Ashland Oil, Inc. v. Delta Resins & Refractories, Inc.

776 F.2d 281, 227 U.S.P.Q. (BNA) 657, 1985 U.S. App. LEXIS 15309

• Court: Court of Appeals for the Federal Circuit
• Decided: October 25, 1985
• Docket: Appeal 84-1779
• Status: Published

Opinion

KASHIWA, Circuit Judge.

Ashland Oil, Inc. (Ashland) appeals from the judgment of the United States District Court for the Eastern District of Michigan, Southern Division, 587 F.Supp. 1406 (1984), holding claims 1, 2, 7 and 10 of U.S. Patent No. 3,485,797 (the ’797 patent), claims 14 and 19 of U.S. Patent No. 3,409,579 (the ’579 patent), and claim 17 of U.S. Patent No. 3,676,392 (the ’392 patent) invalid under 35 U.S.C. § 103. We reverse and remand.

Background

Ashland is the assignee of the three patents involved in this case, which were issued to Dr. Janis Robins. These patents are directed to certain chemical products and processes finding ultimate use in the foundry industry. One method of forming metal castings in the foundry industry involves compacting sand around a pattern to form a sand mold, removing the pattern, and then pouring molten metal into the sand mold. This process often involves the use of internal sand cores around which the molten metal flows to produce various internal configurations.

A chemical binder, for example a phenolic urethane formed by reacting a phenol-formaldehyde resin with a hardener component, such as a polyisocyanate, and a curing agent, such as a tertiary amine, is mixed with the sand, causing the sand-binder mixture to harden at a predetermined rate. After the sand mold mixture has hardened, the mixture retains its shape during the pouring of the molten metal. After the metal solidifies the binder must break down to permit the sand to be readily dislodged from the casting.

. An optimized sand-binder mixture should have a slow or negligible curing period after the initial mixing of the binder with the sand, i.e., the work time, followed by a period of rapid curing. During the work time, the sand-binder mixture remains flowable, due to negligible curing or hardening, to allow easy forming of the mixture to conform to the pattern. Rapid curing after the mold has been formed allows the sand-binder mixture to rapidly reach its hardened state, thus permitting initiation of molten metal pouring.

Ashland sued Delta Resins & Refractories, Inc. (Delta) 1 for infringement of claims 1, 2, 7 and 10 of the '797 patent, claims 14 and 19 of the ’579 patent and claim 17 of the ’392 patent. Claims 1, 2, and 7 of the ’797 patent are directed to a process for producing a phenol-formaldeyhde resin which may be used in producing a chemical binder useful in the formation sand molds. Claim 10 of the ’797 patent is directed to one of the resin products derived from this process.

Claim 1 of the ’797 patent is a broad process claim 2 directed to reacting a phenol and aldehyde in the presence of a catalyst and reads as follows:

1. A process for the preparation of phenol aldehyde reaction products which comprises

reacting a phenol having the general formula

T wherein X, Y, and Z are hydrogen, hydrocarbon radicals, oxyhydrocarbon radicals or halogen, with' an aldehyde having the general formula R’CHO wherein

R1 is hydrogen or a hydrocarbon radical of 1-8 carbon atoms at a mole ratio of aldehyde to phenol of greater than 1,

in the liquid phase under substantially anhydrous conditions with the removal of water above 100° C and

at temperatures below about 130° C in the presence of catalytic concentrations of a soluble divalent metal salt dissolved in the reaction medium.

Claim 10 of the ’797 patent is directed to a phenol-formaldehyde resin (Pep resin) 3 and reads as follows:

10. The phenol formaldehyde resin having the general

radical or halogen, meta to the hydroxyl group of the phenol; m and n are numbers the sum of which is at least two and the ratio of m-to-n is greater than one; and A is a hydrogen or a methylol group, the molar ratio of said methylol group to hydrogen being at least one. 4

Claim 14 of the ’579 patent 5 is a dependent claim 6 directed to a foundry mix which contains sand as the major constituent and up to 10% by weight, based upon the weight of the sand, of a binder composition. The binder composition comprises in admixture the Pep resin as described in claim 10 of the ’797 patent, a hardener component comprising a liquid polyisocyanate containing at least two isocyanate groups, and a curing agent comprising a tertiary amine.

Claim 19 of the ’579 patent is a dependent claim 7 which reads as follows:

19. The process of preparing shaped foundry products which comprises:

(a) forming a foundry mix by uniformly distributing on a foundry aggregate containing sand as the major constituent a binding amount of up to 10% based on the weight of the aggregate of a binder composition obtained by combining [a phenolic resin having the general formula

wherein R is hydrogen or a phenolic substituent meta to the hydroxyl group of the phenol, m and n are numbers the sum of which is at least 2, and the ratio of m-to-n is at least 1, and X is a hydrogen or a methylol group, the molar ratio of said methylol group-to-hydrogen being at least 1] and hardener component of claim 1, said polyisocyanate being employed in a concentration of 10 to 5007» by weight of the phenolic resin;

(b) shaping the foundry mix in a mold; and

(c) contacting the shaped foundry mix with a tertiary amine until the binder is cured.

(bracketed material added, see supra note 7).

Claim 17 of the ’392 patent 8 is a dependent claim 9 which reads as follows:

17. A foundry mix containing sand as the major constituent, and a binding amount of up to 10 percent based on the weight of sand of the resin composition [ , said resin composition comprising in admixture,

a benzylic ether resin which has the general

the hydroxyl group of the phenol, m and n are numbers the sum of which is at least 2, X is an end-group from the group consisting of hydrogen and methylol, and wherein m is at least 1 and the sum of m and the number of methylol end-groups is at least two,

a hardener component comprising a liquid

polyisocyanate containing at least two isocyanate groups and present in an amount equal to 10 to 500 weight percent based on the weight of the resin, and

a curing catalyst having a pK^, value in the

range of about 7 to 11 and present in an amount equal to 0.01 to 10.0 weight percent based on the weight of the resin].

(bracketed material added, see supra note 9).

District Court Proceedings

The district court noted that although the ultimate question of patent validity is a legal one, the determination of obviousness lends itself to several basic factual inquiries, to wit, the scope and content of the prior art, differences between the prior art and the claims at issue, and the level of ordinary skill in the pertinent art. The court recognized that Delta, as the party asserting patent invalidity, has the burden of proving that the claimed inventions in issue would have been obvious by clear and convincing evidence. Once Delta has established a prima facie case of obviousness, the burden of going forward shifts to Ashland to rebut with evidence of nonobviousness, although the burden of persuasion remains with Delta.

The court found that a person of ordinary skill in the art would have a bachelor’s degree in chemistry, several years experience in phenolic and urethane chemistry, and several months exposure to the foundry art.

a. ’797 Resin Claim:

The court stated that Delta relied primarily upon three prior art references to support its argument that the invention set forth in claim 10 would have been obvious: (1) U.S. Patent No. 2,079,633 (the Rothrock patent); (2) N. MEGSON, PHENOLIC RESIN CHEMISTRY (Academic Press Inc. 1958); and (3) R. MARTIN, THE CHEMISTRY OF PHENOLIC RESINS (John Wiley & Sons, Inc. 1956).

Claim 10 of the ’797 patent, the court stated, requires that the sum of m and n must be at least two such that Ashland’s Pep resin must have at least three rings and may have up to forty. The Pep resin contains some three ring and greater molecules, along with a substantial amount of one and two ring adducts. The court found 10 that the process disclosed in the Rothrock patent produces claim 10 material, although having a large portion of ad-ducts and only a small amount of three ring or greater molecules. Ashland denied that the Rothrock process produces the Pep resin of claim 10, and averred that the Rothrock resin is inferior to the Pep resin as a foundry binder because Rothrock’s hydroxyl groups are modified by butyl alcohol solvents.

As to the MEGSON reference, diagram A, see Appendix, illustrates a polybenzylic ether resin within the scope of claim 10. The drawing shows phenol-formaldehyde resin molecules having an ortho-ortho orientation, connected by ether bridges, and an open para position. Claim 10 requires the sum of methylene and ether bridges to be at least two, and while the molecule of the diagram does not show methylene bridges, it is still within the scope of claim 10 because it satisfies the requirement that the sum of methylene and ether bridges with its two ether bridges. Ashland’s position is that MEGSON does not disclose instructions for the preparation of this polybenzylic ether resin, and further that the molecule depicted in diagram A is only a hypothetical structure postulated to be present during curing, not what is present in a room temperature reaction product.

MARTIN discloses a linear polymeric ether resin containing up to thirty-five phenol rings linked together by ether bridges in an ortho-ortho orientation. Jordan Kopac, Delta’s president, 11 testified that MARTIN teaches how ether linkages are formed, and how the number of ether linkages may be increased. Further, as temperature increases, some ether linkages will break down, producing methylene linkages and formaldehyde, the formaldehyde being available to cross-link the resin. The position of Ashland was that the open para position of the Pep resin allows phenol reaction at this site, an important consideration in producing a superior binder, but MARTIN teaches a structure which is para substituted.

Based upon this prior art, the court found that while no single prior art reference rendered the Pep resin of claim 10 obvious, the references taken together would have suggested that resin. See Leinoff v. Louis Milona & Sons, Inc., 726 F.2d 734, 739, 220 USPQ 845, 848-49 (Fed. Cir.1984). Therefore, Delta had sustained its burden of proving by clear and convincing evidence that the Pep resin of claim 10 would have been obvious to one of ordinary skill in the art.

