Full opinion text
OPINION & ORDER STEIN, District Judge. After trial on the merits, this Court finds that Apotex, Inc. and Apotex Corporation (collectively, “Apotex”) have failed to prove by clear and convincing evidence that U.S. Patent No. 4,847,265 is invalid or unenforceable on any of the grounds asserted. Accordingly, Sanofi-Aventis, Sa-nofi-Synthelabo, Inc. and Bristol-Myers Squibb Sanofi Pharmaceuticals Holding Partnership (collectively, “Sanofi”) are entitled to permanent injunctive relief and, as shall be determined by the Court in a future proceeding, damages. This action concerns a patent dispute between Sanofi — which invested in the research and development to patent and bring to market the drug known as Pla-vix® — and Apotex, which seeks to market the generic equivalent of that drug. On the basis of the record established by the parties and the applicable law, the Court enters the following findings of fact and conclusions of law pursuant to Fed. R.Civ.P. 52(a). FINDINGS OF FACT I. History of this Action Plavix®, approved for sale in the United States by the U.S. Food and Drug Administration (“FDA”) in November 1997, is prescribed for the reduction of thrombotic events such as heart attacks and strokes for patients who have recently suffered those events or who have arterial disease or acute coronary syndrome. (See Stipulated Statement of Facts (“Fact Stmt.”), attached as Exhibit A to Joint Pretrial Order dated May 27, 2005 at ¶ 12.) The active ingredient of Plavix® is clopidogrel bisulfate. (Id.) Sanofi obtained a patent claiming clopidogrel bisulfate on July 11, 1989, naming Sanofi employees Alain Ba-dorc and Daniel Fréhel as inventors. (Id. at ¶¶ 8-9.) That patent, U.S. Patent No. 4,847,265 (“the '265 patent”), claims clopi-dogrel bisulfate by its chemical name in Claim 3: “Hydrogen sulfate of the dextro-rotatory isomer of methyl alpha-5(4,5,6,7tetrahydro(3,2-c)thieno pyridyl) (2-chloro-phenyl)-acetate substantially separated from the levo-rotatory isomer.” ('265 patent at col. 12,11. 37-41.) Sanofi-Aventis is the owner of the patent-in-suit, which expires on November 17, 2011. The foreign priority filing date of the '265 patent — the date on which Sanofi filed its earlier application for the corresponding French patent — is February 17, 1987. The '265 patent is exclusively licensed to Bristol-Myers Squibb Sanofi Pharmaceuticals Holding Partnership. (Fact Stmt, at ¶¶ 3,13.) Apotex sought approval from the FDA to manufacture and sell clopidogrel bisul-fate tablets before the expiration of the '265 patent by filing an Abbreviated New Drug Application (“ANDA”) with the FDA in November 2001. (Id. at ¶¶ 14-15.) In the ANDA, Apotex certified pursuant to the requirements of 21 U.S.C. § 355(j) (2)(A)(vii)(IV) that it believed the '265 patent to be invalid. (Id. at ¶ 16; see also Glaxo Group Ltd. v. Apotex, Inc., 376 F.3d 1339, 1344 (Fed.Cir.2004).) Apotex was the first to file an ANDA for clopidogrel bisulfate (Decl. of Dr. Bernard Sherman, dated Aug. 16, 2006 (“Sherman Decl.”) at ¶ 17), thereby securing the right to 180 days of market exclusivity provided by the Hatch-Waxman Act to the first ANDA filer to challenge a patent. See 21 U.S.C. § 355(j)(5)(B)(iv); see also In re Tamoxifen Citrate Antitrust Litig., 429 F.3d 370, 376 (2d Cir.2005). In response to that ANDA filing by Apotex, Sanofi filed this litigation against Apotex on March 21, 2002 pursuant to 35 U.S.C. § 271(e), and asserted that Apo-tex’s filing of the ANDA constituted infringement of the '265 patent, specifically Claim 3. (See Fact Stmt, at ¶ 17.) Apotex counterclaimed, asserting that the '265 patent is both invalid for three separate reasons and unenforceable as well. First, Apotex alleges that the '265 patent is anticipated pursuant to 35 U.S.C. § 102(b) by an earlier patent held by Sa-nofi that covered a genus of chemical compounds called thienopyridines, within which clopidogrel bisulfate falls. (See Fourth Amended Answer and Amended Counterclaim (“Answer”), filed Nov. 17, 2006, at 3, 13.) The earlier patent, U.S. Patent No. 4,529,596 (“the '596 patent”), issued in July 1985 and expired in July 2003. (Fact Stmt, at ¶ 28.) Second, Apotex contends pursuant to 35 U.S.C. § 103 that the subject matter claimed in the '265 patent would have been obvious to a person of ordinary skill in the art at the time the invention was made. (See Answer at 3,13.) Third, Apotex contends that the patent is invalid under the judicial doctrine of obviousness-type double patenting. (Id. at 6, 15.)' Finally, Apotex also asserts that the '265 patent is unenforceable on the basis of Sanofi’s alleged inequitable conduct before the U.S. Patent and Trademark Office (“PTO”). The alleged conduct consists of failing to name Dr. Jean-Pierre Maffrand as an inventor, making false statements to the PTO regarding the unexpected pharmacological properties of clopidogrel bisul-fate, failing to disclose relevant prior research that Sanofi had conducted on a similar chemical compound, and failing to disclose a journal article that Apotex alleges is a material prior art reference. (Id. at 4-6, 8-15.) With regard to infringement, the parties have stipulated that Apotex’s clopidogrel bisulfate product infringes Claim 3 of the '265 patent. (See May 7, 2004 Stipulation and Order; Fact Stmt, at ¶ 18.) The ensuing procedural history of this action — including an account of the extensive settlement negotiations between the parties and the consequences for this litigation — is described in detail in this Court’s Opinion dated August 31, 2006; familiarity with that Opinion is assumed. See Sanofi-Synthelabo v. Apotex Inc., 488 F.Supp.2d 317, 323-26 (S.D.N.Y.2006). In brief, Apotex initiated an at-risk launch of its generic clopidogrel bisulfate product on August 8, 2006. Shortly thereafter, Sanofi moved for a preliminary injunction prohibiting Apotex from distributing its generic product. After an evidentiary hearing was held, the Court ranted Sanofi’s motion for preliminary injunctive relief on August 31, 2006, but denied its request for a recall of the approximately six-month supply of product that Apotex had already shipped to distributors in the United States. Id. at 347-349. Apotex then moved to stay the preliminary injunction, but both this Court and the U.S. Court of Appeals for the Federal Circuit denied that motion. Apotex then appealed the order granting the preliminary injunction to the Federal Circuit, which affirmed the grant of the preliminary injunction in an opinion dated December 8, 2006. See Sanofi-Synthelabo v. Apotex, Inc., 470 F.3d 1368 (Fed.Cir.2006), reh’g denied, 2007 U.S.App. LEXIS 2807 (Fed.Cir. Jan. 19, 2007). The parties then tried the merits of this action before this Court without a jury from January 22 through February 15, 2007. That trial — as well as the evidentiary hearing held in August 2006 — adduced the following facts. II. Sanofi’s Efforts to Develop Antipla-telet Aggregation Drugs A. Ticlopidine In 1972, Dr. Jean-Pierre Maffrand, then a Sanofi chemist in charge of a small research team, was asked by his supervisor to synthesize compounds structurally similar to tinoridine, a drug with known anti-inflammatory properties, in an effort to discover a superior anti-inflammatory drug. (Maffrand Tr. 1570-74.) Tinori-dine is a member of a class of compounds known as thienopyridines. (Maffrand Tr. 1575-76.) The chemical structure of a thienopyridine is characterized by the fusion of a thiophene ring to a