Eli Lilly and Co. v. Barr Laboratories, Inc., 100 F. Supp. 2d 917 (S.D. Ind. 1999)

*918 Jan M. Carroll, Barnes & Thornburg, Indianapolis, IN.

Simon D. Roberts, Shanks & Herbert, Alexandria, VA.

Allen M. Sokal, Finnegan Henderson Farabow Garrett & Dunner LLP, Washington, DC.

Thomas L. Davis, Locke Reynolds Boyd & Weisell, Indianapolis, IN.

Hugh L. Moore, Lord Bissell & Brook, Chicago, IL.

Robert Neuner, Brumbaugh Graves Donohue & Raymond, New York City.

Dennis P. Orr, Mayer Brown & Platt, New York City.

James W. Riley Jr., Riley Bennett & Egloff, Indianapolis, IN.

Maurice B. Stiefel, Bryan Cave, New York City.

Mary Titsworth Chandler, Wooden McLaughlin & Sterner, Indianapolis, IN.

Dan K. Webb Winston & Strawn, Chicago, IL.

BARKER, Chief Judge.

This matter comes before the Court on the parties' cross-motions for partial summary judgment as to Defendants' claim that Plaintiff's patents that are the subject matter of this litigation are invalid for violating the "best mode" requirement of 35 U.S.C. § 112. Defendants contend that Plaintiff failed to disclose the best mode for carrying out the invention with respect to (1) the inventor's preferred method for synthesizing p-trifluoromethylphenol, a starting material for the patented compound fluoxetine hydrochloride and (2) the inventor's preferred recrystallization solvent for purification and recovery of fluoxetine hydrochloride. For the reasons set forth in the Court's discussion below, Defendants' motion for partial summary judgment is denied, Plaintiff's motion for partial summary judgment as to the starting material is granted and Plaintiff's motion for partial summary judgment as to the recrystallization solvent is granted.

Rule 56(c) of the Federal Rules of Civil Procedure provides that a motion for summary judgment shall be granted "if the pleadings, depositions, answers to interrogatories, and admissions on file, together with the affidavits, if any, show that there is no genuine issue as to any material fact, and that the moving party is entitled to a judgment as a matter of law." Fed. R.Civ.P. 56(c). A genuine issue of material fact exists if there is sufficient evidence for a jury to return a verdict in favor of the nonmoving party on the particular issue. See Anderson v. Liberty Lobby, Inc., 477 U.S. 242, 248, 106 S. Ct. 2505, 2510, 91 L. Ed. 2d 202 (1986); Methodist Med. Center v. American Med. Sec., Inc., 38 F.3d 316, 319 (7th Cir. 1994). A disputed fact is *919 material only if it might affect the outcome of the suit in light of the substantive law. See Anderson, 477 U.S. at 248, 106 S. Ct. at 2510; Greenslade v. Chicago Sun-Times, Inc., 112 F.3d 853, 857 (7th Cir.1997). Moreover, "in rendering a decision on a motion for summary judgment, a court must `view the evidence presented through the prism of the substantive evidentiary burden' that would inhere at trial." Monarch Knitting Mach. Corp. v. Sulzer Morat GmbH, 139 F.3d 877, 880-881 (Fed.Cir. 1998) (quoting Anderson, 477 U.S. at 254, 106 S.Ct. at 2513). In considering a summary judgment motion, a court must draw all justifiable inferences in a light most favorable to the opposing party and must resolve any doubt against the moving party. See Anderson, 477 U.S. at 255, 106 S. Ct. at 2513; Spraying Sys. Co. v. Delavan, Inc., 975 F.2d 387, 392 (7th Cir.1992).

While the burden rests squarely on the party moving for summary judgment to show "that there is an absence of evidence to support the nonmoving party's case," Celotex Corp. v. Catrett, 477 U.S. 317, 325, 106 S. Ct. 2548, 2553, 91 L. Ed. 2d 265 (1986), the nonmoving party may not simply rest on the pleadings, but must affirmatively demonstrate by specific factual allegations that a genuine issue of material fact exists for trial. See Celotex, 477 U.S. at 322-23, 106 S.Ct. at 2552-53; Billups v. Methodist Hosp. of Chicago, 922 F.2d 1300, 1302 (7th Cir.1991). Conclusory allegations by a party opposing a motion for summary judgment cannot defeat the motion. See Smith v. Shawnee Library System, 60 F.3d 317, 320 (7th Cir.1995). "The moving party is `entitled to a judgment as a matter of law' [if] the nonmoving party has failed to make a sufficient showing on an essential element of her case with respect to which she has the burden of proof." Celotex, 477 U.S. at 323, 106 S. Ct. at 2553. However, if doubts remain as to the existence of a material fact, those doubts should be resolved in favor of the nonmoving party and summary judgment denied. See Wolf v. City of Fitchburg, 870 F.2d 1327, 1330 (7th Cir.1989).

Barr and Lilly filed cross-motions for summary judgment on Barr's affirmative defense that Lilly's patents are invalid for failure to comply with the "best mode" requirement of 35 U.S.C. § 112.^[1] Barr challenges Lilly's patents on two grounds, asserting that Lilly failed to disclose (1) its method for synthesizing p-trifluoromethylphenol, a starting material for the patented compound fluoxetine hydrochloride and (2) its preferred recrystallization solvent. In order to conduct a thorough analysis of this issue, we will consider the motions together.

On January 10, 1974, Plaintiff, Eli Lilly and Company ("Lilly"), filed a patent application that ultimately resulted in the two patents at issue in this case, U.S. Patent No. 4,314,081 ("the '081 patent") and U.S. Patent No. 4,626,549 ("the '549 patent"). Claim 5 of the '081 patent claims the chemical compound fluoxetine hydrochloride. See Barr Mot. Summ. Judg., Exh. B1. Claim 7 of the '549 patent claims a method of blocking the uptake of serotonin by brain neurons in animals comprising administering to the animal a serotonin-blocking amount of fluoxetine hydrochloride. See Barr Mot. Summ. Judg., Exh. B2. Lilly sued Defendant, Barr Laboratories, Inc. ("Barr"), for infringement of claim 5 of the '081 patent and claim 7 of the '549 patent. Both patents identify p-trifluoromethylphenol as a *920 starting material for making fluoxetine hydrochloride. See Barr Exhs. B1, B2.

Lilly scientists, led by Dr. Bryan Molloy ("Dr. Molloy"), the inventor of the '081 and '549 patents, first made a sample of fluoxetine hydrochloride in February 1973. See Barr Exh. A, Hauser Dep. at 264. In preparing for clinical trials of its new compound, Lilly initially considered purchasing p-trifluoromethylphenol from a chemical supplier, but the management at Lilly refused to approve this purchase request because the material was relatively expensive ($1.00 per gram for one kilogram of material, or $1,000.00) and instead ordered the scientists to manufacture the chemical internally. See Barr Exh. A, Lavagnino Dep. at 63-65. By July 5, 1973, Lilly had created a feasible method of synthesizing p-trifluoromethylphenol. See Barr Exh. B13, Project Team Minutes of 6/20/73 meeting. In April 1977, Dr. Molloy and two of his associates submitted for publication an article detailing their new method, entitled, "An Efficient Synthesis of 4-Hydroxybenzotrifluoride." The authors stated in the article's opening paragraph: "The relatively high cost of 4-hydroxybenzotrifluoride (V) is a limiting factor in its use as a chemical intermediate. Additionally, literature methods for its preparation are cumbersome and not easily adapted to large-scale operations. An improved synthesis of V from readily available, inexpensive 4-chlorobenzotrifluoride (I) in four steps is described." Barr Exh. B14 at 96.