The Rothrock patent, MEGSON, and MARTIN collectively suggest the critical elements of the claimed material, i.e., a phenol-formaldehyde resin containing linear polymers which consist of phenol rings connected by ether bridges or ether and methylene bridges in an ortho-ortho orientation having an unsubstituted para position.

The court held that Ashland’s evidence was insufficient to rebut Delta’s clear and convincing evidence that claim 10 would have been obvious to one skilled in the art in light of the prior art, and that Ashland failed to establish that one of ordinary skill in the art would have been unable to read the prior art references and “discover” the resin claimed by Robins.

b. '797 Process Claims:

The court stated that Delta had sustained its burden of proving by clear and convincing evidence that the prior art discloses reacting phenol with formaldehyde under essentially the same conditions as the Robins patent, finding that the prior art cited by Delta would have suggested to one of ordinary skill in the mid 1960s the possibility of developing the process claimed in claims 1, 2 and 7 of the ’797 patent. The Rothrock patent describes a process for manufacturing phenol-formaldehyde resins, teaching a process using: (1) a formaldehyde/phenol ratio greater than 1; (2) paraformaldehyde (the anhydrous form of formaldehyde), and removal of water in some cases; (3) a temperature range of 100-120° C; and (4) soluble metal salt catalysts, including zinc acetate.

Japanese Patent 13696/60 describes a process for producing phenol-formaldehyde initial condensates, by reacting phenol and formaldehyde: (1) under anhydrous polymerization conditions (starting with paraformaldehyde and removing water); (2) at temperatures above 100° C and as high as 120° C; and (3) using soluble metal salts as catalysts. Although the examples of the Japanese Patent teach a formaldehyde/phenol ratio less than 1, the specification teaches that a formaldehyde/phenol ratio greater than 1 may be used.

The prior art reference of Fraser, Hall & Raum, Preparation of ‘High-Ortho’ Novolak Resins, J.App.Chem. (Dec. 1957), teaches the effectiveness of zinc and lead as catalysts to form ortho-ortho linked phenol-formaldehyde chains, and that benzylic ether bridges are formed at reaction temperatures below 140° C.

There are differences between this prior art and the ’797 process claims. The Roth-rock patent does not teach removal of water above 100° C and the Fraser reference does not teach the removal of water at all. Both Rothrock and the Japanese Patent use butyl alcohol as a solvent, whereas Robins discloses the use of toluene. The ’797 claims in issue, however, do not mention the use of solvents. The butyl alcohol modified resin of Rothrock is not a phenolic resin. Ashland has also argued that neither the Japanese Patent nor Fraser produce compounds with more than two rings.

But, the court stated that Ashland had failed to establish that these differences, in light of Delta’s proof, are great enough to render the inventions in issue non-obvious. The cited references collectively, if not individually, teach: (1) a formaldehyde/phenol ratio greater than 1; (2) anhydrous conditions; (3) a reaction temperature range of 100-120° C; and (4) soluble metal salt catalysts. The differences between the prior art and the claims in issue are insignificant because one of ordinary skill in the art could study the prior art references and come upon the ’797 process claims.

For example, one of ordinary skill could read Rothrock and recognize that varying the solvent in Example 5 and removing water — as Rothrock did in Examples 1 and 7 — yields a process, which could be substantially similar to the ’797 process, for preparing certain phenol-formaldehyde reaction products. Similarly, although the Japanese Patent and the Fraser reference do not produce compounds with greater than two rings, one of ordinary skill reading these pieces of prior art could apply his knowledge and develop a process for preparing a phenolic resin which could be substantially similar to the process of the ’797 patent.

c. The ’579 and ’392 Foundry Binder Claims:

The ’579 and ’392 foundry binder systems consist of the Pep resin, a polyisocyanate hardener, and a tertiary amine catalyst for the ’579 claims or a catalyst with a pkb value of about 7 to about 11 for the ’392 claim.

Delta argued that the prior art, i.e., U.S. Patents Nos. 3,409,571 (’571 patent) and 3,398,122 (’122 patent) issued to Shepard and British Patent 1,031,909, disclosed the use of phenolic urethanes as foundry binders. Dr. Frisch, an expert witness, testified that Robins’ foundry binder claims would not have been obvious to one skilled in the art. One skilled in the art would not expect polyurethanes to work as foundry binders since it was known in the field that reacting phenol with isocyanates results in a blocked phenol forming an unstable urethane which disassociates or reverts to phenol and isocyanates upon heating at temperatures of 140-150° C.

The court’s examination of the prior art led it to conclude otherwise. While recognizing that the British Patent 12 was technically not prior art, it was indicative of what was known to persons of ordinary skill in the art. The patent described reacting novolae resins with highly reactive divalent materials to form a soluble product which can then be thermoset to produce a foundry binder, and such a claim is made in claim 12. The court concluded that the British Patent discloses the use of phenolic urethanes as foundry binders.

The court found the Shepard patents sig-' nificant in that Shepard described a novolae phenolic resin modified with a phosphorous compound, i.e., a soluble thermoplastic. Shepard’s ’571 patent clearly states that thermosetting products can be produced by mixing thermoplastic products with polyisocyanates, and that such thermosetting products are useful as foundry sand binders. Based upon the foregoing, the court concluded that the use of phenolic urethanes as foundry binders was taught by the prior art.

While Ashland argued that Shepard teaches the use of a novolae resin while the ’579 and ’392 claims here in issue use the Pep resin, the court found this difference insufficient to warrant a finding of nonobviousness. One skilled in the art could readily sense that the Pep resin might be substituted into the Shepard patent. It was known in the prior art how ether bridges and OH groups react with polyisocyanates. One skilled in the art could look at MEGSON, MARTIN and the Rothrock patent, analyze their teachings in light of Shepard and the British Patent, and conclude that a polybenzylic ether resin could be plugged into Shepard to produce a phenolic urethane foundry binder.

The court found that Delta had sustained its burden of proving that one of ordinary skill in the art of phenolic chemistry would have found it obvious to use tertiary amines to promote the reaction between the Pep resin and polyisocyanates. J. SAUNDERS & K. FRISCH, POLYURETHANES: CHEMISTRY AND TECHNOLOGY (Interscience Pub. 1962, reprint 1978), teaches that at low temperatures “one normally uses a catalyst such as a tertiary amine or aluminum chloride to promote this reaction.” Further, Shepard’s ’571 and ’122 patents, and U.S. Patent Nos. 3,242,107, 3,282,896, and 3,043,794 disclose that tertiary amines in lieu of or in addition to heat promote the reaction between phenolic resins and polyisocyanates.

The court also found that Delta had sustained its burden of proof to show that the use of a curing agent with a pKb value of about 7 to about 11 to promote the reaction between the Pep resin and polyisocyanates would have been obvious to one of ordinary skill. The SAUNDERS reference disclosed work carried out in the 1940s involving catalysts with the pKb range as claimed in claim 17 of the ‘392 patent which showed that base strength is a controlling factor of a catalyst’s effectiveness in urethane formation. Moreover, the Shepard patents and U.S. Patent Nos. 3,156,659, 3,063,964, and 2,906,717 disclosed the use of various catalysts with a pKb value in the range of claim 17 to promote the reaction between phenolic resins and polyioscyanates.

d. Secondary Considerations:

The court, citing to W.L. Gore & Associates, Inc. v. Garlock, Inc., 721 F.2d 1540, 220 USPQ 303 (Fed.Cir.1983), cert. denied, - U.S. -, 105 S.Ct. 172, 83 L.Ed.2d 107 (1984), and Stratoflex, Inc. v. Aeroquip Corp., 713 F.2d 1530, 218 USPQ 871 (Fed. Cir.1983), stated that it had considered relevant secondary considerations before reaching the conclusion that the Robins’ patents would have been obvious. The court noted that there were no independent secondary considerations relevant to the ‘797 patent apart from its use in Isocure, the commercial foundry mix patented under the ‘579 patent, and Pep Set, the commercial foundry mix patented under the ‘392 patent.

The court found the commercial success of Isocure and Pep Set impressive, noting that Ashland had sold millions of pounds annually of these products, and that both products enjoyed an increasing market share. While noting that Ashland had granted licenses under these patents to Combustion Engineering Company and International Minerals and Chemical Company, the court further found that after Ash-land lost the Milwaukee litigation Combustion Engineering sought a declaratory judgment that the patent claims here in suit were invalid, and subsequently settled for a renegotiated royalty rate decrease from 12.5% to 5%. International Minerals was found to have gone out of the business approximately one year after it was granted a license.

The court noted that Ashland had offered proof that Isocure and Pep Set had received recognition from the foundry industry in the form of awards and write-ups in trade publications. The court, however, found this recognition directed more towards the marketing of, rather than the invention of, these products. The court found it significant that Dr. Robins had not received any recognition from the industry and only $200 from Ashland for his role in developing Isocure and Pep Set.

The court stated that the law is well established that commercial success alone, or combined with other secondary evidence, is insufficient to establish patentability where primary indicia of patentability is lacking. After weighing Isoeure and Pep Set’s secondary considerations including commercial success against the primary indicia of obviousness — disclosures in the pri- or art — the court concluded that all the patent claims in suit were invalid for obviousness.