pyridine ring. (Maffrand Tr. 1575-76.) During 1972 and 1973, Sanofi synthesized a number of thienopyridine compounds. (Maffrand Tr. 1573-74.) Tests on those compounds revealed that unlike tinoridine, they possessed no anti-inflammatory properties. Some of the compounds, however — one of which Sanofi named “ticlopidine” — exhibited antiplatelet aggregation activity. (Maffrand Tr. 1573-75.) Put simply, the compound helped ensure that platelets in the blood would not aggregate together as much as they otherwise would. This was an important discovery. Researchers at Sanofi were convinced that platelets play a major role in events such as myocardial infarction and brain ischemia and Dr. Maffrand and his colleagues were interested to find an anti-platelet aggregation agent that would be “[a] better drug than aspirin.” (Maffrand Tr. 1574-75.) In 1977, Sanofi obtained a patent on ticlopidine; that patent expired in 1994. Ticlopidine was introduced as a drug in France in 1978 and in the United States in 1991, where it was marketed under the brand name “Ticlid.” (Maffrand Tr. 1578-79; 1581-82.) Soon after the launch of ticlopidine in France, Sanofi became aware of rare but potentially fatal side effects associated with ticlopidine, specifically blood disorders known as neutropenia and thrombotic thrombocytopenic purpura (“TTP”). (Schneller Tr. 763, 815-16; Maffrand Tr. 1580-81.) As a result, the FDA required Ticlid to carry a “black box” warning that ticlopidine could cause life-threatening blood disorders. (Maffrand Tr. 1581.) The potential for these serious side effects meant that patients taking ti-clopidine had to be clinically monitored for signs of blood disorders. (Maffrand Tr. 1581.) The sub-optimal side effect profile of ticlopidine — i.e., the risk of developing serious blood disorders — left open the need in the market for a drug that was as effective or more effective than ticlopidine, but with a lower risk of side effects. B. PCR 10SS In 1975, Sanofi synthesized a thienopyri-dine named PCR 1033. That compound is the methyl analog of ticlopidine, which means that one of the two hydrogen atoms on the bridge carbon of a ticlopidine molecule — the carbon linking the thienopyri-dine and phenyl substituents — is substituted with a methyl (CH3) group. (Maffrand Tr. 1584.) The addition of the methyl group makes the bridge carbon asymmetric. Here, some knowledge of the basic principles of stereochemistry is required and has been best set forth by the Federal Circuit in its opinion affirming the preliminary injunction in this action: Stereochemistry refers to the three-dimensional spatial arrangement of a molecule’s constituent atoms. Molecules that have the same chemical substitu-ents, but different spatial arrangements, are referred to as stereoisomers. If they contain an asymmetrical carbon atom, they exist as non-superimposable mirror images of each other and are referred to as enantiomers. Enantiom-ers are optically active because they are capable of rotating plane-polarized light; enantiomers that rotate polarized light to the right are referred to as dextroro-tatory enantiomers, or d-enantiomers; enantiomers rotating polarized light to the left are referred to as levorotatory enantiomers, or 1-enantiomers. A mixture of equal amounts of both types of enantiomers is referred to as a racemic mixture, or racemate, and it exhibits no optical activity. Sanofi-Synthelabo v. Apotex, Inc., 470 F.3d at 1372. Unlike ticlopidine, PCR 1033 is a race-mate composed of the 50:50 mixture of its enantiomers. (Maffrand Tr. 1584, 1599-1600, 1682-83.) To make PCR 1033 suitable for testing, Sanofi chemist Alain Ca-brol attempted to prepare the hydrochloride salt of PCR 1033, but that effort failed. (Maffrand Tr. 1599-1600; 1699; Badorc Tr. 1810.) Instead, Sanofi prepared PCR 1033 as a maleate salt. (Maff-rand Tr. 1600.) When tested, Sanofi found that PCR 1033 was a more potent antipla-telet aggregation agent than ticlopidine, but that it was also less well-tolerated than ticlopidine; side effects were observed during the initial pharmacological testing of PCR 1033 in different species. (Maff-rand Tr. 1587.) Around the beginning of 1976, Sanofi ceased efforts to develop PCR 1033 because of those side effects. (Maff-rand Tr. 1587.) In 1978, Maffrand asked Alain Badorc, a Sanofi chemist, to try to obtain the enan-tiomers of PCR 1033. (Maffrand Tr. 1587-88; Badorc Tr. 1807.) Maffrand testified at trial that “we wanted to try our luck” to see if either enantiomer had a better ratio of antiplatelet activity to tolerance than the racemate. (Maffrand Tr. 1588.) At that time, Maffrand knew that a two-fold increase in a single enantiomer’s therapeutic activity was the quantitatively best possible increase that could be attained through separation of the enantiom-ers of a racemate, and that such a result “is not the most frequent case.” (Maff-rand Tr. 1591-93; Harden Tr. 2304-06.) For that reason, Sanofi’s preferred mode of searching for superior antiplatelet drugs at that time had been to modify the sub-stituents of racemic compounds (thereby creating new compounds) — not to isolate the individual enantiomers of particular racemates — because a modified compound was the more likely path to significantly greater gains in antiplatelet activity. (Harden Tr. 2304; Maffrand Tr. 1592-94.) Badorc successfully separated the enan-tiomers of PCR 1033 on his first attempt by means of a technique known as diaster-eomeric salt formation. (Badorc Tr. 1809; Maffrand Tr. 1702) That method involves combining the enantiomers of a racemic base with an enantiomerieally pure chiral acid to form a diastereomeric salt (i.e., a salt with two chiral centers) in two distinct stereoconfigurations. (Davies Tr. 1934.) The resulting diastereomers, if formed, do not have identical physical properties; their different solubilities in a particular solvent, for example, may facilitate their separation. (Davies Tr. 1934-35; McClel-land Tr. 1105-06.) However, the Court credits the testimony of Dr. Stephen G. Davies, an expert for Sanofi, who stated, “[t]he problem is that it is very hard to find situations where one [diastereomer] crystallizes out and the other one does not .... you are trying to get one to do something while leaving the other one behind.” (Davies Tr. 1935-36.) If the diastereom-ers are successfully separated, a base is added to one of the two separated diaster-eomeric salts to reconstitute the original base, which is then in an enantiomerieally pure form. (McClelland Tr. 1108-1109; Davies Tr. 1936-37.) Badorc prepared the diastereomeric salts of PCR 1033 by using tartaric acid dissolved in ethanol. (Badorc Tr. 1808-09.) The levorotatory enantiomer was designated PCR 3071 and the dextrorota-tory enantiomer was designated PCR 3072. (Maffrand Tr. 1598.) To facilitate further testing, Badorc prepared the hydrochloride salts of PCR 3071 and PCR 3072. (Badorc Tr. 1810.) As noted, previous efforts to prepare the hydrochloride salt of the racemate — PCR 1033 — had failed. (Maffrand Tr. 1600, 1699.) Testing on the enantiomers revealed that PCR 3071 exhibited antiplatelet activity and PCR 3072 was inactive; the active enan-tiomer, however, was less well tolerated than ticlopidine and was not appropriate for administration to humans. (Maffrand 1598-99.) Sanofi discontinued the development of those enantiomers in 1981. (Maffrand Tr. 1599.) C. PCR 354.9 In 1978 — after Sanofi had synthesized PCR 1033 but before it had