One of Dr. Molloy's co-authors, Edward Lavagnino, stated in his deposition that he considered their method "superior" because the starting material for Lilly's in-house synthesis of p-trifluoromethylphenol "was available in tank car quantities, real cheap chemical, and simple transformations," so that the method could be "scaled up to large scale." Barr Exh. A, Lavagnino Dep. at 126-127. Dr. Molloy admitted in his deposition that the method of synthesis he developed in 1973 was the "best" method he knew for preparing p-trifluoromethylphenol at that time. See Barr Exh. A, Molloy Dep. (Vol.III) at 80-81. However, Dr. Molloy did not disclose this method of synthesis in the '081 or '549 patents. See Barr Exh. A, Molloy Dep. (Vol.V) at 600-601.

After first creating fluoxetine hydrochloride, Lilly scientists worked at finding a proper recrystallization solvent for purposes of purification. The recrystallization process involves using a liquid solvent to dissolve the entire sample at a warm temperature and then cooling down the mixture and separating out the desired product in the form of crystals from the solvent and any impurities. See Barr Exh. A, Lavagnino Dep. at 144. Some solvents work better than others at recrystallizing a given compound. See Barr Exh. A, Corey Dep. at 162; Hauser Dep. at 226; Brown Dep. at 135. In order to determine the most effective solvent for a given compound, trial and error experimentation is necessary. See Barr Exh. A, Corey Dep. at 145-147; Brown Dep. at 118; Molloy Dep. (Vol.III) at 190-191; Lavagnino Dep. at 146. Lilly scientists experimented with two solvents, [* * *],^[2] before selecting [* * *] as their preferred solvent for purifying fluoxetine hydrochloride. See Barr Exh. A, Hauser Dep. at 268-269, 286-287; Lavagnino Dep. at 181; Barr Exh. B7, Lavagnino Progress Report for 8/16/73 to 11/15/73. However, Dr. Molloy did not disclose the use of [* * *] as a recrystallization solvent in the patent application. See Barr Exh. A, Molloy Dep. (Vol.VI) at 746; accord Schmiegel Dep. at 387.

Title 35 U.S.C. § 112 provides in relevant part:

The specification shall contain a written description of the invention, and of the manner and process of making and using it, in such full, clear, concise, and exact terms as to enable any person skilled in the art to which it pertains, or *921 with which it is most nearly connected, to make and use the same, and shall set forth the best mode contemplated by the inventor of carrying out his invention.

35 U.S.C. § 112, para. 1 (emphasis added). The "purpose of this [best mode] requirement is to restrain inventors from applying for patents while at the same time concealing from the public preferred embodiments of their inventions which they have in fact conceived." In re Gay, 50 C.C.P.A. 725, 309 F.2d 769, 772 (1962). Determining whether a patent satisfies the best mode requirement involves a two-part factual analysis:

First, it must be determined whether, at the time the patent application was filed, the inventor had a best mode of practicing the claimed invention. This inquiry is wholly subjective and addresses whether the inventor must disclose any facts in addition to those sufficient for enablement. Second, if the inventor had a best mode of practicing the claimed invention, it must be determined whether the specification adequately disclosed what the inventor contemplated as the best mode so that those having ordinary skill in the art could practice it. The latter question is largely an objective inquiry that depends upon the scope of the claimed invention and the level of skill in the art.

United States Gypsum Co. v. National Gypsum Co., 74 F.3d 1209, 1212 (Fed.Cir. 1996); see also Chemcast Corp. v. Arco Indus. Corp., 913 F.2d 923, 927-928 (Fed. Cir.1990). Because patents are presumed valid by statute, Barr must establish a best mode violation by clear and convincing evidence. See Nobelpharma AB v. Implant Innovations, Inc., 141 F.3d 1059, 1064 (Fed.Cir.1998); United States Gypsum, 74 F.3d at 1212; American Hoist & Derrick Co. v. Sowa & Sons, Inc., 725 F.2d 1350, 1358 (Fed.Cir.1984).

Barr asserts that claim 5 of the '081 patent and claim 7 of the '549 patent are invalid for failure to disclose the best mode known by the inventor for carrying out the claimed invention. Claim 5 of the '081 patent claims the compound fluoxetine hydrochloride and claim 7 of the '549 patent claims a method of blocking the uptake of serotonin in brain neurons in animals by administering a serotonin-blocking amount of fluoxetine hydrochloride. Example 1 of the '081 patent identifies the chemical p-trifluoromethylphenol as one of the starting materials for synthesizing fluoxetine hydrochloride. However, Lilly's patents do not claim p-trifluoromethylphenol or a method for synthesizing p-trifluoromethylphenol. Barr claims that Lilly was required to disclose its in-house method for synthesizing the starting material because the inventor, Dr. Molloy, knew that Lilly's process was a superior method of manufacturing ptrifluoromethylphenol, a necessary intermediate of the invention, and thus should have been included in the patent as part of the best mode of carrying out the invention.

Lilly argues that such disclosure was unnecessary because the best mode requirement is limited to the claims of the patent, and Lilly did not claim the starting material itself, the process for synthesizing fluoxetine hydrochloride from the starting material or the process for synthesizing the starting material in the patents. Thus, Lilly contends that the process of making the starting material is an unclaimed element not necessary for satisfactory performance of the claimed invention and so not a part of the best mode for carrying out the invention subject to disclosure in the patent application. Further, Lilly asserts that Dr. Molloy's best mode for carrying out fluoxetine hydrochloride did not include Lilly's method of synthesizing the starting material because there was no difference in the quality of the p-trifluoromethylphenol produced by Lilly's in-house method and material purchased by Lilly *922 from commercial suppliers and Dr. Molloy had no preference as to which source of material he used.

In assessing whether a patent holder has violated the best mode requirement, the Court must distinguish between an inventor's knowledge of a particular method or material that "substantially improves the operation or effectiveness of his invention" on the one hand and "production specifications" on the other hand. Wahl Instruments, Inc. v. Acvious, Inc., 950 F.2d 1575, 1579-1580 (Fed.Cir.1991). The first requires a best mode disclosure, the second does not. See United States Gypsum, 74 F.3d at 1213 ("Disclosure concerning the best mode of practicing an invention must be distinguished for best mode purposes from disclosure for optimum commercial production. The former is required; the latter is not."); see also Young Dental Manufacturing Co. v. Q3 Special Products, Inc., 112 F.3d 1137, 1144 (Fed. Cir.1997) ("the best mode requirement does not apply to `production details'"); Minco v. Combustion Eng'g, Inc., 95 F.3d 1109, 1116 (Fed.Cir.1996). The terms "production specifications" or "production details" refer to "commercial considerations that do not relate to the quality or nature of the invention." Young, 112 F.3d at 1144.