OPINION

A determination of whether the subject matter of claims in issue would have been obvious under 35 U.S.C. § 103 involves factual findings with respect to: (1) the scope and content of the prior art; (2) the level of ordinary skill in the art; (3) the differences between the claimed subject matter and the prior art; and (4) where relevant, objective evidence of nonobviousness, e.g., long-felt need, commercial success, failure of others, copying, unexpected results, i.e., the secondary considerations. Perkin-Elmer Corp. v. Computervision Corp., 732 F.2d 888, 894, 221 USPQ 669, 674 (Fed.Cir.1984); Jones v. Hardy, 727 F.2d 1524, 1527, 220 USPQ 1021, 1023 (Fed. Cir.1984); W.L. Gore & Associates, Inc. v. Garlock, Inc., 721 F.2d 1540, 1550, 220 USPQ 303, 311 (Fed.Cir.1983), cert. denied, - U.S. -, 105 S.Ct. 172, 83 L.Ed.2d 107 (1984). These factual findings serve as the foundation upon which the court bases its ultimate conclusion regarding the obviousness of the claimed subject matter as a whole. Lear Siegler, Inc. v. Aeroquip Corp., 733 F.2d 881, 890, 221 USPQ 1025, 1033 (Fed.Cir.1984). This court reviews the ultimate conclusion of obviousness as one of law on which it must exercise independent judgment. Union Carbide Corp. v. American Can Co., 724 F.2d 1567, 1573, 220 USPQ 584, 589 (Fed.Cir.1984).

A patent is presumed valid, and the burden of establishing invalidity as to any claim of a patent rests upon the party asserting such invalidity. 35 U.S.C. § 282 (1982). The presumption of validity is a procedural device that mandates that the party asserting invalidity bears the initial burden of establishing a prima facie case of obviousness under 35 U.S.C. § 103. Stratoflex, Inc. v. Aeroquip Corp., 713 F.2d 1530, 1534, 218 USPQ 871, 875 (Fed. Cir.1983). Once a prima facie case has been established, the burden shifts to the patentee to go forward with rebuttal evidence showing facts supporting nonobviousness. Ralston Purina Co. v. Far-Mar-Co, Inc., 772 F.2d 1570, 1573 (Fed.Cir.1985); accord, In re Piasecki, 745 F.2d 1468, 1472, 223 USPQ 785, 788 (Fed.Cir.1984). The party asserting invalidity, however, always retains the burden of persuasion on the issue of obviousness until a final judgment is rendered. Hughes Aircraft Co. v. United States, 717 F.2d 1351, 1359, 219 USPQ 473, 478 (Fed.Cir.1983); Stratoflex, 713 F.2d at 1534, 218 USPQ at 875. Each fact forming the factual foundation upon which the court bases its ultimate conclusion regarding the obviousness of the claimed subject matter as a whole must be established by clear and convincing evidence. Lindemann Maschinenfabrik GMBH v. American Hoist & Derrick Co., 730 F.2d 1452, 1459, 221 USPQ 481, 486 (Fed.Cir.1984); SSIH Equipment Co. S.A. v. United States International Trade Commission, 718 F.2d 365, 375, 218 USPQ 678, 687 (Fed.Cir.1983).

On appeal, however, the party subject to the adverse judgment on the issue of validity, in this case the patentee Ashland Oil, bears the burden of showing either that the district court committed reversible legal error in its ultimate conclusion as to obviousness, or that the district court’s probative factual findings underlying its ultimate conclusion on obviousness were clearly erroneous. 13 Fromson v. Advance Offset Plate, Inc., 755 F.2d 1549, 1555, 225 USPQ 26, 30 (Fed.Cir.1985).

A. CLAIM 10 OF THE ‘797 PATENT — PEP RESIN

The district court found that the Pep resin of claim 10 contained some three ring and greater molecules, along with a substantial amount of one and two ring ad-ducts, that the process taught by the Roth-rock patent produces claim 10 material, although having a large portion of adducts and only a small amount of three ring or greater molecules, that MEGSON taught phenol-formaldehyde resins having an or-tho-ortho orientation, connected by ether bridges and having an open para position, and that MARTIN taught a linear polymeric ether resin having up to thirty-five phenol rings linked by ether bridges, some of these ether linkages breaking down at higher temperatures to produce methylene linkages and formaldehyde. Based upon these findings, the court concluded that the Pep resin of claim 10 of the ‘797 patent would have been obvious to one of ordinary skill in the art inasmuch as the Rothroek patent, MEGSON, and Martin collectively suggested the critical elements of Pep resin here at issue.

Before reviewing the factual findings made by the district court with respect to the teachings of each of the individual references, and the propriety of combining the teachings of these references, we find it appropriate to address several statements made by the district court. See Union Carbide Corp., 724 F.2d at 1574, 220 USPQ at 590 (while faulty reasoning may lead to a wrong result, appellant must show not only error in reasoning but error in result).

A.l Combining References

First, the court stated that Ash-land had failed to establish that one of ordinary skill in the art would have been unable to read the prior art references and “discover” the Pep resin claimed by Robins. The law does not impose a burden on Ashland to establish that the combined teachings of the individual prior art references would not have led one skilled in the art to discover the Pep resin of claim 10. The ultimate burden of establishing invalidity rests upon the party espousing such. Stratoflex, 713 F.2d at 1534, 218 USPQ at 875. Where the party asserting invalidity must rely upon a combination of prior art references to establish invalidity, that party bears the burden of showing some teaching or suggestion in these references which supported their use in combination. W.L. Gore, 721 F.2d at 1552, 220 USPQ at 312. It is legal error to place this burden on the patentee.

A.2 35 U.S.C. § 282

Further, if this statement is interpreted to place upon the patentee the burden of establishing the validity of his patents, it is at odds with established case law. Section 282 of Title 35 places the burden for the initial production of evidence, Stratoflex, 713 F.2d at 1534, 218 USPQ at 875, and the ultimate burden of persuasion on the issue of validity on the party asserting patent invalidity. Hughes Aircraft, 717 F.2d at 1359, 219 USPQ at 478; Stratoflex, 713 F.2d at 1534, 218 USPQ at 875. While the burden for the production of evidence shifts to the patentee once a prima facie case of invalidity is established, Ralston Purina, at 5; Piasecki, 745 F.2d at 1472, 223 USPQ at 788, the ultimate burden remains with the party asserting invalidity, in this instance Delta, to establish that the claims of the patents here at issue are invalid. There is no burden on Ashland to establish that the claims of these patents are valid, and it is impermissible for a trial court to shift this burden to the patentee. Jones, 727 F.2d at 1528-29, 220 USPQ at 1025.

A.3 Evidence vis-a-vis Obviousness

The court also held that Ashland’s evidence was insufficient to rebut Delta’s clear and convincing evidence on the obviousness of claim 10 of the ‘797 patent. While on this record we cannot say that this holding by the district court was erroneous, it is open to an interpretation 14 at odds with the established case law, and for this reason we set forth a brief explication of the relevant legal principles. All facts relevant to the issue of obviousness, both the facts established by the party asserting invalidity and the facts established by the rebuttal evidence submitted by the patentee, must be fully considered by the court prior to reaching its conclusion on obviousness. W.L. Gore, 721 F.2d at 1555, 220 USPQ at 314; Stratoflex, 713 F.2d at 1539, 218 USPQ at 879. These facts must be established by clear and convincing evidence. Lindemann Maschinenfabrik, 730 F.2d at 1459, 221 USPQ at 486; SSIH Equipment, 718 F.2d at 375, 218 USPQ at 687.

A.k The Pep Resin

The claims of a patent measure and define the invention. Jones, 727 F.2d at 1528, 220 USPQ at 1024. A § 103 determination requires an evaluation of the prior art references with respect to the claimed invention. Lear Siegler, 733 F.2d at 890, 221 USPQ at 1033; Union Carbide, 724 F.2d at 1574-75, 220 USPQ at 590-91. The claims here in issue are to be read and construed 15 in light of the specification and prosecution history of the patent. ACS Hospital Systems, Inc. v. Montefiore Hospital, 732 F.2d 1572, 1577, 221 USPQ 929, 932 (Fed.Cir.1984). The district court found that the Pep resin of claim 10 contained some three ring and greater molecules, along with a substantial amount of one and two ring adducts. This finding is clearly erroneous, being based upon a misconstruction of the governing law and an interpretation of claim 10 which is erroneous as a matter of law. Cf. Lemelson v. United States, 752 F.2d 1538, 1552, 224 USPQ 526, 534 (Fed.Cir.1985).

The novel phenol-aldehyde resin as claimed in claim 10 is a linear phenolic resin wherein the sum of m and n must be at least two such that the phenolic resin as claimed comprises only molecules having three or more linked-phenol rings. Claim 10 does not claim one or two ring adducts, i.e., dimethylol phenols, benzylic ethers or methylene-bridged diphenols. 16 Relevant prior art for this § 103 determination requires references which disclose phenolic polymers having three or more phenol rings, phenol rings linked by benzylic ether and methylene bridges, and phenol chains having at least one terminal methylol group.

