrogenase — an enzyme used to catalyze the chemical reaction. Examples 3 and 4 explained the interaction between glucose oxidase (another enzyme) and fer-rocene — the mediator which allowed for much faster and more linear testing than the prior art. Example 5 described the construction of an in vitro sensor with a glucose oxidase enzyme and polyviologen mediator. A dialysis membrane was used. The purpose of the membrane was to block larger molecules from passing through to the working chemistry. The sensor was tested in a buffered electromechanical cell. As the amount of glucose in the test solution was increased, the current generated by the sensor grew, thereby indicating that the electrode was acting as a glucose sensor. This same construction was used in Example 6, except chloranil was used as the mediator. Example 7 taught a sensor configured with a glucose oxidase enzyme and a dimethyl ferrocene mediator designed for use in interstitial fluid — ie., skin. Before the sensor was used for testing, the electrode, mediator, and enzyme were dipped into a solution of cellulose acetate, thereby creating a protective membrane over the working chemistry and electrode. The example went on to state: “The small size of such an electrode and its linear response over a large range of glucose concentrations makes it possible to use the electrode for in vivo glucose determination on both severely diabetic and normal individuals” (col.8:54-59). Example 8 — entitled “In vitro sensor”— was the most discussed embodiment at trial (col.8:63). The example began by describing the construction of a sensor with a glucose oxidase enzyme and a ferrocene mediator. No membrane was applied. The example then explained that the sensor was first tested in “nitrogen-saturated buffer solution” (col.9:15). The results for the test in buffer solution were then summarized. A cellulose acetate membrane was then applied to the sensor. The example went on to describe response times for that sensor in buffer and then, in a separate test, in blood. The exact language in the specification stated (col.9:26-33) (emphasis added): With the same buffer, such an electrode modified by a cellulose acetate membrane coating (produced as in Example 7) gave response times of 36 seconds (2 mM) and 72 seconds (6 mM). With blood, this modified electrode gave response times of 36 seconds (blood with a known 2mM glucose content) and 72 seconds (blood at known 6mM glucose content). The sensor constructed in Example 8 was thus tested in two solutions. The example first described was tested in buffer solution. At this point in the specification, no membrane was applied to the sensor. A membrane was then placed on the sensor. The response times of the sensor with a membrane were subsequently set forth for the same buffer solution and then, separately, for blood. It is Abbott’s contention that this example shows that a membrane was in fact required by the invention of the '382 patent when testing in whole or live blood. Example 9 taught the construction of an electrode with a glucose dehydrogenase enzyme and a ferrocene mediator. A dialysis membrane was used to cover the coated electrode. Examples 10 and 11 were minor variations of Example 9. Examples 12 and 13 described further configurations for an electrode with a glucose dehydroge-nase enzyme and ferrocene mediator. Finally, under the '382 claims, Claim 1 covered the sensor electrode coated with the enzyme and mediator. All agree that Claim 1 covered electrodes without limitation to either in vitro or in vivo use. All agree that Claim 1 covered versions with and without membranes. Indeed, dependent Claim 12 narrowed the claim to sensor electrodes having an outermost protective membrane permeable to water and glucose molecules. In sum, the '382 disclosed the basic structure of an active electrode and a faster-acting chemistry, stating that the structure could optionally include a protective membrane as an outer layer and stating that such a membrane was preferable when used with live blood, although the examples involving blood employed a membrane. * * * Turning to the '551 patent in suit, its inventor group was virtually the same as for the '382, with slight adjustments. It was directed to a home-testing kit and more specifically to a two-electrode strip (rather than a three-electrode strip) for one-time, disposable attachment to a han-dheld readout device. The electrodes were coated with enzymes and mediators (“preferably a ferrocene”) — as in the '382 patent. The strip was described as “elongated” for ready handling and assembly. As with the '382, the active electrode was “preferably formed of carbon.” The inventors went on to say that carbon foil available commercially as GRAPHOIL or PA-PYEX was a valuable electrode material. Various “objects” of the invention were described, none of which related to a membrane or lack thereof. Many columns were devoted to construction of the electrodes. The subject of membranes was mentioned only twice in the '551 application. Under “Membrane Cover for Electrode,” the inventors said that “it may be found valuable to exclude the sensor from interfering contact with larger molecules or tissue fluid components” and that this could be done with a “surrounding membrane” (col.6:67-7:13). That passage briefly described how to make a membrane in situ. Later, a step-by-step constructional sequence was given for an electrode strip. Seven steps were listed. Adding a membrane was not