Any process of manufacture requires the selection of specific steps and materials over others. The best mode does not necessarily cover each of these selections. To so hold would be to turn a patent specification into a detailed production schedule, which is not its function .... [T]he best mode inquiry is not so mechanical. A step or material or source or technique considered "best" in a manufacturing circumstance may have been selected for a non-"best mode" reason, such as the manufacturing equipment was on hand, certain materials were available, prior relationship with supplier was satisfactory, or other reasons having nothing to do with development of the invention.

Wahl, 950 F.2d at 1581. The inventor's preferred manufacturing methods and choices are beyond the scope of best mode disclosure when selected for reasons unrelated to the quality or results of the invention, such as "cost/volume" or mass-production reasons, that do not affect or are not thought to affect how the invention works. See Wahl, 950 F.2d at 1581, 1582; see also R2 Medical Sys., Inc. v. Katecho, Inc., 931 F. Supp. 1397, 1421 n. 2 (N.D.Ill. 1996) ("non-best mode factors" include cost and case of manufacturing, in contrast to effect on invention's performance). Thus, in order for Dr. Molloy's failure to disclose his method of making p-trifluoromethylphenol to constitute a best mode violation, Barr must show that Dr. Molloy subjectively believed at the time he filed his patent application that the in-house method of synthesizing p-trifluoromethylphenol in some way improved the quality or nature of his invention, the chemical compound fluoxetine hydrochloride and its patented uses.

In support of its assertion that Dr. Molloy failed to disclose his best mode, Barr relies upon Dr. Molloy's admission that he considered the method he and the other Lilly scientists developed in-house "superior" to previous methods of synthesizing p-trifluoromethylphenol. However, the clear import of Dr. Molloy's statements is that he considered the Lilly in-house method superior to other methods previously used because it was easier and cheaper, but he makes no reference to any improved quality in the p-trifluoromethylphenol product or the ultimate fluoxetine hydrochloride compound. In other words, Dr. Molloy recognizes improvement and superiority in the method, not the product, and Barr produces no evidence reasonably suggesting that Dr. Molloy perceived any resultant difference in the quality or performance of the invention. Thus, the method in question is beyond the scope of the claims of the patents and irrelevant to the final product and constitutes the type of *923 production detail not requiring best mode disclosure.^[3]

The only additional supporting evidence proffered by Barr is the testimony of Dr. Brown, one of Lilly's expert witnesses, who stated in his deposition that "impurities in intermediate compounds can affect the purity of the final compound." Barr Exh. E, Brown Dep. at 260. While Barr's strategy in introducing such testimony is apparently to highlight quality differences between the alternate sources of the starting material, Barr simply does not establish the necessary evidentiary link. Barr does not produce any evidence from which a reasonable fact finder could conclude that there were fewer impurities in Lilly's synthesized starting material than other available starting material or any other evidence of actual quality differences.

In response to Barr's assertions, Lilly produces the affidavit of Dr. Molloy, who attests, "As of January 10, 1974, I had no preference for the source of the p-trifluoromethylphenol starting material used to make fluoxetine hydrochloride, because I regarded the invention as fluoxetine hydrochloride and I did not believe the source of the starting material made any difference to the performance of the fluoxetine hydrochloride. In general, I preferred purchasing commercially available starting materials to making them." Lilly Opp. to Barr Mot. Summ. Judg., Exh. 12, Molloy Decl. ¶ 2.^[4] Barr has produced no evidence casting doubt upon the truth of Dr. Molloy's statement that he had no preference as to the source of the starting material. Thus, for all of the reasons discussed above, we find that Barr's proffer of evidence falls far short of the clear and convincing evidence required to support its best mode challenge for purposes of both Barr's own motion for summary judgment as well as its opposition to Lilly's motion for summary judgment.

Because Lilly has successfully demonstrated that Barr cannot show by clear and convincing evidence that Dr. Molloy preferred the in-house method of synthesizing ptrifluoromethylphenol on the basis of quality considerations, we find that disclosure of this method was not necessary to set forth the best mode of carrying out the invention. Our conclusion is supported by the Federal Circuit's and other federal courts' decisions in the following cases: Nobelpharma, 141 F.3d at 1065 (inventor violated best mode requirement where he failed to disclose details "important" to obtaining a "suitable micropitted implant" and certain machining parameters that were "critical to the production of a functioning implant"); United States Gypsum, 74 F.3d at 1211-13 (best mode violation where inventor failed to disclose "best" material for use in invention that "significantly improved the physical properties of the joint compound," rather than "matter of commercial expediency"); Wahl, 950 F.2d at 1581, 1583 (no best mode violation where embedment molding did not affect how invention worked and inventor made own inserts rather than purchase them on *924 the basis of cost considerations); Northern Telecom Ltd. v. Samsung Electronics Co., 1998 WL 602327, * 5 (N.D.Cal. Sept.4, 1998) (best mode violation where inventor failed to disclose "preferred material for achieving the most beneficial results from the plasma-etching process," such as finer pattern definition and spearless etching); Advanced Semiconductor Materials America, Inc. v. Applied Materials, Inc., 922 F. Supp. 1439, 1447 (N.D.Cal.1996) (inventors violated best mode when they failed to disclose IR (lamp/radiant) heating although it was the only mode with which they were able to obtain temperature uniformity "essential to process"); E.I. DuPont De Nemours and Co. v. Monsanto Co., 903 F. Supp. 680, 717-718 (D.Del.1995) (no best mode violation where inventor claimed compound, not process of manufacturing component substance, and there was no evidence that inventor believed preferred substance produced better results but rather testified "catdye nylon made from caprolactam is chemically indistinguishable from catdye nylon made from polycaprolactam"); Acme Resin Corp. v. Ashland Oil Inc., 20 U.S.P.Q.2d 1305, 1310-1311 (S.D.Ohio 1991) (best mode violation where inventor failed to disclose best compound additive and binder components to produce "optimal results" in bench life extension); Cosden Oil & Chemical Co. v. American Hoechst Corp., 543 F. Supp. 522, 533 (D.Del.1982) (no best mode violation where invention was a compound, not a process, and no evidence that inventor's preferred process produced superior composition to mode disclosed). But see Clayton v. Akiba, 214 U.S.P.Q. 374, 380-381, 1982 WL 50444 (P.T.O. Bd. App.1982) (best mode violation where inventor failed to disclose method of preparing novel intermediate compound for novel patented compound, though inventor's preference was based in part on cost).^[5] Accordingly, as to Barr's claim that Lilly violated the best mode requirement of 35 U.S.C. § 112 by failing to disclose its method for synthesizing p-trifluoromethylphenol, we deny Barr's motion for summary judgment and grant Lilly's motion for summary judgment.