A. 5 Opinion Testimony

While objective factual evidence going towards a § 103 determination is preferable to statements of opinion on the issue, the nature of the matter sought to be established, as well as the strength of the opposing evidence, must be taken into consideration in assessing the probative value of expert opinion. In re Oelrich, 579 F.2d 86, 91, 198 USPQ 210, 215 (CCPA 1978). Opinion testimony rendered by experts must be given consideration, and while not controlling, generally is entitled to some weight. See FED.R.EVID. 701-704; Orthopedic Equipment Co. v. United States, 702 F.2d 1005, 1012, 217 USPQ 193, 199 (Fed.Cir.1983). Lack of factual support for expert opinion going to factual determinations, however, may render the testimony of little probative value in a validity determination. Cf. In re Altenpohl, 500 F.2d 1151, 1158, 183 USPQ 38, 44 (CCPA 1974). While the opinion testimony of a party having a direct interest in the pending litigation is less persuasive than opinion testimony by a disinterested party, it cannot be disregarded for that reason alone and may be relied upon when sufficiently convincing. Cf. In re McKenna, 203 F.2d 717, 720, 97 USPQ 348, 350-51 (CCPA 1953).

The district court found that the process disclosed in the Rothrock patent produced claim 10 material having a large portion of adducts, with only a small amount of three phenol ring or greater molecules. The bases for this finding were the objective teachings disclosed in the Rothrock patent, the opinion testimony given by Jordan Kopac, Delta’s CEO, who although not qualified as an expert was within the category of one skilled in the art as found by the district court, see supra note 11, and the opinion testimony of Dr. Robert Conley, Ashland’s expert witness. The court did not make any explicit credibility determinations with respect to the opinion testimony of Mr. Kopac and Dr. Conley, nor did the court give any indication as to the weight accorded this testimony.

A. 6 The Rothrock Patent

The Rothrock patent disclosed a process for forming a heat-hardening unmodified phenol-formaldehyde resin. 17 There was no disclosure or teaching as to the chemical structure of this phenol-formaldehyde resin, i.e., what product was formed through the use of this process. W.L. Gore, 721 F.2d at 1550, 220 USPQ at 311. The district court did not point to any supporting statements or teachings in the Rothrock patent as a basis for its finding that the process of Rothrock produced a claim 10 phenolic resin.

Each element of a claim is material. Lemelson, 752 F.2d at 1551, 224 USPQ at 533. The process for producing the phenolic resin as claimed in claim 10 of the ‘797 patent requires the removal of water above 100 °C during the process. 18 This removal of water occurs during the condensation stage of the ‘797 process wherein the previously formed phenol-formaldehyde adduct is condensed to form the claim 10 product. The specification of the ‘797 patent teaches the importance of removing water during the condensation step. 19 There was no teaching in the Rothrock patent that water was to be removed during the disclosed process. 20

There is no presumptive correlation that two similar processes form substantially the same product where the processes differ by a materially limiting step. Cf. In re Hoeksema, 399 F.2d 269, 274, 158 USPQ 596, 601 (CCPA 1968) (if the prior art of record failed to disclose a method for making a claimed compound, at the time the invention was made, it cannot be legally concluded that the compound itself was in the possession of the public).

There was no objective evidence to be gleaned from the Rothrock patent which would have supported a factual finding that the Rothrock patent produced claim 10 material. Concomitantly, there was no factual support for Mr. Kopae’s opinion testimony with respect to the Rothrock patent, and consequently, Mr. Kopac’s opinion testimony is of little probative value in a validity determination. Altenpohl, 500 F.2d at 1158, 183 USPQ at 44. Accordingly, the district court committed clear error when it found that the Rothrock process produced claim 10 material. 21

A.7 The MARTIN Reference

MARTIN, the court found, disclosed a linear polymeric ether resin containing up to thirty-five phenol rings linked in an ortho-ortho orientation by ether bridges. A reference, however, must have been considered for all it taught, disclosures that diverged and taught away from the invention at hand as well as disclosures that pointed towards and taught the invention at hand. W.L. Gore, 721 F.2d at 1550, 220 USPQ at 311. While MARTIN taught a polymer having the phenol rings linked together by benzylic ether bridges, as well as at least one terminal methylol group, MARTIN also taught that this polymer had an “R” substituent at the para position. MARTIN taught that this R group was a substituent other than hydrogen. The polymeric resin of claim 10, in contrast, has an open or unsubstituted para position. Not only was the para position of the MARTIN compound blocked, there was no recognition that it would have been advantageous to replace the R substituent with a hydrogen, i.e., unsubstituted para position, to increase the polymer’s reactivity.

A.8 The MEGSON Reference

The court found that the MEGSON reference illustrated a polybenzylic ether resin within the scope of claim 10, the drawing showing a phenol-formaledehyde resin molecule having ether bridge linkages at the ortho-ortho position and an open para position. The court found this molecule satisfied the m, n limitations of claim 10 by having at least two benzylic ether linking bridges, i.e., m equal to or greater than two while n equals zero. 22 But, the MEGSON reference should also have been considered for disclosures that taught away from the invention here at issue. Id., 721 F.2d at 1550, 220 USPQ at 311. The specific disclosure relied upon by the district court depicted two polymers cross-linked by a methylene derivative. Since this particular cross-linking mechanism was postulated to involve a reaction with nuclear hydrogen, the phenols of the polymers were shown as para unsubstituted, i.e., having hydrogen at the para position. The other cross-linking mechanisms, as well as a disclosure of a benzylic ether linked polymer product, depicted polymers having an “R” group at the para position. This R could have stood for hydrogen, or it could have stood for an organic radical. 23 There was no teaching in the explication of the cross-linking mechanism as to what the terminal end groups of these cross-linking structures were, i.e., there was no teaching that these cross-linking structures had a methylol terminal end group. Finally, there was uncontroverted testimony by Ashland’s expert, see supra note 23, that the disclosure in the MEG-SON reference relied upon by the district court was a hypothetical structure.

The test of whether a particular compound described in the prior art may have been relied upon to show that the claimed subject matter at issue would have been obvious is whether the prior art provided an enabling disclosure with respect to the disclosed prior art compound. Cf. In re Donohue, 766 F.2d 531, 533, 226 USPQ 619, 621 (Fed.Cir.1985); Hoeksema, 399 F.2d at 273-74, 158 USPQ at 598-99. Delta did not offer evidence that showed an enabling disclosure for the disclosed structure of MEGSON, while uncontroverted testimony showed the MEGSON structure to be a hypothetical structure.

A. 9 Conclusion

The district court concluded, in light of the Rothrock patent, MEGSON, and MARTIN, that the Pep resin as claimed in claim 10 of the ‘797 patent would have been obvious. Obviousness, however, cannot be established by combining the teachings of the prior art to produce the claimed invention unless there was some teaching, suggestion or incentive in this prior art which would have made such a combination appropriate. ACS Hospital Systems, 732 F.2d at 1577, 221 USPQ at 933; W.L. Gore, 721 F.2d at 1551, 220 USPQ at 311. The district court did not elucidate any factual teachings, suggestions or incentives from this prior art that showed the propriety of combination, nor in fact did the district court even point out what teachings from each of the references, when considered in combination, were relied upon in concluding that the invention of claim 10 would have been obvious. Nor apparently did the district court give any consideration to teachings in these references which would have led one skilled in the art away from the invention of claim 10. We would have to say that the district court used claim 10 of the ‘797 patent as a blueprint, and abstracted individual teachings from the Rothrock patent, MEGSON, and MARTIN to create the Pep resin of claim 10. W.L. Gore, 721 F.2d at 1552, 220 USPQ at 312. This was error as a matter of law. 24

We are not persuaded, based upon the foregoing, that the facts upon which the district court based its legal conclusion that the subject matter of claim 10 of the ‘797 patent would have been obvious were proven by clear and convincing evidence, Lindemann Maschinenfabrik, 730 F.2d at 1459, 221 USPQ at 486; SSIH Equipment, 718 F.2d at 375, 218 USPQ at 687, such that it cannot be said that Delta had satisfied its burden of proof. Nor was there a sufficient basis for the district court to combine the teachings of the Rothrock patent, MEGSON, and MARTIN. ACS Hospital Systems, 732 F.2d at 1577, 221 USPQ at 933. The district court erred as a matter of law in concluding that the invention of claim 10 would have been obvious.

B. THE PROCESS CLAIMS OF THE ‘797 PATENT

The district court found that the Roth-rock patent described a process for manufacturing phenol-formaldehyde resins wherein: (1) a formaldehyde/phenol ratio greater than 1 was taught; (2) the use of paraformaldehyde, the anhydrous form of formaldehyde, was disclosed; (3) a temperature range of 100-120 °C was taught; (4) the removal of water was taught in certain examples; and (5) the use of soluble metal salt catalysts including zinc acetate was disclosed. One point of contention 25 as to what the Rothrock patent disclosed was whether Rothrock taught one skilled in the art the removal of water during the process which was consonant in scope to the water-removal limitation of the ‘797 process claims.