listed as a step (col.8:35-51), an omission since given great weight by Abbott. A later, optional modification stated: “The electrode may then be covered, on both sides, with a semipermeable membrane of cellulose acetate (or polyurethane), not shown, to block large interfering species from contact with the electrode” (col.9:34 — 37). Nowhere in the '551 specification or the original claims was there any suggestion that treating the membrane as optional (or omitting it) was an inventive step. Nonetheless, this order appreciates that a legitimate invention may eventually be found lurking in a disclosure even though the inventors missed it themselves for over a decade. See Newman v. Quigg, 877 F.2d 1575, 1581 (Fed.Cir.1989). So this order accepts Abbott’s contention, at least for purposes of argument, that the '551 specification disclosed an active electrode without a membrane for use with whole blood (as well as disclosing one with a membrane for use with whole blood). The decisive question remains whether or not the same group of inventors (with slight membership changes) had already disclosed in the '382 patent that a membrane was merely preferred for use with live blood and was optional in all other cases. This order now turns to resolving that question. * * * This order accepts Abbott’s proposition that prior to the '382 patent, those skilled in the art typically employed a membrane on a sensor used with live or whole blood, although one exception was already in print. That practice, however, was before the revelation in the '382 patent. The '382 patent expressly stated that a protective membrane was optional in all cases except for live blood, in which case it was preferred In no case did the '382 patent state that a protective membrane was required. In context, it seems clear why this was so. The invention specified a faster-acting ferrocene chemistry. This allowed for shorter response times, i.e., measurement times. This, in turn, reduced the raison d’etre for any membrane. For example, the faster response times reduced the probability of the active chemicals being washed away in the bloodstream and reduced the time within which red blood corpuscles could locate and foul the electrodes. (Fouling refers to the larger red blood cells accumulating on the electrode and blocking the much smaller glucose molecules from reaching the sensor.) The indicated readings took about a minute, even less without any membrane. There was, therefore, less need for any membrane. As a matter of sentence structure, the sentence sets up two cases — an optional case and a preferred case: Optionally, but preferably when being used on live blood, a protective membrane surrounds both the enzyme and the mediator layers, permeable to water and glucose molecules. Italics have been supplied here to illustrate the structure. Ignoring the italicized preferred case, the sentence states: “Optionally, ... a protective membrane surrounds both the enzyme and the mediator layers .... ” That is the general, optional case. The exception, ie., the preferred case, is for live blood. That phrase is italicized. The trial record is clear and convincing that persons of ordinary skill in the art understood the words “optionally” and “preferably” in the same way as the rest of us. There is no doubt that those skilled in the art would have understood that the sentence was trying to say exactly what has been laid out in this paragraph. Abbott contends that skilled artisans simply would not have believed the sentence and would have had no reasonable expectation of reliance on it by reason of a prevalent view that membranes were essential when testing in whole blood. A revelation in a public disclosure cannot be erased from the prior art on the theory that it contradicted the conventional wisdom. The whole point of disclosures in patents is to reveal something new. See Atlas Powder Co. v. Ireco, 190 F.3d 1342, 1347 (Fed.Cir.1999). Abbott has tied itself in knots contorting the grammar to come up with an alternative meaning. One example is Abbott’s Proposed Finding No. 90, which reads: Interpreted in light of the conventional wisdom at the time, the “preferably” language means that the membrane is optional when an in vivo sensor does not contact whole blood but that the membrane is required when the sensor contacts red blood cells in whole blood. This contortion collapses on its own weight. The sentence in question meant just what it said and the ordinary artisan would have so understood it. Contrary to Abbott, Example 8 in the '382 patent was consistent with the plain meaning of this sentence. Example 8 was one of the preferred embodiments. It described a ferrocene-glucose oxidase electrode. In the experiment described in Example 8, the sensor was tested in a buffer solution with two different glucose concentrations, yielding response times of 24 and 60 seconds, depending on the concentration. Then a protective membrane was applied to the sensor. While still testing the buffer concentration, the response times went to 36 and 72 seconds, respectively. The same sensor — again with a protective membrane — was tested in blood samples with the same concentration and 36- and 72-second response times were again obtained. It is true that when the sensor was used in blood a protective membrane was used and that a protective membrane was not used with the first buffer solution. Nothing in Example 8, however, stated that a membrane was required for use in blood. That