As a collateral matter, the parties also dispute whether p-trifluoromethylphenol was commercially available at the time of the patent application in January 1974. Lilly asserts commercial availability as a defense to nondisclosure. As Lilly's failure to disclose a method for synthesizing a material not available to anyone in the market may, in fact, constitute a best mode violation, we will consider this issue as well.^[6] Lilly produces several pieces of evidence showing that p-trifluoromethylphenol was available on the market at the relevant time, chief among them documentation that Lilly itself used commercial p-trifluoromethylphenol in formulating the fluoxetine hydrochloride compound before Dr. Molloy filed the patent application. See Lilly Mot. Summ. Judg., Exh. 9, Notebook 687-A57 at 31, 35, 37, 41-42; Lilly Exh. 7, Cannon Dep. at 93-94, 97-98. In addition, as discussed above, the evidence shows that Lilly synthesized the p-trifluoromethylphenol in-house because the commercial product was too expensive for a large-scale clinical trial. See Barr Exh. A, Lavagnino Dep. at 63-65. Lilly also produces the 1973 catalog of Marshallton Research Laboratories Incorporated, which offered p-trifluoromethylphenol for sale in various quantities. See Lilly Exh. 4; Lilly Exh. 6, Foster Dep. at 20-21, 38; Lilly Exh. 7, Cannon Dep. at 87-88. Marshallton's customers in 1973 included Lilly, Wyeth Laboratories, Fluka AG, SmithKline, E.I. duPont de Nemours, Motorola, Parke, Davis & Co., Pfizer, Inc., Dow *925 Chemical Co., and dozens of university and research institutes. See Lilly Exh. 8; Lilly Exh. 6, Foster Dep. at 43-45. Lilly also attests that there were several methods of synthesizing p-trifluoromethylphenol available in the literature at that time, some of which were tried and rejected by Lilly's scientists in formulating their own method. Barr does not meaningfully challenge any of this evidence.

Barr's primary challenge to the commercial availability of p-trifluoromethylphenol before January 1974 comes from two patents, one issued to Teva Pharmaceuticals, Ltd. in July 1993 and another issued to Lilly in 1985. The Teva patent, U.S. Patent No. 5,225,585, entitled "Production of Fluoxetine and New Intermediates," states, "The substituted phenol (formula V) [p-trifluoromethylphenol] which is an essential building block, is not available commercially." Barr Exh. D1 at col. 2, lines 34-36. The Lilly patent, Spanish Patent No. 535,660, states in relevant part, "Since p-trifluoromethylphenol is not available in the market and it is relatively difficult to synthesize, the manufacturing cost ... is correspondingly high." Barr Exh. D2 at 3. Lilly challenges these statements on the basis that the Teva patent apparently depends upon an earlier Israeli patent and the Lilly patent is a Spanish patent and thus Lilly postulates that p-trifluoromethylphenol may not have been available in Israel or Spain, while available in the United States, the relevant market to this case. However, we find such challenge unnecessary, because even if these patents are reliable sources for establishing commercial availability, these patents speak only to the market in 1984-1985 and 1991-1993, undeniably ten to twenty years after the patent application at issue, January 1974. Thus, Barr's evidence that p-trifluoromethylphenol was not on the market at the time the Teva and Lilly patents issued, without more, is insufficient to create a genuine issue of material fact as to the commercial availability of p-trifluoromethylphenol in the United States in or around January 1974. Further, it is undisputed that Dr. Molloy made his method public in 1977 and so commercial availability would be unnecessary for carrying out Lilly's invention after that point and the market supply of p-trifluoromethylphenol in 1984-1993 appears completely irrelevant to the issue at hand, i.e. whether Dr. Molloy had a duty to disclose his "superior" method in his 1974 patent application.

Barr also attempts to argue through the testimony of its expert witness, Dr. Cannon, that p-trifluoromethylphenol was not commercially available before January 1974. See Barr Resp. Br. to Lilly Mot. Summ. Judg. at 18. However, Dr. Cannon states merely that he doubts the substance was "widely known" at that time and believes it must have been more properly viewed as a "rare chemical." Barr Exh. D3, Cannon Supp. Expert Report at 5. In addition, Dr. Cannon admitted at deposition that he had no personal experience with that compound and concedes commercial availability before January 1974 because it was available then in a catalog from Aldrich, another chemical supplier. See Lilly Exh. 14, Cannon Dep. at 223-225. Thus, Dr. Cannon's testimony does not create a genuine issue of material fact as to commercial availability; if anything, this evidence supports Lilly's contention. Barr also claims that Dr. Molloy's testimony that p-trifluoromethylphenol "was not a readily available intermediate" provides support for their challenge, but this statement was made in response to an inquiry as to what Dr. Molloy meant when he described "difficultly synthesizable phenols" (see Lilly Exh. 7, Molloy Dep. (Vol. III) at 73-74) and does not speak meaningfully to the question of whether p-trifluoromethylphenol was available in the market in January 1974.

Barr also tries to undermine the evidence of the Marshallton catalog proffered by Lilly through the testimony of Marshallton's president, Dr. Foster, but Barr was unsuccessful in casting any real doubt on the fact that Marshallton made such *926 material available for sale and sold it at or before the time the patent application was filed. See Barr Exh. C, Foster Dep. at 38, 43. Barr does not dispute that there were methods of synthesis published and available to the public in the scientific literature of the time, although claiming that many of the methods revealed in the literature were not easily duplicated or contained "nasty reactants." See Barr Reply Br. at 13-14. In addition, Barr does not appear to dispute that Lilly itself purchased quantities of commercial p-trifluoromethylphenol for purposes of creating fluoxetine hydrochloride initially, providing further evidence of commercial availability. In short, Barr does nothing to create a genuine issue of material fact as to whether p-trifluoromethylphenol was commercially available before and around January 1974, further supporting our conclusion that Dr. Molloy's failure to disclose his method of synthesizing p-trifluoromethylphenol in his patent application did not violate the best mode requirement of 35 U.S.C. § 112.

Barr next contends that Lilly's patents are invalid for violating the best mode requirement because Dr. Molloy did not disclose his preferred recrystallization solvent for purifying fluoxetine hydrochloride, [* * *]. Barr argues that the patent claims a pharmaceutically-acceptable form of fluoxetine hydrochloride and thus the purification process is a claimed element. In the alternative, Barr asserts that choosing an effective recrystallization solvent is a necessary and important process requiring scientific experimentation and so should have been disclosed in order to set forth the best mode of carrying out the invention. Lilly argues that the patent does not claim either a purification process or a particular level of purity in the fluoxetine hydrochloride compound requiring disclosure of Lilly's preferred recrystallization method or solvent. Lilly also asserts that choosing a recrystallization solvent is a routine detail that is not required to be disclosed as part of the inventor's best mode.

First we will address Barr's contention that the patents claim a pharmaceutically-pure form of fluoxetine hydrochloride. Claim construction is a matter of law exclusively for the Court to determine. See Markman v. Westview Instruments, Inc., 517 U.S. 370, 384, 116 S. Ct. 1384, 1393, 134 L. Ed. 2d 577 (1996). "To ascertain the meaning of claims, we consider three sources: The claims, the specification, and the prosecution history." Markman v. Westview Instruments, Inc., 52 F.3d 967, 979 (Fed.Cir.1995) (citing Unique Concepts, Inc. v. Brown, 939 F.2d 1558, 1561 (Fed.Cir.1991)).