discontinued the development of PCR 3071 and PCR 3072 — Sanofi synthesized the ethyl analog of ticlopidine, which was designated PCR 3233. (Maffrand Tr. 1601-02.) Attempts to make the hydrochloride salt of PCR 3233 — an oily base — failed, but, after several failed attempts, Badorc eventually obtained its nitrate salt, which was designated PCR 3549. (Maffrand Tr. 1604; Badorc Tr. 1812.) Testing revealed that PCR 3549 was more potent than ticlopi-dine and better tolerated than PCR 1033, but was still less well tolerated than ticlo-pidine. (Maffrand Tr. 1604-05.) In November 1978, Maffrand asked Badorc to prepare the enantiomers of PCR 3549 to see if either enantiomer had a better rislqbenefit profile than the racemate. (Maffrand Tr. 1605-06; Badorc Tr. 1811.) Badorc tried to separate the enantiomers of PCR 3549 using diastereomeric salt formation — the technique that he had used successfully to resolve the enantiom-ers of PCR 1033 — but several attempts failed. (Maffrand Tr. 1606; Badorc Tr. 1813-14.) Badorc then tried to obtain the enantiomers of PCR 3549 by a different method — chemical asymmetric synthesis. (Maffrand Tr. 1606; Badorc Tr. 1819-20.) That method involves taking a precursor compound that is enantiomerieally pure and modifying the compound to obtain the desired compound, but without altering the stereochemical configuration of the precursor compound. (Davies Tr. 1932-33; Badorc Tr. 1819-20.) At that time, Badorc believed — and he testified at trial that he still believes — that he successfully-synthesized the enantiomers of PCR 3549 by that method. (Badorc Tr. 1823-24.) Badorc further testified that he confirmed his successful synthesis of the enantiom-ers of PCR 3549 in 1978 by measuring the rotatory power of each enantiomer, which were equal in absolute value (Ba-dorc Tr. 1824); that result — the equal and opposite optical rotation of two products— is evidence of the presence of two enan-tiomers. (Davies Tr. 1981; Hendrickson Tr. 1474.) In addition, the nitrate forms of each enantiomer had equivalent melting points that were each higher than the melting point of PCR 3549. (Davies Tr. 1981.) Furthermore, “the chance is minimal, if not nonexistent,” that two parallel asymmetric syntheses — such as those Ba-dorc performed — would result in products demonstrating the same level of rotatory power and the same melting points. (Davies Tr. 1819.) The Court credits the testimony from Dr. Davies that this data provided strong evidence that Badorc had indeed obtained the enantiomers of PCR 3549 by means of asymmetric synthesis. (Davies Tr. 1981-82.) Sanofi designated the levorotatory enan-tiomer PCR 3620 and the dextrorotatory enantiomer PCR 3621 and tested them for platelet aggregation inhibition activity. (Badorc Tr. 1826.) That testing showed that the platelet aggregation inhibiting activities of PCR 3620 and PCR 3621 were each comparable to PCR 3549, meaning neither improved on the racemate. (Maff-rand Tr. 1615-16.) For that reason, Sano-fi discontinued development of PCR 3620 and PCR 3621 and focused its efforts on PCR 3549. (Maffrand Tr. 1616-17.) PCR 3549, however, proved to be less well-tolerated than ticlopidine at similar dosage levels, and lower doses of PCR 3549 were insufficiently therapeutic. For those reasons, Sanofi abandoned its efforts to develop PCR 3549. (Maffrand Tr. 1616-17.) D. PCR J/,099 Sanofi synthesized several other thieno-pyridines in addition to PCR 1033 and PCR 3549. After having created two car-boxylic acid derivatives of ticlopidine — in 1976 and 1978, respectively — which showed no platelet inhibition, Sanofi synthesized a new compound in 1980 that had an ethyl ester (0=C-0-CH2-CH3) substi-tuent on the bridge carbon. (Maffrand Tr. 1637-39.) That compound — PCR 3935— proved active in tests of antiplatelet aggregation activity (Maffrand Tr. 1638), and, after attempts to prepare it as both hydrochloride and bisulfate salts had failed, Sa-nofi prepared it as a hydrobromide salt. (Maffrand Tr. 1640.) Concerned, however, that PCR 3935 might be converted in the human body into carboxylic acid — which could render the compound therapeutically inactive — Sanofi synthesized a new compound in July 1980 by replacing the ethyl ester substituent of PCR 3935 with a methyl ester (0=C-0-CH3) substituent. (Maffrand Tr. 1638-40.) Sanofi designated that compound — a racemic mixture prepared as a hydrochloride salt — as PCR 4099. (Maffrand Tr. 1640-41.) The non-salt (free base) form of PCR 4099 is described by the chemical formula methyl cu-tí,5,6,7-tetrahydro-thieno(3,2-c)-5-pyri-dyl)-o-ehlorophenyl-acetate, or “MATTP-CA.” The nomenclature “methyl a-(4,5,6,7-tetrahyd ro-thieno(3,2-c)-5-pyri-dyl)-o-chlorophenyl-aeetate” is equivalent to “methyl alpha-5(4,5,6,7-tetrahydro(3,2-c)thieno pyridyl) (2-chlorophenyl)-ace-tate.” (Fact Stmt, at ¶ 31.) Testing showed that PCR 4099 exhibited more antiplatelet activity and was better tolerated than ticlopidine. (Maffrand Tr. 1641.) Nonetheless, Sanofi continued to synthesize and test other racemic carbox-ylic acid and ester derivatives of ticlopi-dine, as well as derivatives with amide (0=C-NR2) substituents. (Maffrand Tr. 1641-43.) In total, Sanofi synthesized approximately 70 carboxylic acid, ester, and amide derivatives of ticlopidine, all of which were racemic mixtures. (Maffrand Tr. 1636-1637; 1642.) Many of these compounds were subjected to a battery of tests of pharmacological activity, including four tests known as the bleeding-time test, the silk-thread test, the collagen test, and the AJDP test. (Maffrand Tr. 1641; Hanson Tr. 2226, 2232-33, 2238-39, 2243, 2246-67.) These were screening tests “designed to quickly screen a large number of candidate compounds in order to identify a manageable number of compounds [for] more sophisticated testing.” (Hanson Tr. 2226-27; see also Maffrand Tr. 1644, 1646.) Sanofi’s expert Dr. Stephen R. Hanson gave testimony at trial — which the Court credits — that conclusions drawn from such tests are “tentative” and that “it would be risky to extrapolate the results of these tests” — performed only in rats — to other animals or humans. (Hanson Tr. 2227.) III. The '596 Patent and Its Nonr-U.S. Counterparts In July 1982 and July 1983, Sanofi applied for French and U.S. patents, respectively, on the class of thienopyridines having carboxylic acid, amide, and ester substituents. An internal Sanofi memorandum authored by Dr. Maffrand and dated February 11, 1982 stated that the patent would cover “PCR 4099 and its analogues.” (Defendant’s Exhibit (“Def. Ex.”) 632.) The memorandum also noted that based on pharmacological testing, “[t]he most interesting esters were: PCR 4099, 4316, SR 24593, [and] SR 24597” and “[t]he most interesting amides were: PCR 4317, 4535, [and] 4516.” (Id.) The application that was filed in the United States matured into the '596 patent, which issued on July 16, 1985 and listed Maffrand, Aubert, and Ferrand, among others, as inventors. The '596 patent is entitled “Thieno [3,2-c] Pyridine Derivatives and Their Therapeutic Application,” and describes a genus of compounds that exhibit blood-platelet aggregation inhibition and anti-thrombotic activity. (Fact Stmt, at ¶ 27; '596 patent at col. 8, 11. 