a membrane was added seems to have been little more than a way to investigate the time effect of adding a membrane. It is also true, as Abbott urges, that no test recited in the preferred embodiments included a test on blood without a membrane. There were, however, too few blood examples among the embodiments to warrant any inference from this happenstance. No doubt, the broad teaching of the sentence in question went beyond the specifies of the preferred embodiments. That is often true in patents. Broad teachings do not have to be supported by specific experimental examples in order to qualify as prior art. The '382 sentence was then and remains correct, a fact that even Abbott does not challenge. Membranes were never part of the electrochemistry itself. Rather, they offered certain mechanical advantages, provoked by two different concerns. The first was human safety. For in vivo use, toxic materials might break away from the coated sensor and pollute the bloodstream. To protect against this possibility, a membrane immobilized the active ingredients, i.e., retained them in place and thus reduced the risk of breakaway. Reduced response times from faster chemistry, however, reduced the breakaway risk — for the sensor could be removed sooner than before. The second concern was the risk of “fouling.” This was the risk that red blood particles would stick to the active electrode and prevent glucose from interacting with the chemicals coated onto the electrodes. If enough “fouling” occurred, the signal would be diminished below an acceptable level and an erroneous readout would occur. Fouling might occur in live blood or whole blood. In these proceedings, the supposed problem of fouling has been exaggerated by Abbott. After the faster chemistry disclosed in the '382 patent, the risk became more theoretical than practical, especially for one-use, disposable applications. Subsequent diabetic kits using the faster sensors have deleted the membrane with acceptable results. In sum, the '382 statement in question was then and remains correct. Abbott’s idea that skilled artisans would have read the sentence in question and disbelieved it in 1983 is not plausible on the trial record. Skilled artisans would have known that deleting the membrane would simply have deleted their mechanical advantages. They would have known, however, that the electrochemistry would still have worked. They would have known that the degree of fouling would have depended on how long the sensor was exposed to blood. They would have known that the risk of fouling would have been reduced for faster-acting chemistry and reduced even more for sensors used only once, ie., disposable sensors with no accumulation of residue. They would have known that omitting the filter would have had the further advantage of speeding up the test time even more. To be sure, in making these findings in the preceding paragraph, the Court has relied on trial testimony and materials outside the four corners of the patent and prior-art references. This, however, is because Abbott itself has resorted to extrinsic evidence and “conventional wisdom.” That is, to overcome the '382 prior-art sentence in question, Abbott has resorted to extrinsic evidence, arguing that skilled artisans would not have understood the sentence in light of prevailing practices. Therefore, it is entirely appropriate for the other side to likewise resort to extrinsic evidence as to how those skilled in the art would have taken the '382 sentence in question. Abbott next argues that one skilled in the art would have read the “optionally, but preferably” passage of the '382 as mere “patent phraseology.” Notably, the passage in question stated that a membrane was preferable “when being used on live blood” (col.4:63-64). Defense Expert Turner testified that even today’s implantable electrochemical sensors used for testing glucose in live blood would use a membrane to ensure that toxic materials were not released into the blood stream (Tr. 333). He even went as far as saying that the FDA would likely not approve an implantable sensor without a membrane for safety reasons, which was the reason a protective membrane was preferable for live blood. Unlike implantable sensors used to test live blood, however, one-time disposable sensors for in vitro testing (as disclosed in the '551 patent) had no such safety concerns. With in vitro testing, a membrane was motivated only by the potential of fouling the electrode. As to them, the '382 sentence in question taught that the membrane was merely optional. The examiner was persuaded by Abbott’s view as a result of two considerations. One was the presence or absence of a membrane in the '382 examples, particularly in Example 8. This argument is unpersuasive and rejected by this order, for the reasons stated above. The other reason was based on an extrinsic evidentiary declaration without which the examiner said no allowance would be made. This was the now-controversial declaration of Dr. Gordon Sangh-era. Although he was not a co-inventor, he had worked at Medisense and had become an Abbott employee at the time of his declaration. The entirety of his substantive statement to the examiner was as follows (TX 443): 3. THAT he is familiar with U.S. Patent No. 4,545,382 and with the history of the development of the technology disclosed in this patent. In particular he is familiar with the beliefs and concerns of those skilled in the art in 1981 when the first application leading to this patent was filed as well as in 1983 when the first application leading to the present application was filed. 