After examining the claim language in question, we agree with Lilly's assertion that the term "pharmaceutically-acceptable" as used in the patents does not apply as a restriction on the purity of the fluoxetine hydrochloride compound itself, but rather defines the acids that may be used to form the salts of the claimed compounds: pharmaceutically-acceptable or non-toxic acids which form pharmaceutically-acceptable or non-toxic salts. In other words, the claim language refers to toxicity of the components used to create the compound and the toxicity of the resultant compound, rather than a particular level of purity in the final product, as may be achieved through a process like recrystallization.

For example, claim 5 of the '081 patent, which claims fluoxetine hydrochloride, depends upon claim 4, which claims fluoxetine "and pharmaceutically-acceptable acid addition salts thereof formed with non-toxic acids." Claim 4 depends in turn upon claim 1, which claims a family of compounds that includes fluoxetine "and acid addition salts formed with pharmaceutically-acceptable acids." The '549 patent claims use the same language as the '081 patent. In addition, the description portions *927 of the patents, which are identical, set forth:

This invention provides [amine compounds] of the formula ... and acid addition salts thereof formed with pharmaceutically-acceptable acids.

['081 patent at col. 1, lines 22-23, 43-44]

* * * *

Also included within the scope of this invention are the pharmaceutically-acceptable salts of the amine bases represented by the above formula formed with non-toxic acids. These acid addition salts include salts derived from inorganic acids such as: hydrochloric acid, nitric acid, phosphoric acid ... and the like, as well as salts of non-toxic organic acids including aliphatic mono and dicarboxylates ... etc. Such pharmaceutically-acceptable salts thus include: sulfate, pyrosulfate ... and the like salts.

['081 patent at col. 2, lines 31-58]

* * * *

... [I]t is preferred to use an acid addition salt of the compound formed with a pharmaceutically-acceptable non-toxic acid.

['081 patent at col. 16, lines 48-50]

Lilly Opp. Exh. 8. In addition, Lilly provides evidence that the common meaning of the term "pharmaceutically-acceptable addition salts formed with non-toxic acids" in the patent context merely refers to limiting claims to salts that are non-toxic. See Lilly Opp. Exhs. 3, 4, 5.^[7] As Lilly argues, "The claim language merely excludes salts of toxic acids such as hydrogen cyanide. It has nothing to do with the form or purity of the claimed compounds." Lilly Opp. Br. to Barr Mot. Summ. Judg. at 14. Barr's argument to the contrary is clearly unavailing.

Next we must consider whether Dr. Molloy's preferred method of recrystallization is necessary to the practice of his best mode of carrying out his invention so that its disclosure in the patent application was required. Barr argues that the claims make "crystal clear that the compounds are intended for a pharmaceutical purpose, which means that compound purity is critical." Barr Rep. Br., Mot. Summ. Judg. at 12. In addition, Barr contends that Dr. Molloy had a duty to disclose his recrystallization process as part of his preferred method of making fluoxetine hydrochloride. However, Lilly asserts that the patents do not claim a method of making fluoxetine hydrochloride and so purification is an unclaimed element that must be disclosed only if it is necessary to carry out the claimed invention. Lilly contends that recrystallization is not necessary to make fluoxetine hydrochloride but only an optional step to further purify the compound, if desired, after it has been made. Further, Lilly argues that recrystallization is a "routine production detail" not subject to best mode disclosure, stating that the patent teaches that the salts disclosed, including fluoxetine hydrochloride, can be purified by recrystallization and that "chemists of ordinary skill in the art knew how to select suitable solvents to recrystallize such salts in 1974 when the patent application was filed by simple, straightforward, standard experiments with common solvents." Lilly Opp. Br. to Barr Mot. Summ. Judg. at 17.

*928 We have already determined that the claims of the patent do not include a particular level of purity of fluoxetine hydrochloride and thus we conclude that the purification process is not a claimed element subject to disclosure. However, we acknowledge the validity of Barr's assertion that chemists generally prefer to use purer compounds, particularly when the compound is used for pharmaceutical products for humans. There is sufficient evidence to support a finding that Dr. Molloy preferred a purified form of fluoxetine hydrochloride as a finished product, especially considering the fact that Lilly's long-range plan was to use the invention for pharmaceutical purposes. See Barr Exh. A, Molloy Dep. (Vol.III) at 141 ("it is better to have pure compounds than impure compounds"), 143 ("methods of synthesis that yield the purest compounds are preferable to methods of synthesis that yield less pure compounds, all other things being equal"); Barr Exh. A, Brown Dep. at 55-56, 60 (impurities can be toxic to humans and may interfere with the intended use of the compound). Thus, we find that there is, at least, a genuine issue of material fact as to whether Dr. Molloy's best mode for carrying out his invention includes purification of the claimed compound, to the extent preferred for pharmaceutical use, even though purification is an unclaimed element, because it is necessary to the satisfactory performance of the invention. See Applied Med. Resources Corp. v. United States Surgical Corp., 147 F.3d 1374, 1377 (Fed.Cir.1998) ("An applicant is obliged to disclose nonclaimed elements necessary to the operation or carrying out of the invention to which the patent is directed."); see also Chemcast Corp., 913 F.2d at 928 ("where the invention relates only to a part of, or one aspect of, a device, an applicant is not required to disclose a nonclaimed element necessary to the operation of the overall device, but not necessary to the operation of the invention to which the patent is disclosed").

Our inquiry does not stop here, though. Having established that there is a genuine dispute as to whether Dr. Molloy had a best mode that included purification of fluoxetine hydrochloride, the second level of inquiry is whether this best mode is adequately disclosed in the patent application, for if it is adequately disclosed, there is no best mode violation. Lilly argues that the recrystallization process is a routine detail that would have been known to a chemist of ordinary skill and thus disclosure was not required. Similar to production details, "routine details" are not subject to best mode disclosure:

Routine details are details that are apparent to one of ordinary skill in the art. They are appropriately discussed separately from production details because routine details do relate to the quality or nature of the invention. Nevertheless, they need not be disclosed because, by definition, their disclosure is not required under the second inquiry of the best mode determination. In other words, to satisfy the best mode inquiry of the best mode test, an inventor need only disclose information about the best mode that would not have been apparent to one of ordinary skill in the art. Because routine details are apparent to one of ordinary skill, they need not be disclosed.

Young, 112 F.3d at 1144; see also Great Northern Corp. v. Henry Molded Prods., 94 F.3d 1569, 1572 (Fed.Cir.1996); Wahl, 950 F.2d at 1581. "What is required is an adequate disclosure of the best mode, not a guarantee that every aspect of the specification be precisely and universally reproducible." Amgen, Inc. v. Chugai Pharm. Co., 927 F.2d 1200, 1212 (Fed.Cir.1991).

Lilly cites in support of its argument Warner-Jenkinson v. Allied Chemical Corp., 477 F. Supp. 371 (S.D.N.Y.1979), in which the district court found that the inventor did not need to disclose an alcohol-wash purification step for the patented dye to satisfy the best mode requirement. The court found conclusive testimony that the alcohol-wash procedure was "something *929 that almost any organic chemist should know" and that "a chemist with ordinary skill in the manufacture of food colors could pick up the Red 40 patents, read the disclosed method of preparation, and have no problem in preparing the product with the degree of purity required for FDA certification." Warner-Jenkinson, 477 F. Supp. at 395. Lilly argues that purifying fluoxetine hydrochloride in [* * *] is analogous to the routine purifying alcohol-wash described in Warner-Jenkinson.