26-39.) Sanofi also obtained European patent EP099802 (“the '802 patent”) and Canadian Patent No. 1,194,875 (“the '875 patent”) over the same class of compounds. These patents were also published more than a year before February 17, 1987, the foreign priority filing date of the '265 patent. A. The Specification of the '596 Patent The general formula described in column 1 of the '596 patent specification— which has two variables (X and Y) — depicts a racemate in non-salt (free base) form. The general formula covers an extremely large number of compounds covered by providing for various substituents in place of the X and Y groups. (McClel-land PI Tr. 494; Davies Tr. 1913-14). Each of the X and Y variables can be one of a number of enumerated substituents, with thirty-seven possibilities for the X variable and 1710 possible choices for Y. Multiplying these variables by the number of pharmaceutically acceptable acid addition salts yields an even higher number. Dr. Maffrand testified at trial that the patent covers “millions of compounds if we take into account all the combinations possible of products and salts.” (Maffrand Tr. 1771.) Similarly, Sanofi’s expert Dr. Davies testified that the patent covers “several millions of compounds.” (Davies Tr. 1914.) The general formula does not describe any individual enantiomer and does not describe how to obtain the enantiomers of any racemate; as drawn, the general formula describes a racemic mixture only. (Davies Tr. 1913.) However, the specification states that with respect to the general formula, “[t]he invention relates both to each enantiomer and their mixture.” ('596 patent at col. 1, 11. 40-41.) The specification also sets forth that the thienopyridine compounds within the genus covered by Claim 1 can be made into addition salts with pharmaceutically acceptable mineral or organic acids or a mineral base. ('596 patent at col. 1,11. 42-51.) In addition, the '596 patent specification gives twenty-one examples of compounds included within the genus compound “to exemplify and to illustrate the different substituents which were claimed in the general formula.” (Maffrand, PI Tr. 138; '596 patent at cols. 3-8.) These different compounds are described in the patent as different salts forms — including hydro-bromide, hydrochloride, and bisulfate salts. (Maffrand, PI Tr. at 139-140.) Specifically, three of the examples are shown as hydrochloride salts and four of the examples are shown as bisulfate salts. (Snyder Tr. 236.) The first of the twenty-one examples in the '596 patent — Example 1 — is PCR 4099, prepared as a hydrochloride salt. (Maffrand, PI Tr. at 139.) Neither Example 1, nor any other example set forth in the specification, refers to the individual or separated enantiomers of PCR 4099 or, for that matter, of any other exemplified compound. Furthermore, there is no example in the '596 patent of a methyl ester — such as PCR 4099 — prepared as a bisulfate salt. (Banker Tr. 1328.) In addition, the pharmaceutical formulations provided as non-limiting examples at the end of the specification do not show a bisulfate salt in the solid dosage form. (McClelland Tr. 1178-79; '596 patent at col. 12,11. 3-28.) The specification of the '596 patent also describes the compounds covered by the patent as exhibiting “an excellent tolerance and a low toxicity.” ('596 patent at col. 8, 11. 42^13.) In the 1980s, Sanofi scientists commonly used the terms “tolerance” and “tolerated” to denote a compound’s relative toxicity and described compounds that performed well in sub-acute and acute toxicity tests as “not very toxic” or “better.” Dr. Maffrand testified that he considers “toxicity” to be “included in tolerance.” (Maffrand Tr. 1664.) Dr. Frédéric La-cheretz, the deputy head of Sanofi’s toxicology department in the mid-1980s, testified that the terms “ ‘more toxic than’ [or] ‘less well-tolerated than’ ... mean the same thing.” (Lacheretz Tr. 2392.) In addition, Dr. Hanson testified that the terms toxicity and tolerance are “not necessarily mutually exclusive.” (Hanson Tr. 2224.) Similarly, Apotex’s expert, Dr. Gilbert Banker, agreed that “not toxic” means that a substance “won’t make you sick, won’t affect your physiologic functions, won’t deleteriously affect your organs or even produce untoward side effects,” and that non-toxic has the same meaning as “well tolerated.” (Banker Tr. 1294.) The Court credits Dr. Hanson’s testimony that “[u]nless you’re talking about a formally-described toxicology or tolerance study, the term tolerance ... is commonly used as a general term that can describe toxic side effects as well as less severe side effects.” (Hanson Tr. 2225-26.) Furthermore, the Court finds that the use of the terms “well-tolerated” and “less well-tolerated” is not restricted to descriptions of a drug’s performance in tests at therapeutic doses only. (Graham Tr. 2424.) Apotex has failed to persuade the Court that those terms cannot appropriately be used to describe performance in tests at supra-therapeutic doses, including subacute and acute toxicity tests. (See Graham Tr. 2424; Hanson Tr. 2224; Lacheretz Tr. 2359; Ro-dricks Tr. 2495.) B. The Claims of the '596 Patent As described above, Claim 1 sets forth a general formula that covers millions of compounds, including “their addition salts with pharmaceutically acceptable mineral or organic acids ... including both enan-tiomeric forms or their mixture.” ('596 patent at col. 12,11. 30-68, col. 13,11. 1-19.) Claim 8 covers the same genus of compounds as Claim 1, but claims their “therapeutic compositions.” Claim 8 reads: A therapeutic composition having blood-platelet aggregation inhibiting activities and anti-thrombotic activities containing an effective amount of a compound of claim 1, or an addition salt thereof with a pharmaceutically acceptable mineral or organic acid with mineral bases, or one of the two enantiomers or their mixture and a pharmaceutically acceptable carrier. ('596 patent at col. 14,11. 5-11.) Claims 2 through 7 correspond to particular racemates, and Claim 2 specifically claims MATTPCA, or the free base of PCR 4099: Methyl a-(4,5,6,7-tetrahydro-thieno(3,2,c)-5-pyridyl)-o.chlorophe-nyl-acetate. ('596 patent at col. 13,11.19-20.) Claims 2 through 7 do not include the language “both enantiomeric forms or their mixture” or “their addition salts.” Claim 2, therefore, does not expressly claim any salt form of MATTPCA. (Davies Tr. 1917-19; Byrn Tr. 2134.) In addition, Claim 2 does not include a stereo-chemical descriptor (e.g., “d-” or “£-”) that identifies it as referring to an individual enantiomer. C. Tests of Pharmacological Activity Reported in the '596 Patent As noted above, prior to Sanofi’s application for the '596 patent, several racemic compounds covered by the general formula of Claim 1 were tested for platelet aggregation inhibition activity. Results of those tests — the bleeding-time test, the silk-thread test, the collagen test, and the ADP test — were reported in the specification to the '596 patent. (Hanson Tr. 2226, 2232-33, 2243, 2246.) Table I reports the results of the ADP test in which nine compounds, including PCR 4099 — which is listed as Derivative 1, were tested; most of the compounds demonstrated significant activity. At the doses at which they were tested, Derivative 4 demonstrated activity that was not meaningfully different from that of PCR 4099. (Snyder Tr. 633-35; '596 patent at col. 9, 11. 31-55.) In the collagen test reported in Table II, twelve compounds were tested, and, again, most of the tested compounds demonstrated significant activity. At the doses that were tested, Derivatives 9, 4, and 10 demonstrated activity quantitatively superior to that of PCR 4099, although Sanofi’s expert Dr. Hanson described the activity levels of those compounds as comparable to that of PCR 4099. (Hanson Tr. 2244-46; '596 patent at col. 10, 11. 