4. THAT he is familiar with the teachings of U.S. Patent No. 4,987,173 to Nankai et al. and in particular with the teachings of Examples 3 and 4 with regard to the construction of sensors for use with serum and whole blood samples. 5. THAT based on his historical knowledge he is confiednt [sic] that on the filing date of the earlist [sic] application leading to the present application on June 6, 1983 and for a considerable time thereafter one skilled in the art would have felt that an active electrode comprising an enzyme and a mediator would require a protective membrane if it were to be used with a whole blood sample. Therefore he is sure that one skilled in the art would not read lines 63 to 65 of column 4 of U.S. Patent No. 4,545,382 to teach that the use of a protective membrane with a whole blood sample is optionally or merely preferred. 6. THAT Examples 3 and 4 of U.S. Patent No. 4,897,173 provide evidence that this concern about unprotected active electrodes for whole blood samples persisted until at least the June 21, 1985 filing date of the earliest application leading to this patent. The fact that the Example 3 teaching a sensor for use with serum samples has no protective membrane but Example 4 teaching a sensor for blood has a polycarbonate membrane is evidence that the authors of this technical disclosure still believed that active electrodes could not be directly exposed to whole blood samples. With the exception of the '173 Nankai patent, the declaration was conclusory and unsupported. The '173 Nankai patent was more specific. It did, indeed, happen to use a filtration layer with whole blood and did not use one with serum, as Abbott states. The Nankai PCT filing date was June 19, 1986. Nankai did tend to support the “conventional wisdom” argument advanced by Abbott. But Nankai was and remains subject to a very important and overriding caveat. The Nankai specification made no reference to the '382 patent and said nothing about the “optionally, but preferably” sentence. Nankai was silent on the key sentence. Nankai did not purport to construe it. Although for obviousness purposes, the hypothetical person skilled in the art is presumed to have full knowledge of all prior art, that in no way means that we must presume Nankai knew of the '382 sentence in question. Nankai teas simply one practitioner, not someone presumed to be omniscient. His patent in no way addressed the meaning of the key sentence. He may have been unaware of the key sentence, for all the record shows. By contrast, for our obviousness purposes, we must presume the hypothetical artisan knew all of the prior art, including the key sentence at issue. See Custom Accessories, Inc. v. Jeffrey-Allan Ind., Inc., 807 F.2d 955, 962 (Fed.Cir.1986). The decisive fact remains that those skilled in the art, had they read it, would have understood the '382 sentence as stating that a protective membrane was preferred in the case of live blood and optional in all other cases. They would have understood it as disagreeing with any viewpoint that membranes were necessary when testing whole or live blood. The very purpose of a patent is to disclose new information to persons skilled in the art. This order finds that the '382 patent taught those skilled in the art that — at least when faster chemistry was employed — a protective membrane was optional in all cases except the case of live blood, in which case the protective membrane was preferred — but not required. The trial evidence and the plain language of the disclosure are clear and convincing on this point. Abbott’s “conventional wisdom” evidence is rejected. The foregoing is sufficient. To this, it must be said that the information withheld from the examiner, discussed momentarily, eviscerates any vestige of plausibility to Abbott’s extrinsic evidence, for that information from the applicants themselves now shows that they knew full well the meaning of the very “optionally, but preferably” sentence at the heart of this suit. This evidence, revealed for the first time in these proceedings, also decidedly supports this order’s invalidity conclusion. This order rejects the Sanghera declaration and its supposed conventional wisdom. * * * There is a different aspect to Abbott’s entire theory that deserves comment. Deletion of a feature from a prior-art device with a corresponding deletion of its function is not an invention. For example, if the prior art already discloses a pencil with an eraser, one may not delete the eraser and claim an eraserless pencil as an invention. The reason is that the deletion of the eraser would also mean a deletion of its function. This would be true even if the conventional wisdom was that all pencils came with erasers. See Richards v. Chase Elevator Co., 159 U.S. 477, 486, 16 S.Ct. 53, 40 L.Ed. 225 (1895). Similarly, deletion of the protective membrane was not inventive in the '551 patent because there was a corresponding deletion of its function. The loss of this function was tolerable because the chemistry was fast enough (at least by the time of the '382 prior-art disclosure) to obtain acceptable results without a membrane. But assuming arguendo that skilled artisans had uniformly believed that a membrane was necessary (despite the '382 patent), the mere deletion of the membrane with a corresponding loss of its functions would not