B.l Robins’ ‘797 Process

Process claim 1 of the ‘797 patent, an independent claim, describes a material claim limitation which requires the phenolic-resin producing reaction to be conducted “under substantially anhydrous conditions with the removal of water above 100 °C.” The materiality of this limitation is disclosed in the ‘797 specification wherein it is stated that the “failure to continuously remove water not only affects the activity of the catalysts, but also the structure of the product formed, in permitting, for example, para-substitution”, and that “the presence of water results in reaction products which cannot be cured to mechanically strong resins by the use of acidic agents at room temperature.” (Our emphasis). The specification further discloses that the “process of the present invention is carried out in equipment which will provide for the continuous removal of water from the reaction mixture.” (Our emphasis). The water-removal limitation of the process claims of the ‘797 patent, therefore, requires that water be continuously removed during the polymer formation stage of the reaction, ACS Hospital Systems, 732 F.2d at 1577, 221 USPQ at 932 (the claim here in issue is read and construed in light of the specification), i.e., at temperatures above 100 °C where the phenol-formaldehyde adduct — a mixture of dimethylol phenols, benzylic ethers and methylene-bridged phenols — is condensed to form the three-ring or greater phenolic resin.

B.2 The Rothrock Process

In contrast, the Rothrock patent in general did not teach or suggest the removal of water during the process described therein, nor was there any teaching or suggestion that the removal of water during the process was a critical limitation. More particularly, there was no teaching or suggestion that water was to be removed during the phenolic resin formation stage of the reaction, i.e., at temperatures above 100 °C. Only Example I of Rothrock disclosed a removal of water, teaching that a small amount of water which had formed on the sides of the reaction flask was removed from the resulting clear liquid formed by the process. 26 This water was removed, however, after the resin of the process had been formed, i.e., at temperatures well below 100 °C. The equipment for the process of Rothrock as disclosed in the examples did include an air condensor. 27 Mr. Kopac testified that there was no disclosure or teaching in the Rothrock patent of the function being performed by the air condensor, and that it could not be said with any certainty that the air condensor functioned to remove or retain water in the reaction zone during the Rothrock process. Examples II, III, IV and VII of the Rothrock patent disclosed that the reaction mixture was heated at reflux. 28 This disclosure would have suggested to one skilled in the art that water was not removed, nor was there any necessity for doing so, during the reaction process of Rothrock. W.L. Gore, 721 F.2d at 1550, 220 USPQ at 311 (a reference must have been considered in its entirety, for disclosures which taught away from an invention as well as disclosures which directed one to the invention). There was no clear teaching or suggestion in Rothrock that water was to be removed at any step during the condensation process disclosed therein, but rather only after the phenolic condensation was completed. Moreover, the Rothrock reference as a whole suggested that water was retained during the reaction process. See supra note 27 and text following, and note 28.

The district court found that the Rothrock patent had not taught the removal of water above 100° C. Yet, the district court subsequently found that one of ordinary skill could have read Rothrock and recognized that varying the solvent in Example V and removing water — as Rothrock had done in Examples I and VII 29 — yielded a process which could have been substantially similar to the ‘797 process. These findings by the court are in direct conflict. Since the ‘797 process claims contain a material limitation directed to “the removal of water above 100° C”, Lemelson, 752 F.2d at 1551, 224 USPQ at 533, and since the district court found that Rothrock did not teach removal of water above 100° C, there was no basis for the court’s finding that the Rothrock process was substantially similar to the ‘797 process since the Roth-rock process lacked this material limitation. See supra note 21. Accordingly, we hold that this finding of the district court is clearly erroneous.

B.3 The Japanese Patent

The Japanese Patent, the court found, described a process for producing phenol-formaldehyde initial condensates, by reacting phenol and formaldehyde: (1) under anhydrous polymerization conditions, i.e., starting with paraformaldehyde and removing water; (2) at temperatures above 100 °C and as high as 120 °C; and (3) using soluble metal salts as catalysts. Further, the district court found that although the examples in the Japanese Patent taught a formaldehyde/phenol ratio less than 1, the specification taught that this ratio could be greater than 1.

While the process disclosed in the Japanese Patent did teach the use of paraformaldehyde, or another substance having the same effectiveness, and the continuous removal of water, the Japanese Patent should have been considered in its entirety, with due consideration given to disclosures that diverged or taught away from the invention here at issue, as well as disclosures which directed one skilled in the art to the invention. W.L. Gore, 721 F.2d at 1550, 220 USPQ at 311.

The Japanese Patent disclosed a process for producing a phenol-formaldehyde initial condensate in which phenol was added to a polar solvent such as ordinary alcohol, paraformaldehyde or a substance having similar effectiveness was dissolved into the phenol directly in a molar ratio range of 0.5-1.5 of formaldehyde/phenol, and this mixture was reacted in the presence of a weak alkaline catalyst 30 with the continuous elimination of water to produce a liquid mixture of 2-methylolphenol or 2-methylol-phenol and 2, 6-dimethylolphenol. This resulting liquid mixture was then acidified to induce a condensation reaction to form 2, 2’-dihydroxydiphenyl methane and/or other methylolation products. The specification further taught that the hydroxide induced the ortho orientation of formaldehyde in a non-water system and that alkoxyphenoxymethane accelerated the ortho linkage during condensation. 31 The specification further taught that the process reaction could have been effected without dissolving the phenol in a polar solvent, i.e., a solvent need not be used.

In contrast, the ‘797 specification teaches that the soluble metal salt catalyst is a metal ionically bonded to a salt radical, and that this salt radical should be that of a stronger acid, one having a dissociation constant greater than 10-8, to prevent cross-linking during the formation of the reaction product. Further, while the process claims of the ‘797 patent do not incorporate a specific limitation calling for the use of a solvent, the specification discloses that in the preferred embodiment of the process a non-polar organic solvent is utilized. While the Japanese Patent taught that the use of a solvent was optional, it also taught that when a solvent was used, it must have been a polar solvent. W.L. Gore, 721 F.2d at 1551, 220 USPQ at 311.

To the extent that the district court concluded that the Japanese Patent by itself would have rendered the subject matter of the process claims of the ‘797 patent obvious, this conclusion is erroneous as a matter of law. The Japanese Patent clearly taught, at a minimum, 32 that an alkaline catalyst, to induce the ortho orientation of the formaldehyde, was a material element of the process. See supra note 31. Moreover, the reaction of this process was a two stage reaction wherein the reactants were first exposed to an alkaline catalyst, and the resulting liquid mixture was then acidified to produce the final reaction product. The process claims of the ‘797 patent, in contrast, do not use an alkaline catalyst, nor is the ‘797 process a two stage reaction for the formation of a phenolic resin which requires initial reaction of the phenol and formaldehyde with an alkaline catalyst, and then the acidification of the resulting liquid mixture.

To the extent that the district court concluded that the teachings of the Japanese Patent could have been combined with the teachings of the Rothrock Patent to reach the conclusion that the subject matter of the ‘797 process claims would have been obvious, this conclusion is erroneous as a matter of law. See supra note 24. The district court did not point to any teachings or suggestions in either reference which would have led one skilled in the art to perceive an advantage to be derived from their combination. ACS Hospital Systems, 732 F.2d at 1577, 221 USPQ at 933; W.L. Gore, 721 F.2d at 1551, 220 USPQ at 311. In point of fact, the teachings of the two references would have led one skilled in the art away from their combination. Roth-rock taught that his process could not be practiced using an alkali earth metallic hydroxide catalyst. In contrast, the Japanese Patent required such a catalyst. Rothrock required the use of a solvent, which could have been either polar or non-polar while the Japanese Patent taught that the use of a solvent was optional. Further, the Japanese Patent taught that when a solvent was used, it must be polar. The Japanese Patent also taught that the formaldehyde/phenol ratio could have a range of 0.5-1.5 while Rothrock required that the ratio must be greater than 1.

B.4 The Fraser Reference

The district court found that the Fraser reference taught the effectiveness of zinc and lead as catalysts to form ortho-ortho linked phenol-formaldehyde chains and that ether bridges were formed at reaction temperatures below 140 °C. The court also found, however, that the Fraser reference did not teach the removal of water above 100 °C and that the method of Fraser did not produce compounds having more than two phenol rings.

A reference, however, should also have been considered for its antithetical teachings. W.L. Gore, 721 F.2d at 1550, 220 USPQ at 311. One of the critical teachings of Fraser was that the formaldehyde/phenol ratio must be less than 1, i.e., a molar excess of phenol was required for the method. It would be error as a matter of law if the district court concluded that the subject matter of the process claims of the ‘797 patent would have been obvious in view of the Fraser reference alone. Nor was there any teaching or suggestion that would have led one skilled in the art to combine the teachings of the Fraser reference with either the teachings of the Roth-rock patent or the teachings of the Japanese Patent, or both. ACS Hospital Systems, 732 F.2d at 1577, 221 USPQ at 933.

B.5 Conclusion

We held supra that the district court’s finding that the Rothrock process was substantially similar to the process of the ‘797 patent was clearly erroneous. Further, we are not persuaded that the other facts, discussed supra, which the district court relied upon in reaching its legal conclusion that the subject matter of claims 1, 2, and 7 of the ‘797 patent would have been obvious, were proven by clear and convincing evidence, Lindemann Maschinenfabrik, 730 F.2d at 1459, 221 USPQ at 486; SSIH Equipment, 718 F.2d at 375, 218 USPQ at 687, such that it cannot be said that Delta had sustained its burden of proof before the district court. It was error as a matter of law for the district court to conclude that processes of claims 1, 2 and 7 of the ‘797 patent would have been obvious. 33

C. THE CLAIMS OF THE ‘392 AND ‘579 PATENTS

Claim 17 of the ‘392 patent is directed to a foundry mix having sand as the major constituent and up to 10 percent by weight of sand of a resin composition. The resin composition comprises in admixture a benzylic ether resin substantially similar to the phenolic resin claimed in claim 10 of the ‘797 patent, 34 a hardener component defined as a liquid polyisocyanate, and a curing catalyst defined as a base having a pKb value in the range of about 7 to about 11. Claim 14 of the ‘579 patent is directed to a foundry mix similar to that of claim 17, except that the phenolic resin of the binder composition is the resin as claimed in claim 10 of the ‘797 patent, and the curing agent is defined as a tertiary amine. Claim 19 of the ‘579 patent is directed to a process for preparing foundry shapes using the foundry mix of claim 14.