1-31.) In the bleeding-time test, reported in Table III, six compounds were tested. All the tested compounds, including PCR 4099, demonstrated maximal inhibition of blood clotting activity. (Snyder Tr. 641; Hanson Tr. 2232-33, 2237-38; '596 patent at col. 10, 1. 55, col. 11, 1. 5.) In the silk-thread test reported in Table IV, most of the six tested compounds demonstrated significant activity. At the doses that were tested, two of the six compounds, Derivative 3 and Derivative 10, demonstrated activity quantitatively superior to that of PCR 4099. (Snyder Tr. 641-42; '596 patent at col. 11, 11. 35-51.) Again, Sanofi’s expert Dr. Hanson testified, however, that “[f]or compounds No. 3, 10, [PCR] 4099, and No. 2, all of which did very well, it is not possible to further rank their effects.” (Hanson Tr. 2241.) As noted above, Dr. Hanson testified that the pharmacological activity data presented in Tables I-IV of the '596 patent is not an appropriate scientific basis on which to rank the relative performance of the compounds — except in cases where a high dose produces little or no activity. (Hanson Tr. 2226-27, 2241, 2246.) These were screening tests used to make rough determinations of activity versus inactivity. In addition, because the screening tests were performed in rats, inferences as to how the tested compounds would perform if administered to humans would be tentative because many drugs behave differently in rats and human beings. As a result, it “would be risky to extrapolate the results of these tests” to any other animal or to human beings. (Hanson Tr. 2227-29.) Furthermore, the tests did not provide enough data points to construct a dose-response curve for each compound; a dose-response curve is a tool used to compare the relative potency of different compounds. (Hanson Tr. 2235.) The lack of data points made it impossible to derive an “ED50” value, the dose at which each compound produces half its maximal response, which is a similar tool of comparison. (Snyder Tr. 630-32; Maffrand Tr. 1646; Hanson Tr. 2236-37.) Finally, nearly all of the compounds were tested at only one or two dose levels (Snyder Tr. 642-43), making it impossible to draw conclusions about how the compounds would perform at lower doses. (Hanson Tr. 2238.) For these reásons, the Court finds that a person of ordinary skill in the art would not discern a preference for PCR 4099 based on the pharmacological testing data in the '596 patent. (Hanson Tr. 2250.) D. The Canadian '875 Patent As noted above, the '875 patent is the Canadian counterpart to the '596 patent. (Def. Exs. 206; 1235.) Claim 1 describes a “procedure for preparing derivatives of general formula (I)” — the same general formula claimed in Claim 1 of the '596 patent. With regard to the general formula, the '875 patent also claims “the two enantiomers or their mixture.” In addition, after describing the procedure for preparing certain compounds, the claim states that “then the corresponding derivative sought is obtained, which is isolated and, if desired, its enantiomers separated and/or it is salified by mineral or organic acid action.” (Def. Ex. 1235.) E. The European '802 Patent U.S. Patent No. 5,989,578 (“the '578 patent”) — a patent held by Sanofi which issued in 1999 — states that clopidogrel is “described” in EP 099 802, the European counterpart to the '596 patent. Two factors militate against giving that statement any credence. First, the application for the European '802 patent was filed in July 1983, well before Sanofi’s separation of PCR 4099 into its enantiomers or the formulation of the dextrorotatory enantiomer of PCR 4099 — clopidogrel—as a bisulfate salt, as described below. (Maffrand Tr. 1672-73.) Second, a named inventor of the '578 patent — Dr. Pierre Savi — testified that the statement in that patent that clo-pidogrel is described by the European '802 patent is “incorrect,” and that he had not personally read the European patent until his deposition relating to this action. (Savi 8/21/03 Dep. 164-65, 314-17.) Dr. Savi also testified that other named inventors listed on the '578 patent are biologists, not chemists. (Id. at 317.) Accordingly, the Court does not credit the alleged disclosure in the '578 patent and does credit Dr. Savi’s unequivocal later statement that clo-pidogrel is not described by the '802 patent. IV. The Development of PCR 1099 and Its Enantiomers Sanofi’s decision to further develop PCR 4099 required clinical testing, which in turn required Sanofi to synthesize large quantities of the drug in tablet form. From 1980 until 1987, PCR 4099 was the subject of more than fifty different types of tests in animals and humans, and Sanofi spent “tens of millions of dollars” on those tests. (Maffrand PI Tr. 144.) Sanofi also performed additional pharmacological and toxicological studies, as well as metabolic and pharmacokinetic studies. (Maffrand PI Tr. 143.) Between 1983 and 1987, Sa-nofi conducted acute and long-term chronic toxicity tests on PCR 4099 in rats, mice, and baboons. In each test and in each species, convulsions were observed at certain dose ranges. (Lecheretz Tr. 2362-70; Rodricks Tr. 2481-83.) A one-year chronic toxicity study in baboons in 1987 showed convulsions at all doses, including the lowest dose tested, 25 mg/kg, and demonstrated that more convulsions were observed at higher doses. (Rodricks Tr. 2481-82.) During 1985, Sanofi representatives, including Dr. Maffrand, made poster presentations and distributed corresponding abstracts concerning their work on thieno-pyridines — and, in particular, PCR 4099— at conferences of the International Society for Thrombosis and Hemostasis held in San Diego and Jerusalem. (Maffrand Tr. 1680; see also Def. Exs. 183, 184, 185, 186, 188, 224, 308, 418, 427.) The abstract prepared for the San Diego conference regarding PCR 4099 stated that the compound “exhibits the same broad spectrum of antiaggregating effect as ticlopidine in animals but is 40 times more potent in rats and 10 times in baboons.” (Def. Ex. 184.) According to Sanofi’s expert Dr. Hanson, the abstracts prepared for both conferences informed a person of ordinary skill in the art that “PCR 4099 exhibits activity in both humans and animals against platelet aggregation, and specifically the activity seems to be specially active toward the ADP pathway of platelet aggregation.” (Hanson Tr. 2219-20.) Dr. Maffrand testified that Sanofi’s “main goal was to inform the scientific community of the potent antiplatelet and antithrombotic activities of PCR 4099” and “to show that the compound was much more effective than ticlopidine.” (Maffrand Tr. 1680.) PCR 4099 was the only compound covered by the '596 patent that was publicized at the San Diego and Jerusalem conferences. (Maffrand Tr. 1736, 1748-49.) The posters and abstracts did not, however, discuss in any way the stereoselectivity of platelet inhibition by the enantiomers of PCR 4099 or, for that matter, make any reference at all to the enantiomers of PCR 4099. (Maffrand Tr. 1681-86; Hanson Tr. 2220.) Accordingly, a person of ordinary skill in the art would not have drawn any inference from those materials concerning the stereoselectivity of platelet inhibition by the enantiomers of PCR 4099. (Maff-rand Tr. 1681-86; Hanson Tr. 2220.) To the contrary, based on the posters and abstracts, a person