Lilly contends that other forms of purification besides recrystallization are known, that many other recrystallization solvents to purify fluoxetine hydrochloride are known and have been used and that the trial-and-error process of selecting an effective solvent for recrystallization was a routine matter within the knowledge of a chemist of ordinary skill in the art. See Lilly Exh. 2, Corey Report at para. 11-14; Lilly Mot. Summ. Judg., Table 1; Lilly Exh. 12, Teva Patent at col. 6, lines 9-17; Lilly Exh. 10, DDX 245 at KT 363 12, 16; Lilly Exh. 5, Geneva Resp. to Lilly Req. for Admissions, response to request 4 (recrystallization with water); Lilly Exh. 6, Geneva letter dated October 26, 1998 (recrystallization with water and methyl ethyl ketone). Lilly also notes that the '081 patent teaches recrystallization as a purification method, as shown in the working examples discussing recrystallization from various solvents. See Lilly Exh. 9, col.6, lines 7-12, 19-22, 25-27; col. 7, lines 36-38; col. 8, lines 1-3, 12-14; col. 9, lines 63-65. Accordingly, Lilly argues, a skilled chemist reading the patent would have understood that fluoxetine hydrochloride could be purified using recrystallization. In addition, two solvents disclosed in the '081 patent, ethyl acetate and [* * *], have since proven to be suitable recrystallization solvents for fluoxetine hydrochloride. See Lilly Exh. 9, col. 6, line 27; col. 7. lines 37-38; col. 8, lines 3, 13-14; Lilly Exh. 10 at KT 363 12, 16. Thus, Lilly contends, a skilled chemist testing only the disclosed solvents would have succeeded in recrystallizing fluoxetine hydrochloride.

Lilly also avers that Barr's expert, Dr. Cannon, listed nine "traditional recrystallization solvent systems that the average skilled chemist would try first" to recrystallize hydrochloride salts of amines such as fluoxetine hydrochloride, three of which Lilly claims have already been used to recrystallize fluoxetine hydrochloride successfully, while only one has been shown to be a poor recrystallization solvent. See Lilly Exh. 16, Cannon Expert Report at 6; Lilly Mot. Summ. Judg., Table 3. Dr. Cannon admitted that [* * *], in addition to the nine solvents he listed, was also known as a recrystallization solvent for amine salts. See Lilly Exh. 16, Cannon Expert Report at 6; Lilly Exh. 4, Cannon Dep. at 204. As further support of its argument that purification methods are routine, Lilly asserts that publications and patents often do not specify the solvents used for recrystallization. See Lilly Mot. Summ. Judg., Table 2; Lilly Exh. 3, Cannon Supp. Report at 4. Thus, Lilly argues, the evidence clearly shows that an ordinary skilled chemist would have been able to select an effective recrystallization solvent for the invention, even [* * *].

The foremost question before the Court, then, is whether the selection of a suitable recrystallization solvent is a routine detail within the knowledge of a chemist of ordinary skill. One of Lilly's witnesses contends that "it's a trivial thing for organic chemists to do recrystallization ... it's sophomore organic chemistry.... [I]t's a simple process that you learn, one of the first things that you learn in organic chemistry." Lilly Opp. Exh. 6, Lavagnino Dep. at 144-145. Mr. Lavagnino characterized his own selection process for purifying fluoxetine hydrochloride as merely "fooling around with recrystallization solvents." Lilly Opp. Exh. 6, Lavagnino Dep. at 143-144. In fact, Mr. Lavagnino asserts, "As I testified at pages 144 to 145, experimenting to determine an acceptable recrystallization solvent is elementary and routine and is one of the first things that one *930 learns in organic chemistry. I spent no more than a half hour testing at most four or five solvents and easily determined that [* * *] was acceptable." Lilly Opp. Exh. 15, Lavagnino Decl. ¶ 2. Another Lilly expert, Dr. Elias J. Corey, testified at deposition that:

fluoxetine hydrochloride and other amine hydrochlorides of that sort, rather than mediumsize molecules amine hydrochlorides, they are generally quite easy to purify by recrystallization.... Let's take the case of fluoxetine, the free amine. Greg Reichard, ... a graduate student in my group, synthesized R & S fluoxetine. We converted the free amine fluoxetine into the crystalline hydrochloride simply by treating it with an ethereal solution of hydrogen chloride, and the crystals came down immediately. So, it's a nicely crystalline compound. In that case, it's easy to purify and easy to get crystalline material.

Lilly Exh. 7, Corey Dep. at 117-118. Dr. Corey stated that while the selection process requires trial-and-error experimentation, "it's very straightforward ... [I]t's elementary and it's one of the very first things that a synthetic chemist masters. ... It's one of the quickest, easiest things you can do in the laboratory.... For a person, as I have defined it, who is of ordinary skill in the pharmaceutical industry, it is literally a snap.... It's a slam dunk." Lilly Exh. 7, Corey Dep. at 147-148. Another of Lilly's witnesses opines similarly that "recrystallization of fluoxetine hydrochloride is a routine matter." Lilly Exh. 1, Cordes Dep. at 63. Further, Dr. Molloy states in his affidavit:

I recall generally that selecting a recrystallization solvent for fluoxetine hydrochloride was not a problem. If it had been, I would have remembered it. Lilly's laboratory notebook records demonstrate that the total time spent selecting a recrystallization solvent was on the order of hours, not months.

Lilly Opp. Exh. 12, Molloy Decl. ¶ 8.

In disagreement with Lilly's experts, Barr's expert, Dr. Cannon, opines that this selection process cannot be considered routine because:

selecting recrystallization solvents often requires time-consuming trial and error experimentation. Picking the right solvents does not, as Lilly's experts suggest, require the mere invoking and application of simple rules about polarity.

* * * *

The selection of a suitable solvent would not necessarily be straightforward and it may involve some real difficulty. This is because the goal of identifying a recrystallization solvent usually involves considerably more than just identifying a solvent in which the compound is soluble in the hot solvent and insoluble in the cold. This says nothing about the ability of this solvent to effect purification of the subject compound.

* * * *

Even when a chemist determines that more than one solvent can be useful for recrystallizing a particular compound, scientists typically prefer a particular solvent based on a variety of considerations, including, for instance, yield, crystal purity, toxicity of the solvent, the residual amounts of the solvent in the compound crystals, and the cost of the solvent.

Lilly Exh. 3, Cannon Supp. Report at 1, 2, 3.