C.l General Level of Skill in the Art

The district court found that British Patent No. 1,031,909, although technically not prior art, was indicative of what was generally known during the relevant time frame to persons of ordinary skill in the foundry art. Finding that the British Patent described reacting novolac resins with highly reactive divalent materials 35 to produce soluble, fusible polymers which may be employed “as binders for sand (foundry resins)”, the court concluded that the British Patent disclosed the use of phenolic urethanes as foundry binders. 36 The court also found that U.S. Patent Nos. 3,398,122 and 3,409,571 (the ‘122 and ‘571 patents, respectively), issued to Shepard, were significant for a teaching that phenolic urethanes were useful as constituents of foundry binders. Shepard, the court found, described a soluble thermoplastic, i.e., a novolac phenolic resin modified with a phosphorous compound, which could be mixed with polyisocyanates to form thermosetting products useful in foundry sand binders. Based upon the district court’s findings with respect to the Shepard references, we cannot say that the district court erred in concluding that phenolic urethanes were taught as having utility in foundry binders.

C.2 The Shepard Patents

The district court found that one skilled in the art “could also readily sense that the ‘Pep’ resin might be substituted into the Shepard patent.” Since it was known in the prior art how ether bridges and hydroxyl groups reacted with polyisocyanates, 37 one of ordinary skill in the art could have looked at the Rothrock patent, MEGS ON, and MARTIN, analyzed their teachings in light of the Shepard patents and the British Patent, 38 and concluded that a polybenzylic ether resin could have been plugged into Shepard to produce a phenolic urethane foundry binder.

The district court’s conclusion that the Pep resin might have been substituted for the phenolic condensate used in the Shepard patent is erroneous as a matter of law, for several alternative reasons. 39 First, the district court failed to consider the ‘571 patent in its entirety, in particular for those teachings therein that would have led one skilled in the art away from the subject matter of the ‘392 and ‘579 patents. W.L. Gore, 721 F.2d at 1550, 220 USPQ at 311. The specification of the ‘571 patent taught that the thermoplastic products of the invention could be used to produce thermosetting products by curing the thermoplastic products with agents, such as, inter alia, polyisocyanates. The specification disclosed further that these thermoplastic products and/or thermosetting products were useful as foundry sand binders.

But, the district court failed to consider what these thermoplastic products were disclosed to comprise. The ‘571 specification disclosed that these thermoplastic products were polymeric esters characterized in that:

of the member of the (1) a major portion of the moiety phosphorus family has the formula:

in which the unsatisfied bonds are attached to aryl nuclei of the same phenolic condensate, and in which M is an atom of the phosphorus family, Y is oxygen or sulfur, and X is halogen, hydroxyl, mercapto, hydrocarbyl, hydrocarbyloxy, halogen-substituted hydrocarbyl, halogen-substituted hydrocarbyloxy, or an aryloxy radical of the same phenolic condensate to which M is attached;

(2) at least 60 percent of the phenol-aldehyde or phenol-ketone condensate has o,o1-alkylidene linkages, and;

(3) the phenolic condensate has an average number of aryl nuclei per molecule in the range of 2.2 to 8.

Thus, the ‘571 specification taught that the thermoplastic product was more than a phenolic condensate. It was a phenolic condensate wherein a phosphorus-containing moiety had its unsatisfied bonds attached to the aryl nuclei of the same phenolic condensate. Even assuming arguendo that one skilled in the art might readily have sensed that the Pep resin of the ‘797 patent might have been used as the phenolic condensate called for by the Shepard patent, Shepard taught only that it was the polymeric ester or thermoplastic product, i.e., the phenolic condensate in combination with the bonded phosphorus-containing moiety, that could be reacted with polyisocyanates or tertiary amines to produce thermosetting products having utility in foundry sand binders. But, the teachings of Shepard did not disclose to one skilled in the art whether the phenolic condensate, by itself, would have had utility as a thermosetting product.

The district court found that one of ordinary skill in the art would have looked at the Rothrock patent, MEGSON, and MARTIN, analyzed their teachings in light of the Shepard and British patents, 40 and concluded that a polybenzylic ether resin could have been plugged into Shepard to produce a phenolic urethane binder. Based upon our holdings in Section “A. CLAIM 10 OF THE ‘797 PATENT — PEP RESIN”, supra, this conclusion is erroneous as a matter of law. In Section A we held that the Roth-rock patent, MEGSON, and MARTIN, considered singly would not have led to the conclusion that the Pep resin as claimed in claim 10 of the ‘797 patent would have been obvious, and that there was no basis for combining the teachings of these references, such that one skilled in the art would not have had knowledge of the Pep resin as of the critical period. Therefore, there was no proper basis for the district court to conclude that one skilled in the art, even with knowledge of the teaching of the Shepard patent with respect to the utility of phenolic urethanes as foundry binders, would have had knowledge of a phenolic resin substantially similar to the phenolic resin as claimed in claim 10 of the ‘797 patent. Therefore, it was erroneous to conclude that the Pep resin of the ‘797 patent could have been substituted into the Shepard patent for use in producing a phenolic urethane foundry binder as taught in the Shepard patent.

C.3 Combining Prior Art with the Shepard Patents

Moreover, assuming for the sake of argument that the Rothrock patent, MEG-SON, and MARTIN would have led one skilled in the art to a phenolic resin substantially similar to the phenolic resin as claimed in claim 10 of the ‘797 patent, the ‘571 patent contained relatively little in the way of positive suggestion or inference which would have led one skilled in the art to combine the teachings of these references with the Shepard patent. ACS Hospital Systems, 732 F.2d at 1577, 221 USPQ at 933. The ‘571 patent taught two methods for the preparation of phenolic condensates having a high percentage of ortho-ortho alkylidene linkages.

The preferred phenolic condensate was prepared by reacting an excess of phenol with formaldehyde, i.e., a phenol/formaldehyde ratio greater than 1, in the presence of an inorganic alkali catalyst. In contrast, the Rothrock patent taught a formaldehyde/phenol ratio of 1:1, or a range of 1:63 to 1:1. Further, the Rothrock patent taught that the resins thereof could not be produced in the presence of alkali catalysts.

The alternative process for producing the phenolic condensates described in the ‘571 patent involved reacting an aldehyde with a phenol in the presence of an acid catalyst. The ratio of formaldehyde/phenol was described as being in the range of 0.5 to 1.0, with the preferred range being 0.7 to 0.9. Thus, while the ‘571 patent taught a formaldehyde/phenol ratio range wherein the upper bound minimally overlapped the lower bound of the Rothrock patent formaldehyde/phenol ratio, the preferred range taught in the ‘571 patent diverged away from the ratio as taught in Rothrock. The ‘571 patent also taught that the acid catalyst could be hydrochloric, sulfuric or oxalic acids, and was silent as to the need for a solvent to effect condensation. The Roth-rock patent, in contrast, taught that condensation occurred in the presence of both a mild acid catalyst, such as zinc acetate, boric acid, or copper acetate, and a completely volatile, non-gum-forming solvent, and further that resins could not be produced in the presence of strongly acidic catalysts, such as hydrochloric acid. Since the ‘571 patent appeared to have suggested stronger acid catalysts than those usable in the Rothrock patent, and was silent as to the use of a solvent, an uncertainty would have arisen as to whether the teachings of the Rothrock patent could have properly been combined with the teachings of the Shepard patent.

A further point is that the ‘571 patent taught that the phenolic condensate of the novel ester was one having at least 60 percent ortho-ortho alkylidene linkages. The specification disclosed that the term alkylidene expressed the structural relationship of the substituted methylene residues of the aldehyde to the phenolic nuclei of the phenolic condensates and that the term was intended to be generic to all such substituted methylene groups defined within the scope of the invention, and further taught that the phenolic condensates most useful in the invention were characterized by R2-C-R2 linkages, wherein R2 could be independently selected from the group consisting of hydrogen, a hydrocarbon radical, and a halogen-substituted hydrocarbon radical. This teaching would have seemed to preclude the teachings of the Rothrock patent being combined with the teachings of the ‘571 patent inasmuch as the phenolic resin of the Rothrock patent which would have been substantially similar to the resin as claimed in claim 10 of the ‘797 patent would have had a majority of ether linkages, 41 not alkylidene linkages.

D. SECONDARY CONSIDERATIONS

The district court stated that it had considered relevant 42 secondary considerations prior to reaching its conclusion that the subject matter of the claims in issue of the ‘392 and ‘579 patents would have been obvi ous. Thus, the district court seemingly recognized the holdings of this court vis-avis secondary considerations, to wit, that all relevant evidence going to the issue of obviousness/nonobviousness, which includes properly presented evidence on secondary considerations, must' have been considered prior to reaching a conclusion on obviousness/nonobviousness. Fromson, 755 F.2d at 1556-57, 225 USPQ at 32; W.L. Gore, 721 F.2d at 1555, 220 USPQ at 314.