of ordinary skill in the art would have concluded that PCR 4099 was under development as a promising racemic drug with several positive qualities and no significant reported negative qualities. (Snyder Tr. 290-91.) Dr. Maffrand, however, also was aware as of 1985 that testing had showed that PCR 4099 had potential negative side effects in humans. (Maffrand Tr. 1651.) Various studies — including acute toxicity, dose-range finding, and chronic oral toxicity studies — conducted in 1983 and 1985 had demonstrated the tendency of PCR 4099 to cause convulsions in animals at particular dose levels. (See Plaintiffs Exhibits (“PL Exs.”) 114, 115, 116, 117; Ro-dricks Tr. 2481-82.) For example, the report of a four week oral toxicity study of PCR 4099 conducted in baboons in 1983 included the observation that “[djeath was preceded by convulsions and was most likely not accidental.” (Pl. Ex. 113.) Furthermore, as a general matter, Sanofi was aware of the possibility that drugs that appeared safe in pre-clinical ,and clinical development might nonetheless show rare but serious side effects after launch — as had recently been the case with ticlopidine. (Maffrand PI Tr. 144; Maffrand Tr. 1651.) With the goal of finding a compound with a better profile than PCR 4099, Dr. Maff-rand decided in November 1985 to have the enantiomers of PCR 4099 separated and tested. (Maffrand Tr. 1651.) A. Predicting the Biological Properties of the Enantiomers of PCR 4099 Whether it was possible for Sanofi to predict the biological properties of the en-antiomers of PCR 4099 — and whether one could predict that the properties of a single enantiomer would be superior to those of its opposite enantiomer — is hotly contested by the parties to this litigation. However, in view of the prior art and the testimony given in this action, the following facts are clear. A person of ordinary skill in the art in the mid-1980s would have known that the enantiomers of a racemate could exhibit different biological activity — including different levels of both therapeutic activity and toxicity. (Snyder Tr. 169.) As of 1984, for example, Ariéns taught that “[o]f-ten, only one isomer is therapeutically active, but this does not mean that the other is really inactive. It may very well contribute to the side-effects.” E.J. Ariéns, “Stereochemistry, a Basis for Sophisticated Nonsense in Pharmocokinetics and Clinical Pharmacology,” Eur. J. Clin. Pharm. 663 (1984) (Def. Ex. 138); see also id. at 664. Similarly, Williams & Lee taught that: Basic pharmacological data on the differences in activity between enantiomers suggest that there are many other drugs for which an increase in therapeutic index might be obtained by using the appropriate enantiomer rather than the ra-cemic drug. Kenneth Williams & Edmund Lee, “Importance of Drug Enantiomers in Clinical Pharmacology,” Drugs 333, 348 (1985) (Def. Ex. 164); see also Goldstein, et al., Principles of Drug Action (2d Ed.) 754 (1974) (Def. Ex. 1305). Dr. Maffrand testified at the preliminary injunction hearing that literature in the late 1970s taught that the therapeutic properties of stereoisomers could “show no obvious difference” or “could have different affinity and different activity for their known receptor.” (Maffrand PI Tr. 170-71.) Where there is variation, the extent of that variation is not predictable and can be weak, moderate, or strong — a view confirmed by experts from both parties and credited by this Court. (Harden Tr. 2295-96.) Dr. Robert Snyder, an expert for Apotex, testified that without separating and testing the enantiomers of a particular racemate, a person of ordinary skill in the art could not know what the degree of difference — if any — between the properties of the enantiomers of a racemic compound would be. (Snyder Tr. 556.) The prior art, in fact, suggested that “weak” stereo-selectivity- — i.e., a difference in activity of ten-fold or less between two stereoisom-ers — was fairly common and that strong stereoselectivity — i.e., a difference in activity of 100-fold or more between stereoi-somers — was less prevalent. (Harden Tr. 2295-99; Hanson Tr. 2215-16; Snyder Tr. 557-58.) In particular, the following prior art reference from Lehmann is relevant: When the biological activities of stereoi-somers (enantiomeric and diastereomeric ...) are compared, it is sometimes found that only one member of each pair is very active, rarely that both members are equally active. Very frequently though, both members exhibit the same type of activity but to a different degree. Several thousand such cases have been recorded .... The differences in activity for the members of any one pair ... vary enormously, viz. from ca. 1 up to nearly 10. P.A. Lehmann, et al., “Stereoselectivity and Affinity in Molecular Pharmacology,” Prog. Drug. Res. 101, 104 (1976) (Def. Ex. 268). The Court finds, furthermore, that a person of ordinary skill in the art would have known that absolute stereoselectivity — -meaning the presence of a particular kind of activity exclusively in one stereoi-somer and its total absence in the other— was uncommon. (Harden Tr. 2299-2300; Lehmann (1976) at 122.) It was also rare, moreover, if both enantiomers exhibited precisely the same levels of activity. (Snyder Tr. 172, 556-58; Lehmann (1976) at 104.) Experts from both parties agreed that even today, no scientific principles afford a basis for predicting to what degree, if any, a pair of stereoisomers will exhibit different levels of therapeutic activity and different levels of toxicity. (Snyder Tr. 563-65; Davies Tr. 2018-19; Hanson Tr. 2214.) As Dr. Davies testified, “Anything is possible .... You can’t predict anything unless you do the experiments.” (Davies Tr. 2019.) The prior art confirms this testimony — as least as it relates to the separation of enantiomers as of the mid-1980s. See Lehmann (Def. Ex. 268) at 104 (“In spite of extended efforts ... it has so far not been possible to explain these ratios; every biological system, sometimes every individual pair, seems to constitute a casus sui generis. ”); W. Soudjin, “Advantages and Disadvantages in the Application of Bioactive Race-mates or Specific Isomers as Drugs,” in Stereochemistry and Biological Activity of Drugs 87, 100 (1983) (Def. Ex. 495) (noting the need for “extensive pharmacological, toxicological and clinical pharmacological research” to determine “whether it is advantageous to use racemates or enantiom-ers in clinical practice”). There are numerous drugs marketed as racemates in which there is no clear difference between the activities of stereoisom-ers, or where even though one enantiomer shows a greater level of activity, the drug is nonetheless marketed as a racemate. Dr. Thomas K. Harden, an expert for Sa-nofi, described fluoxetine, gatifloxacin, amphetamine, and sotalol as drugs in the latter category. (Harden Tr. 2306, 2310-15.) At the time that Sanofi decided to investigate the enantiomers of PCR 4099, Dr. Maffrand was aware that in some cases, the stereoisomers of a drug have greatly different activities (e.g., ibuprofen, thalidomide), but their metabolism in vivo — as opposed to in vitro — leads to ra-cemization or interconversion to a different enantiomeric or achiral compound. (Maff-rand PI Tr. 171-72.) Racemization is a process whereby a compound consisting of a single enantiomer is converted to a one-to-one mixture of that enantiomer and its opposite (i.e., the racemate) by the cleavage and reformation of a chemical