It is important to note that Dr. Cannon does not dispute that the methods for selecting a suitable recrystallization solvent are well-known to those of ordinary skill in the art. Rather, he objects to Lilly's experts' characterization of the selection process as "routine" and "straightforward" because the process can be time-consuming and laborious. Dr. Cannon opines at length on the many variables that can complicate the process of finding a suitable recrystallization solvent. However, Dr. *931 Cannon provides very little information directly speaking to actual difficulties encountered in recrystallizing fluoxetine hydrochloride that would elevate the process from a routine procedure within the average chemist's general skill. At best, Dr. Cannon challenges the purportedly "routine" nature of the process by asserting that "Lilly engaged in a great deal of experimentation before finally settling on [* * *] as the solvent of choice," asserting that the fact that Lilly devoted research time to this selection process indicates that the Lilly scientists did not regard the process as "routine" and "trivial." Barr Exh. D4, Cannon Supp. Report at 4. However, Lilly produced evidence from witnesses with direct knowledge of Lilly's in-house procedures in question (Dr. Cannon undisputedly was not involved) who testified that the selection of [* * *] as a recrystallization solvent at Lilly was not time-consuming and difficult. See Lilly Opp. Exh. 6, Lavagnino Dep. at 143-144; Lilly Opp. Exh. 15, Lavagnino Decl. ¶ 2; Lilly Opp. Exh. 12, Molloy Decl. ¶ 8.

The evidence before the Court, virtually undisputed, shows that the process of purifying chemical compounds in general and recrystallizing chemical compounds with the aid of liquid solvents in particular was well-known by those of ordinary skill in chemistry and further, that [* * *] was one of a dozen or so common liquid solvents known in the art as suitable recrystallization solvents for hydrochloride salts (like fluoxetine hydrochloride). There is some dispute between the parties as to how long Lilly worked at selecting an effective recrystallization solvent and how long a chemist of ordinary skill in the art reading Lilly's patents would have to work to select a proper solvent, but these questions do not meaningfully detract from the fact that a skilled chemist could have purified the compound for purposes of practicing the invention without Lilly's disclosure of [* * *] as its preferred recrystallization solvent. While we do not announce a categorical rule that the amount of time and labor required to practice an undisclosed procedure that is nonetheless within the art is irrelevant, we find that the particular facts of this case lead to the inescapable conclusion that the selection of a satisfactory purification method or a suitable recrystallization solvent for fluoxetine hydrochloride was a routine detail that would have been known by a chemist of ordinary skill in the art. Cf. Amgen, 927 F.2d at 1211 (no best mode violation where "isolation of the preferred strain was a `routine limited dilution cloning procedure[]' well know in the art"); Hybritech Inc. v. Monoclonal Antibodies, Inc., 802 F.2d 1367, 1384-1385 (Fed.Cir.1986) (evidence that screening methods known in the art were "labor-intensive and time-consuming" insufficient proof of concealment of a best mode); International Telephone and Telegraph Corp. v. Raychem Corp., 188 U.S.P.Q. 214, 221 (D.Mass.1975) (no best mode disclosure required where "the selection of a material as an extrusion aid would have been within the ordinary skill of the art at that time").

However, this finding does not reach the primary issue raised by Barr: whether Dr. Molloy's best mode extends to the specific purification method and recrystallization solvent used in Lilly's production of fluoxetine hydrochloride, [* * *], such that it should have been disclosed in the patent application. After considering the evidence before the Court, we find that it does not. Lilly has shown that there are many methods of purification, recrystallization among them, and many different recrystallization solvents available for purification purposes, including [* * *]. In addition, we found that selecting purification methods and suitable recrystallization solvents are routine matters for a chemist of ordinary skill in the art. Further, Lilly has produced evidence that the particular purification method used does not meaningfully impact the final fluoxetine hydrochloride product. Lilly's expert, Dr. Corey, sets forth in his expert report:

*932 After purification of fluoxetine hydrochloride by recrystallization, the product performance would be the same regardless of the exact procedure used. By carrying out a few standard test experiments, a person of ordinary skill could quickly ascertain a suitable solvent and suitable conditions of temperature and concentration to effect purification of fluoxetine hydrochloride. Indeed, that person would likely find that any of several alternative procedures could be employed to produce pure fluoxetine hydrochloride. ... Although [* * *] is, and would have been considered in 1974, a perfectly reasonable solvent for recrystallization and purification of fluoxetine hydrochloride, there are a considerable number of solvents and solvent mixtures which would have been (and still would be) perfectly satisfactory. It is simply untrue that [* * *], a commonly used solvent, is the only solvent capable of producing fluoxetine hydrochloride by recrystallization.

* * * *

The choice of a recrystallization solvent for the purification of fluoxetine hydrochloride does not have a meaningful effect on the pharmacological properties of the pure crystallized compound, since polymorphic forms (i.e. different crystal forms) have never been reported despite an enormous body of work on this compound.

* * * *

Theoretically, the ideal purification is one in which the purified material is obtained in 100% purity with no material loss whatsoever (that is, 100% yield or efficiency). In practice, this ideal is usually unattainable and a product is recrystallized only the minimum number of times required in order to generate product of satisfactory purity. Often, 95-98% purity is acceptable....

Lilly Exh. 2, Corey Report at para. 6, 19, 20 (emphasis added). Barr's response to this expert testimony is the supplemental report of its own expert, Dr. Cannon. This report announces Dr. Cannon's disagreement with Lilly's experts as to the "routine" nature of recrystallization and expounds at length about the various complexities of choosing appropriate recrystallization solvents for achieving satisfactory results, as discussed above. Unfortunately for Barr, Dr. Cannon's opinion does not contradict or call into doubt Dr. Corey's assertion that using different purification methods does not meaningfully affect the properties or performance of the compound (see Lilly Exh. 3, Cannon Supp. Report), a factor we consider critical to determining whether the use of [* * *] is part of the inventor's best mode as a method that is thought to "substantially improve the operation or effectiveness" of the patented compound or rather functions as a non-substantive production detail. Wahl, 950 F.2d at 1579-1580; see also Cosden Oil, 543 F. Supp. at 533 (no best mode violation where no evidence that special process conditions known within the art "were required to obtain the advantages of the invention").

In addition to failing to show that the use of [* * *] is the type of production step that must be included as part of the inventor's best mode because of its importance in achieving the desired results in the invention and thus indirectly prove that the inventor contemplated a best mode, Barr similarly fails to show that Dr. Molloy subjectively had a best mode that included recrystallizing fluoxetine hydrochloride from [* * *]. While Barr has produced sufficient evidence showing that Lilly scientists formed a preference for recrystallizing the patented compound with [* * *] and rejected water and benzene through their experimentation with these solvents, Barr does not show or even raise a genuine issue of material fact that the inventor shared this preference as part of his best mode. Barr contends in its motion brief that "Molloy determined that the use of a specific recrystallization *933 (purification) solvent named `[* * *]' gave the best results" (Barr Summ. Judg. Mot. Br. at 3), but supports this assertion only by the undisputed fact that Dr. Molloy was aware of the Lilly's scientists' experimentation with the recrystallization process, rather than direct evidence of Dr. Molloy's preference. Barr highlights the fact that Dr. Molloy was the inventor as well as team chemist for Lilly's project team for fluoxetine hydrochloride. See Barr Summ. Judg. Mot. Br. at 10-11, citing Barr Exh. A, Molloy Dep. (Vol.I) at 114. Dr. Molloy admits that he was aware that Lilly scientists had used [* * *] (see Barr Exh. A, Molloy Dep. (Vol.VI) at 746) and that the scientists likely made him aware of their preference for this solvent, though having no specific recollection of this fact (see Barr Exh. A, Molloy Dep. (Vol.V) at 446, 448-449).