The district court, however, also averred that the law was well established that commercial success alone, or considered in combination with other secondary considerations, is insufficient to establish patentability where primary indicia of patentability was lacking. 43 Just as it is legal error for a district court to fail to consider relevant evidence going to secondary considerations, Lindemann Maschinenfabrik, 730 F.2d at 1461, 221 USPQ at 488, it may be legal error for a district court to presuppose that all evidence relating to secondary considerations, when considered with the other Graham 44 indicia relating to the obviousness/nonobviousness issue, cannot be of sufficient probative value to elevate the subject matter of the claimed invention to the level of patentable invention. Fromson, 755 F.2d at 1556-57, 225 USPQ at 32; Union Carbide, 724 F.2d at 1573, 220 USPQ at 589 (this court reviews the issue of obviousness as one of law on which it must exercise independent judgment — we must be convinced not only that the decision maker engaged in faulty analysis in applying the law to the facts, but also that a correct application of the law to those facts would bring a different result).

The objective evidence of secondary considerations may in any given case be entitled to more or less weight, depending upon its nature and its relationship to the merits of the invention. Stratoflex, 713 F.2d at 1539, 218 USPQ at 879. Secondary considerations may be the most pertinent, probative, and revealing evidence available to the decision maker in reaching a conclusion on the obviousness/nonobviousness issue. W.L. Gore, 721 F.2d at 1555, 220 USPQ at 314.

While it is incumbent upon the decision maker to recognize that evidence of secondary considerations need not be necessarily conclusive on the obviousness/nonobviousness issue, Fromson, 755 F.2d at 1557, 225 USPQ at 32, the decision maker must also bear in mind that, under certain circumstances, the evidence of secondary considerations may be particularly strong and entitled to such weight that it may be decisive. For example, in Simmons Fastener Corp. v. Illinois Tool Works, Inc., 739 F.2d 1573, 1575-76, 222 USPQ 744, 747 (Fed.Cir.1984), the trial court concluded that the claimed invention would have been obvious in light of the teachings of the prior art. But, the trial court had failed to consider the evidence going to secondary considerations in arriving at this conclusion. Holding that the trial court erred as a matter of law by failing to consider the evidence of secondary considerations prior to arriving at its legal conclusion, this court, after considering the teachings of the prior art and the secondary considerations, reversed the decision of the district court.

While the district court in this case found the evidence of commercial success of the chemical sand binders claimed in the ‘392 and ‘579 patents impressive, the court found countervailing considerations in the evidence that: (1) after the Milwaukee litigation, see supra note 5, one Ashland licensee had effected a downward renegotiation of its royalty payment after initiating a declaratory judgment action against Ash-land on the patent claims here in issue; (2) another Ashland licensee went out of the business approximately one year after the grant of the license; (3) industry recognition of the chemical sand binders as claimed in the ‘392 and ‘579 patents was directed more towards the marketing of these products rather than the invention thereof; and (4) Dr. Robins, the inventor listed in the ‘797, ‘392, and the ‘579 patents, had not received any recognition from the industry and only a small monetary consideration from Ashland for his role in developing these inventions. On the record before this court, we cannot say that these factual findings made by the district court were clearly erroneous. See supra note 13.

The record also reveals, however, that Ashland proferred other evidence of secondary considerations, to wit: (1) affidavits from its industrial customers attesting to the long-felt need satisfied by these chemical sand binders produced by Ash-land; (2) the unexpected results, i.e., the “S” or “Z” cure curve, achieved by the ‘392 and ‘579 foundry mixes; and (3) the alleged copying by Delta of these chemical sand binders. The opinion rendered by the district court did not discuss these secondary considerations, see also supra note 42, and they apparently were not accorded any probative value or entered into the final calculus on the issue of obviousness/nonobviousness. This was error as a matter of law. Lindemann Maschinenfabrik, 730 F.2d at 1461, 221 USPQ at 488.

Where the evidence of record is unchallenged as to secondary considerations ignored by the decision maker, this court may, as a matter of law, consider this objective evidence in reviewing the ultimate conclusion of obviousness/nonobviousness entered by the trial court. Id., 730 F.2d at 1461, 221 USPQ at 488. However, where this evidence of record is controverted, as it is in this case, this court will normally remand to the district court for its initial consideration of this evidence. Under the circumstances of this case, however, where we have held that the prior art of record is insufficient to support the legal conclusion of obviousness rendered against the ‘392 and ‘579 patents, remand for the district court’s consideration of the other evidence going to secondary considerations is not necessary. Nor is it necessary for this court to determine whether a proper consideration of this evidence would have resulted in a different conclusion as to the obviousness of the ‘392 and ‘579 patents by the district court. Union Carbide, 724 F.2d at 1573, 220 USPQ at 589.

E. CONCLUSION

We have held that the district court committed reversible error in combining the teachings of the Rothrock patent, MEG-SON, and MARTIN to reach the conclusion that the subject matter of claim 10 of the ‘797 patent would have been obvious, and further, that these references, considered individually, would not have supported a conclusion that the subject matter of claim 10 would have been obvious.

We have further held that the district court committed reversible error in combining the teachings of the Rothrock patent, the Japanese patent, and Fraser to conclude that the invention of process claims 1, 2, and 7 of the ‘797 patent would have been obvious, and that these references, considered singly, would not have supported a conclusion that the invention of these process claims would have been obvious.

We have also held that, inasmuch as one of ordinary skill in the art would not have had knowledge of the Pep resin as claimed in claim 10 of the ‘797 patent, there was no basis to substitute this resin into the Shepard patent and conclude that the foundry binder inventions of the ‘392 and ‘579 patents would have been obvious. Further, as a matter of law there was no basis in either the Shepard patent or the other prior art relied upon by the district court which would have led one skilled in the art to combine the teachings of this prior art with the teachings of the Shepard patent.

Finally, we have held that the district court erred as a matter of law in failing to consider all evidence going to the secondary considerations. And further, that the district court erred by failing to determine if there was the requisite nexus between the proferred evidence of secondary considerations and the merits of the claimed inventions of the ‘392 and ‘579 patents.

Accordingly, the decision of the district court that claims 1, 2, 7 and 10 of the ‘797 patent, claims 14 and 19 of the ‘579 patent,

and claim 17 of the ‘392 patent are invalid is reversed.

The case is remanded for consideration of the infringement issue.

REVERSED AND REMANDED

APPENDIX

MEGSON, PHENOLIC CHEMISTRY, 29 (Academic Press Inc. 1958)

DETAILED SURVEY

with a dimethylene ether. It is believed, however, that methylene-bridged phenols are important constituents of resins resulting from the hardening of the alcohols, and if these are formed from the benzyl ethers, then elimination of formaldehyde should be much greater than that actually observed in the second stage of heating.

It appears, therefore, that much of the formaldehyde must recombine, and three methods are mentioned by which it can do this. The first involves reaction with nuclear hydrogens to give cross-linked methylene derivatives (A); the second involves reaction with phenolic hydroxyls to yield cross-linked ethers (B); the third involves reaction with methylene groups (when formed) to give cross-linked compounds of a different type (C).

All these reactions would explain the low evolution of formaldehyde and the high evolution of water during the second stage of hardening.

4 . The left hand phenol depicted supra has been labeled so that the ortho (2, 6), meta (3, 5) and para (4) positions are identifiable. CH2-0-CH2 identifies a benzylic ether bridge, CH2 identifies a methylene bridge, OH is the chemical designation for a hydroxyl group, and CH2OH is the chemical designation for a methylol group.

5 . In prior litigation between Ashland and Delta, claims 1, 13, 15, 16 and 18 of the '579 patent were invalidated for obviousness. See Ashland Oil, Inc. v. Delta Oil Products, 212 USPQ 508 (E.D.Wis.1981), rev’d in part, 685 F.2d 175 (7th Cir.1982), cert. denied, 460 U.S. 1081, 103 S.Ct. 1769, 76 L.Ed.2d 343 (1983). None of these prior invalidated claims involved the Pep resin of claim 10 of the ‘797 patent.

6 . Claim 14 incorporates the binder composition of claim 6, claim 6 in turn being dependent upon claim 1, the broad binder composition claim of the ‘579 patent.

7 . Claim 19 has been redrafted in the format of claim 15, with the addition of the relevant portion of claim 6, to facilitate review. Claim 19 is dependent upon claim 15, and incorporates the phenolic resin of claim 6, i.e., the Pep resin of claim 10 of the '797 patent.

8 . In the prior action between Ashland and Delta, see supra note 5, the district court had held all claims of the ‘392 patent invalid for obviousness. The Seventh Circuit reversed in part, holding that only claims placed in issue, in that case claims 1 and 16, were subject to invalidation. The invalidated claims were not directed to the use of the Pep resin.

9 . Claim 17 incorporates the resin composition of claim 1, with the additional limitation of claim 7 wherein the phenolic resin component is a benzylic ether resin. The relevant portion of claim 7 has been incorporated into claim 17 to facilitate review.