bond at the chiral center of the molecule. (Davies Tr. 1956.) The Court finds that a person of ordinary skill in the art would also have been aware of these cases, and would have known that the risk of racemization in the body potentially eliminates any pharmacological advantage derived from preparing the individual enantiomers. (Maffrand Tr. 1656-57.) In addition, Sanofi’s work on thienopyri-dines heightened Dr. Maffrand’s uncertainty regarding the pharmacological potential of PCR 4099’s enantiomers. Dr. Maffrand knew, for example, that thieno-pyridines were active only after administration to live animals, but were not active after in vitro exposure to isolated platelets. (Maffrand Tr. 1684-85; Harden Tr. 2325-26.) On that basis, Dr. Maffrand and his colleagues inferred that PCR 4099 required metabolic conversion — which takes place in vivo — to an active metabolite. (Maffrand Tr. 1684-85; Harden Tr. 2325-26; Snyder Tr. 586.) That active metabolite might not be chiral, and, even if chiral, its mechanism might not show stereoselec-tivity. (Maffrand Tr. 1685-86; Hanson Tr. 2213-14; Harden Tr. 2325-26; Snyder Tr. 585-86.) If so, this would mean that pursuing the enantiomers of PCR 4099 might have no benefit. This, in turn, would suggest that development efforts focus on the racemate instead of the enantiomers. Furthermore, Sanofi’s prior work on the enantiomers of racemic thienopyridine compounds did not suggest that the separated enantiomers would exhibit different types or levels of activity. (Maffrand Tr. 1696-1700). Although Sanofi had previously observed that only one of the enan-tiomers of PCR 1033 exhibited antiplatelet activity, the enantiomers of PCR 3549, by contrast, had exhibited equivalent levels of activity. (Maffrand 1598-99, 1604-05.) Furthermore, the Court credits Dr. Davies’ testimony that a medicinal chemist would not expect PCR 1033 and PCR 4099 to behave in the same manner in the body on the basis of any structural similarity; PCR 4099 has an ester group, which allows for significant hydrogen bonding, whereas PCR 1033 does not. (Davies Tr. 2009-14.) In sum — as experts from both parties agree — it was not possible to predict whether either enantiomer would be more or less therapeutically active — and more or less toxic — than the other. (Snyder Tr. 591-92; Davies Tr. 2014-16; McClelland PI Tr. 505-06.) B. Predicting Whether the Enantiom-ers of PCR U099 Could Be Prepared In addition, Dr. Maffrand and his colleagues could not predict with any reasonable degree of certainty that they would be able to obtain the enantiomers of PCR 4099. Up until the decision to attempt the resolution of PCR 4099, Sanofi had only twice attempted — albeit successfully — to obtain the enantiomers of racemic thieno-pyridines, even though Sanofi had synthesized approximately 1500 thienopyridines, of which roughly 600 were racemates with chiral carbon atoms. (Maffrand Tr. 1618-19.) As of 1987, there were at least ten different methods by which a chemist could try to obtain the individual enan-tiomers of a chiral compound. (Davies Tr. 1921-22.) After choosing a particular technique, a chemist would also have to make numerous decisions: the choice of solvents, of reagents, of concentrations, of temperature, and of time. (Davies Tr. 1939-43.) Sanofi’s expert Dr. Davies testified on direct examination that there is no way to predict which method or methods will succeed if no prior art deals with the compound at hand; practitioners will begin with a method with which they are familiar or a method for which the necessary reagents are readily available. (Davies Tr. 1939.) As noted above, Sanofi had used diaster-eomeric salt formation to separate the en-antiomers of PCR 1033, but had failed when trying to use that method to separate the enantiomers of PCR 3549. As of 1987, the prior art — including standard textbooks of organic chemistry — was replete with discussions of how to resolve enantiomers by diastereomeric salt formation, which was first identified by Louis Pasteur in 1853 and is sometimes referred to as “the classical method.” (Davies Tr. 1925; Snyder Tr. 149-51, 155-56). The Feiser, Karrer, and Jacques references make clear that diastereomeric salt formation was a method of enantiomeric resolution that was well known in the art at the time that Sanofi chemists faced the task of obtaining the enantiomers of PCR 4099. See Louis F. Feiser & Mary Feiser, Advanced Organic Chemistry 85-89 (1961) (Def. Ex. 1322); Paul Karrer, Organic Chemistry 98-99 (1946) (Def. Ex. 162); Jean Jacques, et al., Enantiomers, Race-mates, and Resolutions, 378-88 (1981) (Def. Ex. 492). However, Dr. Davies explained that although that method is “one of the oldest methods,” that does not mean that it is the best method, the easiest method to apply, or “the first one you should try.” (Davies Tr. 1925-26.) As a general matter, obtaining the first crystals of any new compound through separation of diastereomeric salts is difficult. Jacques, for example, teaches that “[n]o infallible recipe exists for overcoming the resistance of a diastereomer to crystallize for the first time.” (Jacques (Def. Ex. 492) at 386.) The Eliel reference confirms this point: Unfortunately, resolution is, in this respect, still very much a matter of trial and error, and even in the papers of experienced investigators one is apt to find, from time to time, a statement that a certain compound resisted resolution by any one of a large number of combinations of resolving agents and solvents that were tried. Ernest L. Eliel, Stereochemistry of Carbon Compounds 50 (1962) (Def. Ex. 1930). See also McClelland Tr. 1117-21; Davies Tr. 1945-48. Although screening techniques would have allowed a person of ordinary skill in the art to improve the odds of a successful separation, diastereomeric salt separation was — and remains even today— a “paradigm of trial and error.” Ton Vries, et al., “The Family Approach to the Resolution Of Racemates,” J. Angewandte Chemie 2349 (1998) (PI. Ex. 852); see also Jacques, Def. Ex. 492 at 380; McClelland Tr. 1137-39. Experts from both parties agreed that although a chemist can search the prior art for useful examples of prior resolutions of similar compounds, the precedents afford absolutely no assurance that the same configuration will succeed with a different compound. (McClelland Tr. 1118; Davies Tr. 1975; see also Eliel (Def. Ex. 1930) at 50.) Furthermore, the literature tends not to report failed resolution attempts. (McClelland Tr. 1135.) In the particular case of PCR 4099, the prior art offered no examples of the resolutions of chiral thienopyridines, and Apotex’s expert Dr. McClelland — in preparing for his testimony — did not find any reference that contained the set of conditions that Sanofi ultimately used to obtain the enantiomers of PCR 4099. (McClelland Tr. 1071.) In fact, Dr. McClelland did not find any relevant instances of the resolution of thieno-pyridines in the prior art. (McClelland Tr. 1127; 1230-32.) In addition, a person of ordinary skill in the art would have recognized the possibility that it would be more difficult to prepare the enantiomers of PCR 4099 than it had been to prepare the enantiomers of PCR 1033 and PCR 3549. First, the meythl ester group attached to the chiral center in PCR 4099 made that molecule more susceptible than PCR 3549 to racem-ization. (Hendrickson Tr. 1498; Maffrand Tr. 1656, 1700; Badorc Tr. 1832-33.) Indeed, a person of ordinary skill in the art