warrant a patent. It would be different if the '551 patent disclosed a specific configuration that preserved the membrane’s function but without the membrane. Exactly what was disclosed in the '551 patent that compensated for the deletion of the membrane and guarded against fouling? The Court asked this question several times during the bench trial. Clearly, the '551 specification and prosecution history were totally silent on this point. At the closing argument, Abbott’s counsel argued — for the first time — that the '551 disclosed use of certain materials for constructing the electrodes and that these materials were less sensitive to oxygen. Whereas the '382 patent had taught carbon as a preferred electrode, the '551 patent recommended carbon foil available commercially as GRAPHOIL or PAPYEX. The argument emerged that normally red blood cells (and their oxygen content) posed a noise hazard but that the electrodes specified in the '551 were less sensitive to oxygen. Thus, it was said, a membrane could be safely deleted from an electrode constructed from GRAPHOIL. Nowhere in the specification, nowhere in the prosecution history, and nowhere in the trial evidence was this point made. It surfaced for the first time at closing argument. Still, the Court has fully considered it. It is easy to see why it has taken so long to invent this line of argument. Both the '382 and '551 patents disclosed electrodes that exhibited immunity from oxygen. For example, the '382 specification stated that the electrodes exhibited “very low oxygen sensitivity.” This would allow “omission of the dilution step involved in blood analysis in current instruments,” the '382 specification said (col.5:20-22). For its part, the '551 patent stated that, for carbon foil, “oxygen interference is minimal, there being less than 4% change in signal between anaerobic and fully aerobic samples” (col.7:15-20). Given that the '382 had already disclosed “very low oxygen sensitivity,” the later statement in the '551 patent was no improvement on that score. The '551 statement was a passing comment on a design consideration (concerning oxygen sensitivity) that had been covered in the earlier patent and was covered again in the later patent. Since the earlier patent had already achieved “very low” oxygen sensitivity, it is far-fetched to argue that the later patent somehow solved that problem, much less solved it in a way that specifically dispensed with the need for a membrane, a nexus nowhere made until at closing argument in 2008. To return to the main point, the clear-cut fact remains that to the extent the '551 dispensed with the membrane, it also dispensed with its function and thus no invention was disclosed at all. This point would hold even if we indulged Abbott’s view of the conventional wisdom about membranes at the time. The main invalidity issue is the no-membrane limitation. The foregoing resolves that key component. The inequitable-conduct issue is also anchored in the no-membrane limitation. This order, therefore, will now proceed directly to that issue for ease of reader convenience and return later to the less controverted limitations and complete the obviousness analysis. Inequitable Conduct Turning to the defense of inequitable conduct, the “optionally, but preferably” sentence remains at center stage. When Abbott acquired the pending application that led to the '551 patent, its in-house lawyer, Lawrence Pope, took over the prosecution. That was in 1997. He replaced Fish & Richardson, who had been unsuccessful for twelve years in obtaining allowance of any claims. Examiner Shay had repeatedly rejected all proposed claims over the '382 patent. Abbott “brainstorming” sessions were held to find a way to win claims on the '551 application. These sessions included Dr. Gordon Sanghera. The original inventors were not included. By this point, Abbott’s competitors were beginning to sell diabetic home-testing kits in competition with the Exactech, the Medisense-Abbott product. Although Dr. Sanghera denied it at trial, this order finds that Dr. Sanghera and Attorney Pope were motivated, in part, by marketplace developments to find a claim to suppress competition. The very day the '551 patent issued, for example, Abbott asserted it in a patent-infringement action against a home diabetic kit made by Lifescan, Inc. There is, however, nothing wrong with seeking a patent in order to stifle competition, at least under the patent laws, so long as the patent is lawfully obtained. The brainstorming sessions produced an argument never before advanced by the inventors or by prior counsel, namely that the '551 specification taught that a protective membrane was not necessary when testing whole blood. This argument was then presented to Examiner Shay in an oral interview by Attorney Pope in November 1997. With respect to novelty and the prior art, they expressly discussed the '382 sentence. For convenience, this now-familiar sentence is repeated: Optionally, but preferably when being used on live blood, a protective membrane surrounds both the enzyme and the mediator layers, permeable to water and glucose molecules. More specifically, the Interview Summary (TX 469) referenced the Higgins '382 and Pace '410 patents and stated: Applicant indicated that he would like to submit claims specifically covering a compound specific electrode with the filtering membrane absent. The Higgins, et al. (’382) disclosure was discussed especially] the paragraph spanning