However, this evidence, while showing that Dr. Molloy knew of the work and conclusions of the scientists working on purification, does not speak directly to whether Dr. Molloy subjectively preferred [* * *] so as to constitute part of his best mode. In fact, Dr. Molloy directly contradicts Barr's assertion in his affidavit, stating:

I disagree with the statement on page 3 of Barr's Memorandum in Support of Its Motion for Summary Judgment on the '081 and '549 Patents Based Upon a Best Mode Violation that I determined that using [* * *] to recrystallize fluoxetine hydrochloride gave the best results. The experimentation using [* * *] as the recrystallization solvent for fluoxetine hydrochloride was done by others, not me. Before January 10, 1974, Lilly used a variety of solvents on a variety of lots of fluoxetine hydrochloride prepared by different synthetic routes by several different Lilly scientists. Before January 10, 1974, no Lilly scientist performed a systematic, objective comparison of recrystallization solvents for fluoxetine hydrochloride. Without such a comparison, it could not have been concluded that [* * *] was the best recrystallization solvent. I do not recall having a preferred solvent for recrystallizing fluoxetine hydrochloride as of January 10, 1974.... Although I have no specific recollection, I doubt that I had a preference for any particular recrystallization solvent for fluoxetine hydrochloride as of January 10, 1974.

Lilly Opp. Exh. 12, Molloy Decl. ¶¶ 5-7.^[8] Besides serving as direct evidence that Dr. Molloy's subjective perception of the best mode of his invention did not contemplate purification by means of [* * *], Dr. Molloy's testimony also provides indirect evidence that he did not consider [* * *] significant to practicing the invention. A clear inference may be drawn from Dr. Molloy's affidavit testimony that he was not unduly concerned about developing a particular purification process or recrystallization solvent as part of his best mode; otherwise, he would have been more involved in and interested in, and perhaps would have had a better recollection of, the recrystallization procedure being developed by the other scientists at Lilly. Instead, the evidence of his relative detachment from the process indicates that he considered the method of purification a detail ancillary to his invention.

After considering all the evidence and arguments presented by the parties, we find that Barr has not produced evidence sufficient to support its clear and convincing evidentiary burden at trial to show that the inventor, Dr. Molloy, believed that *934 using [* * *] to purify fluoxetine hydrochloride enhanced the performance or provided any other significant benefit such that the method should have been disclosed in the patent application as part of his best mode of carrying out the invention. Rather, the evidence presented by Lilly and largely uncontroverted by Barr shows that any possible variance in the purity of the final fluoxetine hydrochloride product recovered by different purification methods, or even various recrystallization solvents, is not so significant in carrying out the invention as to constitute part of the inventor's best mode, particularly because the patents do not claim a specific level of purity, and that Dr. Molloy did not have a subjective preference for [* * *] as a part of his best mode. Having concluded that Barr cannot show by clear and convincing evidence that the use of [* * *] in purifying fluoxetine hydrochloride was part of Dr. Molloy's best mode for carrying out his invention subject to disclosure under 35 U.S.C. § 112 and that purification methods, including recrystallization, were known to a chemist of ordinary skill, we find that purification was adequately disclosed in the patent application. Thus, we hold that the '081 and '549 patents are not invalid for violating the best mode requirement and accordingly deny Barr's motion for summary judgment and grant Lilly's motion for summary judgment on this claim.

Defendants challenge the validity of Lilly's '081 and '549 patents on the grounds that the inventor, Dr. Molloy, failed to satisfy the best mode requirement of 35 U.S.C. § 112 by not disclosing in his patent application (1) his in-house method of synthesizing p-trifluoromethylphenol, a starting material for the patented compound fluoxetine hydrochloride and (2) his use of [* * *] as a recrystallization solvent for fluoxetine hydrochloride. The parties filed cross-motions for summary judgment on these issues. For the reasons discussed in the Court's discussion above, we deny Defendants' motion for summary judgment as to the best mode defense, grant Plaintiff's motion for summary judgment as to Defendants' best mode challenge for failure to disclose the method of synthesizing the starting material and grant Plaintiff's motion for summary judgment as to Defendants' best mode challenge for failure to disclose [* * *] as a recrystallization solvent.

[1] Defendants Geneva Pharmaceuticals, Inc., Apotex, Inc. and Bernard C. Sherman move to join and adopt Barr's motion for summary judgment on the best mode defense, Barr's reply to its motion for summary judgment on the best mode defense and Barr's responses to Lilly's motions for summary judgment on the best mode defense. This motion is granted and thus the Court's rulings on these summary judgment motions with respect to Barr will also apply to these defendants.

[2] Text redacted pursuant to Court order dated January 14, 1999.

[3] Barr contends that the impact of the synthesized p-trifluoromethylphenol on the final product is irrelevant to Dr. Molloy's duty to disclose the method of synthesis (see Barr Summ. Judg. Rep. Br. at 15), but this argument is clearly misplaced in light of the governing case law outlined above that an inventor has no duty to disclose production details. Barr also argues that Lilly produces no evidence to support its assertion that there was no difference in the quality of the final product (see Barr Summ. Judg. Rep. Br. at 15), apparently forgetting that Barr, not Lilly, bears the burden of proving its defense by clear and convincing evidence.

[4] Barr objects to the introduction of this affidavit, claiming that Dr. Molloy cannot contradict his prior testimony with a self-serving affidavit that fails to recognize and explain the contradiction. See Barr Summ. Judg. Rep. Br. at 20-21. However, as explained above, Dr. Molloy's prior deposition testimony is not in conflict with his affidavit, as he never mentioned any preference for or superiority of the in-house p-trifluoromethylphenol product. His subsequent affidavit merely serves to clarify any potential ambiguity and set forth evidence of his subjective belief at the time of the patent application.

[5] Akiba can be distinguished on the ground that p-trifluoromethylphenol was not a "novel compound" itself, as set out in the Court's discussion of p-trifluoromethylphenol's commercial availability, below.

[6] This appears to us to be more an issue of enablement, which Barr does not challenge. However, we find the issue may also be relevant to a best mode analysis, as the parties argue.

[7] Lilly Opp. Exh. 3 is an excerpt from Howard I. Forman, The Law of Chemical Metallurgical and Pharmaceutical Patents (1967), in which the author explains, "Since claims in pharmaceutical patents are sensitive to being rejected as being too broad, there has evolved some terminology which has found acceptance by the Patent Office to avoid excessive breadth in claims." For example, if you claim a compound of the group consisting of A and "the salts thereof" the Examiner will in all likelihood reject the claim as "unduly broad in salts," in that "this includes toxic as well as nontoxic materials.... But if you have basis therefor in the specification, he will permit you to overcome this rejection by using the widely employed expression[] `the acid addition salts thereof formed with a pharmaceutically acceptable acid'...."

[8] Barr objects to the introduction of this affidavit, claiming that Dr. Molloy cannot contradict his prior testimony with a self-serving affidavit that fails to recognize and explain the contradiction. See Barr Summ. Judg. Rep. Br. at 20-21. However, Dr. Molloy's prior deposition testimony is not in conflict with his affidavit, as he never mentioned any preference for or superiority of [* * *] as a recrystallization solvent. His subsequent affidavit merely serves to clarify any potential ambiguity and set forth evidence of his subjective belief at the time of the patent application.