would have known that phenylglycine and its derivatives — of which PCR 4099 is one — are susceptible to racemization. (Badorc Tr. 1832; Davies Tr. 1956-60.) Second, because of the presence of the ester substituent on PCR 4099, the hydrogen atom located on the chiral carbon could easily be removed through the action of a base or an acid (which would be required in the course of resolution of the enantiomers). A free hydrogen in solution could then reform a bond to the same carbon atom, but it would bond on either side of that atom with equal likelihood, thereby yielding a racemate. (Davies Tr. 1958-61; 2002-04.) Third, the enantiom-ers of PCR 4099 would be susceptible to racemization in the acidic environment of the stomach by the same mechanism. (Davies Tr. 2004; Hendrickson Tr. 1463.) Accordingly, neither the chemists at Sa-nofi nor a person of ordinary skill in the art could have reasonably expected that the separate enantiomers of PCR 4099 could be obtained at the time that Sanofi was contemplating whether to investigate them and, if obtained, they could not have predicted by what method and configuration. C. Resolution of the Enantiomers of PCR W99 In November 1985, Alain Badorc and Dr. Daniel Fréhel were asked to attempt to obtain the enantiomers of PCR 4099 in order to test whether either enantiomer would make a better drug than the race-mate. (Maffrand Tr. 1650-51.) As the outset, Badorc and Dr. Fréhel decided not to attempt resolution through the formation of diastereomeric salts. The Court credits Badorc’s testimony that that decision was influenced by the fact that PCR 4099 was “only barely basic and, therefore, only barely capable of carrying out a separation using diastereomeric salts.” (Ba-dorc Tr. 1829.) In addition, Badorc believed that there were few commercially available acids that were likely to form diastereomeric salts in sufficient quantities to permit resolution. (Badorc Tr. 1830.) Instead of attempting diastereomeric salt separation, Badorc decided to attempt chemical asymmetric synthesis, the method that he had successfully used to prepare the enantiomers of PCR 3549. This choice was facilitated by the commercial availability of the required starting materials. (Badorc Tr. 1829-31.) Badorc’s lab notebook and a June 12, 1986 report show that Badorc reacted the starting materials to form the enantiomers of an intermediate compound — referred to as “OCBATH.” (PL Ex. 50.) He then used the same “reaction sequence” that he had used in the synthesis of the enantiom-ers of PCR 3549. (Hendrickson 1497-98; Badorc 1836-37; Pl. Ex. 50 at S 03977-78; Pl. Ex. 51 at S 58990-91.) With PCR 4099, however, the sequence failed because the second reaction in the sequence, a condensation, yielded a racemic form of the desired product, not a pure enantiomer. (Badorc Tr. 1836-37; Pl. Ex. 50 at S 03978; PL Ex. 51 at S 58991.) Badorc then tried a different approach. Starting from the racemic OCBATH, he attempted to separate that intermediate’s enantiomers through the formation of diastereomeric salts with tartaric acid in isopropanol. After several crystallizations, Badorc obtained enantiomers with the same melting point and equivalent levels of optical rotation. (Badorc Tr. 1837-38; PI. Ex. 50 at S 0379-80; PI. Ex. 51 at S 58992-93.) Badorc then attempted to convert the enantiomers of OCBATH into those of PCR 4099 using the same cyclization reaction conditions used to prepare the enantiomers of PCR 3549. That product, however, also racemized. (Badorc Tr. 1838; PI. Ex. 50 at S 03980; PI. Ex. 51 at S 58993.) Badorc also tried using diastereomeric salts to resolve the enantiomers of the acid precursor of PCR 4099, a compound designated as PCR 3068. To that end, he attempted resolution with several different chiral bases, but those efforts failed because no crystals formed. (Badorc Tr. 1839-42; PL Ex. 50 at S 03891.) When Badorc attempted to form diastereomeric salts of PCR 3068 with ehirally pure d-camphorsulfonic acid in ethanol, crystals did form, but that effort failed because the crystals contained salts of both enantiom-ers, rather than just of one. (Badorc Tr. 1841-42.) At that point — in January 1986 — Badorc revisited the technique that he had originally been reluctant to use and attempted to resolve the enantiomers of PCR 4099 directly by diastereomeric salt formation. (Badorc Tr. 1842.) To determine what combination of acid, solvent, and concentration would yield crystals, Badorc set up a “screening” experiment with approximately thirty test tubes containing various combinations of a chiral acid (e.g., cam-phorsulfonic, toluoyltartarie, dibenzoyl tartaric, mandelic, and tartaric) with PCR 4099 in various concentrations and in various solvents. (Badorc Tr. 1842-43.) Ba-dorc obtained a solid material in only one of the test tubes — a gummy substance at the bottom of the tube containing a 1:1 ratio of camphorsulfonic acid and PCR 4099 in acetone. (Badorc Tr. 1844.) No crystals were obtained in any test tube. (Badorc Tr. 1844.) In mid-February, Badorc began an additional screening test of several test tubes with different combinations of acid, solvent, and concentration. (Badorc Tr. 1844-45.) Badorc checked the test tubes each day for crystals, the presence of which would indicate the formation of a potentially separable diastereomeric salt. (Badorc Tr. 1846.) Approximately one month after beginning the experiment, Ba-dorc detected crystals in only one tube— the tube that contained ( + ) camphorsul-fonic acid and PCR 4099 in a 4:10 ratio, dissolved in acetone. (Badorc Tr. 1846-47.) As noted above, that same combination using the acid and PCR 4099 in a 1:1 ratio had not yielded crystals. (Badorc Tr. 1843-44.) Because the small test tube reaction had not yielded a sufficient quantity of crystals for further testing, Badorc carried out a larger version of the experiment on March 18, 1986. (Badorc Tr. 1847-48; Pl. Ex. 72 at S 02014.) Risking the loss of the crystals that he had obtained, Badorc used those crystals to seed the larger version, which — after several trials — yielded a significant quantity of crystals of the levoro-tatory enantiomer of PCR 4099 in the cam-phorsulfanate form. (Badorc Tr. 1848-49; Pl. Ex. 72 at S 02014.) Using the same conditions — but substituting chirally pure 10-eamphorsulfonic acid of opposite rotation — Badorc also resolved the dextrorota-tory enantiomer (i.e., clopidogrel). (Ba-dorc Tr. 1850-51; Pl. Ex. 50 at S 03986; Pl. Ex. 74 at S 02019.) Finally, to form the free base of both resolved enantiomers, Badorc reacted the camphorsulfonate salt forms with sodium bicarbonate, a base. (Badorc Tr. 1852.) Badorc verified the optical purity of the free base compounds to confirm that he had resolved the enan-tiomers and that they had not racemized at any stage. (Badorc Tr. 1852-53.) Nothing in the prior art — and nothing in his own experience with PCR 1033 and PCR 3549 — directed Badorc to employ the precise configuration of acid, solvent, and concentration that was ultimately successful with PCR 4099. D. Preparation of Clopidogrel as a Pharmaceutically Acceptable Salt To permit the testing and comparison of the properties of the enantiomers of PCR 4099 to those of the racemate, which had only been tested as a hydrochloride salt, Badorc prepared the hydrochloride salts of each enantiomer. (Badorc Tr. 1852-53.) As described below, pharmacological and toxicological testing showed that clopido-grel hydrochloride exhibited more antipla-telet a