columns 4 & 5. It was determined that since Higgins et al. appear to require the membrane for use with whole blood (see example 8) an affidavit or other evidentiary showing that at the time of the invention such a membrane was considered essential would overcome this teaching. A box was checked stating that an agreement had been reached. In short, the examiner agreed to permit an evidentiary showing to overcome the presumed teaching of the “optionally, but preferably” sentence. To this end, Attorney Pope prepared a sworn declaration for the signature of Abbott’s Dr. Sanghera. Although he was skilled in the art by the time of the declaration, Dr. Sanghera had not been skilled in the art at the time of the invention (and, as stated, had not been one of the inventors). This, of course, was not a requirement for a declaration. Dr. Sanghera read, understood, and signed the declaration, knowing its purpose and knowing that it would be submitted to the PTO to overcome the presumed teaching of the sentence. The declaration is quoted above. In brief, it stated that Dr. Sanghera was sure that one skilled in the art at the time of the invention would not have read the sentence in question to teach that the use of a membrane with a whole-blood sample was optional or even preferred. To this end, Dr. Sanghera did not consult with any of the inventors to learn what had been considered optional, preferred, or essential despite the fact he still had a good relationship with at least Inventor Hill. He limited his research to literature. The declaration was filed for Examiner Shay along with an amendment and remarks by Attorney Pope. The amendment cancelled all prior claims and proposed new claims, soon allowed. The attorney’s remarks (TX 470) are now set forth at length with italics on the passages of particular relevance: At the interview the applicants’ undersigned representative explained that a new set of claims would be presented which focus on the feature that the active electrode is directly exposed to a whole blood sample without the intervention of a barrier material such as a membrane or gel which filters out larger molecules or other blood components expected to interfere with the active electrode’s operation. It was agreed that this embodiment was one of the options clearly disclosed in the present application. It was also agreed that the art generally taught the use of such protective barriers on the effective filing date of the present application. The applicants’ representative pointed out that U.S. Patent No. U, 54-5,382 to Higgins et al teaches that active electrodes designed for use with whole blood require a protective membrane. He noted that the general teaching to this effect at lines 63 to 66 of column 4 of this patent was amplified and supported by the specific working examples. In each working example in which an active electrode was prepared for use with a whole blood sample it was provided with a protective membrane by either deposition of a cellulose acetate film or attachment of a dialysis membrane. Example 8 at columns 8 and 9 was noted as being particularly instructive in this regard. An active electrode was constructed by successively coating the end of a carbon rod with ferrocene and then glucose oxidase. This unprotected active electrode was first tested in nitrogen saturated buffer and then in an air saturated buffer to establish the impact, if any, of oxygen on the reaction; the impact appears to have been minimal. Then at fines 22 to 33 the effect of a cellulose acetate membrane on response time was investigated when the sample was buffer and when it was blood. In both cases the response time appears to have increased by as much as 50%, e.g., from 24 to 36 seconds for a low level of glucose. Nevertheless all the succeeding examples utilized a protective membrane. The dear implication is that the use of protective membrane caused a slower response time but nonetheless was needed for a whole blood sample. The art continued to believe that a barrier layer for whole blood sample was necessary for a considerable period. For instance, U.S. Patent No. 4,897,173 to Nankai et al (copy accompanies this response), which claims priority from 1985, describes the production of electrodes for the measurement of glucose. In Example 3 at columns 4 and 5 an electrode structure for serum (see fine 6 of column 5) is described which does not involve a protective membrane. In contrast Example 4 at columns 5 and 6 directed to an electrode for use with whole blood (see fines 61-62 of column 5) teaches a filtration layer 21 with a pore size of one micron. One skilled in the art would not have read the disclosure of the Higgins patent (U.S.Jp,545,382) as teaching that the use of a protective membrane with ivhole blood samples was optional. He would not, especially in view of the working examples, have read the optionally, but preferably language at line 63 of column 6 as a technical teaching but rather mere patent phraseology. This is supported by the Declaration under 37 C.F.R. 1.132 of Gordon Sanghera which accompanies the present amendment. The Examiner is respectfully requested to indicate the allowability of the currently pending claims and issue a Notice of Allowance. The applicants have established that a new claim limitation supported by the present application provides a patentable distinction over U.S. Patent No. 4,545,382, the key reference in the prosecution of the present application and its predecessors. There is no teaching or suggestion of unprotected active electrodes for use with whole blood specimens in this patent or the other prior art of record in this application. Furthermore, the present claims are patentably distinct from the claims of U.S. Patent No. 5,682,884. Therefore, this case is in condition for allowance. In sum, Attorney Pope’s remarks stated that the sentence in question would have been regarded as “mere patent phraseology” rather than a “technical teaching” and that the art believed that a membrane was “required” even for a considerable period after the '382 patent, closing with: “There is no teaching or suggestion of unprotected active electrodes for use with whole blood specimens in this patent or the other prior art of record in this application.” In reliance on the submission, Examiner Shay allowed the new claims and the '551 issued. :|: * * At the time of the interview and the submission, Attorney Pope and Dr. Sangh-era were well aware of previous representations based on the same “optionally, but preferably” sentence made by Medisense to the European Patent Office in 1994-95. Attorney Pope and Dr. Sanghera, however, made a conscious and deliberate decision to withhold disclosure to the PTO of these prior statements. This much is conceded. Abbott contends, however, that there was no duty to disclose the earlier statements and that there was no intent to deceive. On these latter points, the following was proven at trial by clear and convincing-evidence. The '636 patent — the European counterpart to the '382 — had been revoked in 1993 in an opposition proceeding based on a German prior-art reference called Dl. In 1994, Medisense appealed, arguing that D 1 was distinguishable on two grounds. The centerpiece of Medisense’s appeal relied on the very sentence in question — the “optionally, but preferably” sentence. Overall, the '636 and '382 specifications were virtually identical. In both, the “optionally, but preferably” sentence and its immediate context were completely identical. Before the EPO, however, Medisense had an incentive to advance the sentence as an important teaching over the Dl reference. Medisense submitted that the “optionally, but preferably” sentence demonstrated that the '382/'636 invention did not need a membrane for measuring glucose in blood, whereas the D 1 device had required one. Specifically, Dl had disclosed an enzyme electrode usable for glucose and covered by a semipermeable membrane. Before the EPO, Medisense argued that the Dl membrane was essential to the Dl invention. By contrast, Medisense stated that the '382/'636 membrane was merely optional. Medisense relied on the “optionally, but preferably” sentence as follows (TX 311 at AL54151): 10. The above object is solved by a glucose sensor as defined in claim 1 of the patent in suit ['382/'636j. Apart from the important feature of utilizing a ferrocene or ferrocene derivative as mediator, another important difference over Dl resides in that the claimed glucose sensor — contrary to that of Dl which requires a membrane — does not have and must not have a semipermeable membrane within the meaning of Dl. Contrary to the semipermeable membrane of Dl, the protective membrane optionally utilized with the glucose sensor of the patent [in] suit is not controlling the permeability of the substrate (as set forth above under IV.2), in the membrane of Dl the permeability for the substrate must be kept on a low value to achieve a linear relationship between the measures [sic] currency and the substrate concentration in the test solution. Rather, in accordance with column 5, lines 30 to 33 of the patent in suit: “Optionally, but preferably when being used on live blood, a protective membrane surrounds both the enzyme and the mediator layers, permeable to water and glucose molecules.” See also claim 10 of the patent in suit as granted according to which the sensor electrode has an outermost protective membrane (11) permeable to water and glucose molecules. Finally, see Example 7 in column 10, lines 19 to 26 reporting that by using such a protective membrane the response time did not increase but from 24 to 60 sec. (without membrane) to 36-76 sec. (with membrane). Accordingly, the purpose of the protective membrane of the patent in suit, preferably to be used with in vivo measurements, is a safety measurement to prevent any course particles coming off during use but not a permeability control for the substrate. The passage indented the “optionally, but preferably” sentence for emphasis, just as set out above. The bolded words were bolded by Medisense, just as set out above. The foregoing quotation is exactly the way it was made by Medisense in January 1994. The “safety” purpose stated in the quotation helped to show, it deserves to be said, why a protective membrane was merely “preferred” for live blood, i.e., in vivo testing. It was optional in all cases but when placed in a human bloodstream, a membrane was advisable to retain the chemistry aboard the electrode and, thusly, prevent toxic particles from circulating within the patient. In the same submission (TX 311 at AL54154), Medisense stated that Dl was “strongly teaching away from the subject matter as claimed [in the '382/'636] which not only does not require a membrane but must not have a membrane. In other words, with the claimed subject matter, rather than keeping the permeability for the substrate at a low level, there is free access of the substrate