10 . Ashland has argued that the district court made few factual findings, that the court was merely presenting the arguments of the parties. While the format of the court’s opinion lends some credence to this argument, we note that the court properly recognized that it was required to make factual' determinations on the scope and content of the prior art and the differences between the prior art and the Robins patents. See Graham v. John Deere Co., 383 U.S. 1, 17, 86 S.Ct. 684, 693, 15 L.Ed.2d 545, 148 USPQ 459, 467 (1966). Thus, the court’s exposition appears to be the court’s short-hand method of examining the prior art and the differences between the prior art and the patents in issue. Consequently, we review these statements by the court as its factual findings.

11 . During pretrial discovery Delta had identified Dr. Raymond Wentland as its only expert witness. Dr. Wentland was not called to testify during trial. Instead, Delta chose to rely primarily on the testimony of Mr. Kopac, who although possessing the qualifications of one of ordinary skill in the art, had not been qualified to testify as an expert witness.

12 . The application on the British patent was filed in London on 5 November 1963 and published on 2 June 1966. This application was based upon a prior United States application, Serial No. 241,131, filed 30 November 1962. The application on the 797 patent was filed on 14 March 1962, the application on the ‘392 patent had a continuation-in-part filing date of 14 March 1962 and the application on the ‘579 patent had a continuation-in-part filing date of 1 August 1966.

13 . A finding is clearly erroneous when the appellate court, after reviewing the entire record, is left with the definite and firm conviction that a mistake has been made, even though there is some evidence in the record to support such a finding. United States v. U.S. Gypsum Co., 333 U.S. 364, 395, 68 S.Ct. 525, 541, 92 L.Ed. 746, 76 USPQ 430, 444 (1948); Lindemann Maschinenfabrik GMBH v. American Hoist & Derrick Co., 730 F.2d 1452, 1458, 221 USPQ 481, 485 (Fed.Cir.1984).

Further guidance as to the role of appellate review under the clearly erroneous standard has been provided by the Supreme Court in Anderson v. City of Bessemer City, N.C., - U.S. -, 105 S.Ct. 1504, 1511, 1512, 84 L.Ed.2d 518 (1985), wherein the Court stated that:

This standard plainly does not entitle a reviewing court to reverse the finding of the trier of fact simply because it is convinced that it would have decided the case differently-

******

If the district court’s account of the evidence is plausible in light of the record reviewed in its entirety, the court of appeals may not reverse it even though convinced that had it been sitting as the trier of fact, it would have weighed the evidence differently. Where there are two permissible views of the evidence, the factfinder’s choice between them cannot be clearly erroneous.

14 . One possible interpretation of the district court’s holding is that the district court evaluated the facts established by Ashland’s rebuttal evidence solely on the basis of its ability to overcome or knockdown the legal inference of obviousness, i.e., Delta's facts establishing a prima facie case for obviousness. This approach was rejected in In re Rinehart, 531 F.2d 1048, 1052, 189 USPQ 143, 147 (CCPA 1976), wherein the court stated that facts established by rebuttal evidence must be evaluated along with the facts on which the earlier conclusion of obviousness, i.e., the prima facie case, was reached, not against the conclusion itself. See Ralston Purina Co. v. Far-Mar-Co, Inc., 772 F.2d 1570, 1573 (Fed.Cir.1985). The ultimate conclusion on obviousness must rest upon an evaluation of all facts that have been established by clear and-convincing evidence.

15 . Claim interpretation is a legal matter subject to review free of the clearly erroneous standard applicable to fact findings. Raytheon Co. v. Roper Corp., 724 F.2d 951, 956, 220 USPQ 592, 596 (Fed.Cir.1983), cert. denied, - U.S. -, 105 S.Ct. 127, 83 L.Ed.2d 69 (1984).

16 . The specification of the ‘797 patent teaches the formation of novel phenolic compositions characterized as a phenol-formaldehyde adduct or modified resole resin, which comprises one-ring dimethylol phenols and two-ring benzylic ethers and methylene-bridged phenols. The specification, however, also discloses that the novel phenolic resin of claim 10 is a higher molecular weight product formed by the condensation of the phenol-formaldehyde adduct. Further, the specification teaches that the “formation of the [phenol-formaldeyde] Adduct ... is a precursor to the novel phenolic resins [as claimed in claim 10]."

While the specification is silent as to the actual chemical composition of phenol aldehyde reaction products produced by practicing the process disclosed in the ‘797 patent, Dr. Robins, the inventor, testified that the phenol aldehyde reaction product resulting from this process would comprise approximately 5-10% phenol formaldehyde resin as claimed in claim 10, with the remainder being monomers such as dimethylol phenols, benzylic ethers and methylene-bridged phenols.

The actual chemical composition of the reaction products produced by the ‘797 process, however, is irrelevant to a § 103 determination with respect to the product claimed in claim 10. The metes and bounds of claim 10 define the relevant product for the § 103 determination. Lear Siegler, 733 F.2d at 890, 221 USPQ at 1033.

17 . The Rothrock patent taught only that the phenol-formaldehyde resin produced by the process disclosed therein was cured or heat-hardened by the addition of thermal energy to the resin. There was no disclosure that the resin was curable at room temperatures by the addition of an acid catalyst. The specification of the '797 patent, in contrast, teaches that the Pep resin is curable at room temperatures by the addition of an acid catalyst, or may be cured by the addition of thermal energy to the resin. W.L. Gore, 721 F.2d at 1550, 220 USPQ at 311 (a reference must have been considered in its entirety, for disclosures which taught way from the invention as well as disclosures which directed one skilled in the art towards the claimed subject matter).

18 . See supra note 16 and accompanying text.

19 . The presence of water results in reaction products which cannot be cured to mechanically strong resins by the use of acidic reagents at room temperature. The presence of water affects not only the activity of the catalyst, but also the structure of the product formed, for example, permitting substitution at the para position.

20 . This statement will be more fully addressed in Section "B. THE PROCESS CLAIMS OF THE '797 PATENT", infra.

21 . The district court’s factual finding that the process of the Rothrock patent produced claim 10 material is not plausible in light of the entire record. Anderson v. City of Bessemer City, N.C., - U.S. -, 105 S.Ct. 1504, 1512, 84 L.Ed.2d 518 (1985). There is no objective evidence disclosed in the Rothrock patent as to the nature of the product formed by the process disclosed therein. Delta did not proffer any objective evidence as to the type of phenolic resin produced by the Rothrock process. Mr. Kopac's opinion testimony as to what the Rothrock process produced is not substantiated by any objective evidence, and therefore can have probative value only as conjecture of one skilled in the art. See FED.R.EVID. 701. The evidence of record does not support the view that the Rothrock process produced claim 10 material. Id.

22 . While we have reservations about this interpretation of claim 10, we will not, on the record before us, say it is erroneous as a matter of law. ACS Hospital Systems, 732 F.2d at 1577, 221 USPQ at 932. Under the court’s interpretation when n is equal to zero, i.e., no methylene bridges, the ratio of m to n is undefined. But, whenever methylene bridges are present, i.e., n is a real number unequal to zero, the ratio of m to n is a finite real number. It is questionable whether the ratio of m to n would be a finite real number in all circumstances except one, where it is undefined.

The specification of the '797 patent indicates that the majority of linkages between the phenol rings will be benzylic ether such that the remainder of the linkages will be methylene. The specification also teaches that, at the temperatures at which condensation occurs to form the claim 10 material, some methylene linkages are formed.

Dr. Robins, however, gave testimony that n could equal zero, but this testimony was elicited in connection with a discussion as to whether the dimethylol phenols, benzylic ethers and methylene-bridged phenols are within the scope of claim 10. Dr. Robins also testified that the scope of claim 10 did not encompass the one or two phenol-ring adducts, but only polymers having three or more phenol rings. Dr. Conley, Ashland’s expert, also testified that n could be equal to zero. Anderson v. City of Bessemer City, N.C., - U.S. -, 105 S.Ct. 1504, 1512, 84 L.Ed.2d 518 (1985) (where there are two permissible views of the evidence, the fact finder’s choice between them cannot be clearly erroneous). The court’s interpretation of this claim appears to have been predicated on the factual testimony proffered by Ashland’s witnesses.

23 . Dr. Conley testified that Megson’s original work on this subject taught that the R group was an alkyl or aryl constituent. He also testified that the drawing cited by the district court, as well as the other two drawings relating to the cross-linking mechanism, were postulated or hypothetical structures to explain the cross-linking mechanism.

24 . To properly combine references A and B to reach the conclusion that the subject matter of a patent would have been obvious, case law requires that there must have been some teaching, suggestion, or inference in either reference A or B, or both, or knowledge generally available to one of ordinary skill in the relevant art, which would have led one skilled in the art to combine the relevant teachings of references A and B. See, e.g., ACS Hospital Systems, 732 F.2d at 1577, 221 USPQ at 933; W.L. Gore, 721 F.2d at 1551, 220 USPQ at 311; In re Sernaker, 702 F.2d 989, 994, 217 USPQ 1, 5 (Fed.Cir.1983). The decision maker's determination as to what objective evidence in reference A or B, or both, or generally available to one of ordinary skill in the relevant art, is of the nature of a factual finding.

The decision maker, however, after making findings as to the objective evidence, must subjectively analyze these factual findings to determine whether the teachings of references A and B could have been combined. Thus, the ultimate determination as to whether references could have been combined is a legal conclusion.

Where the district court fails to set forth the objective bases for its conclusion that references could have been combined, this court will review the determination as a matter of law.

25 . The Rothrock patent taught that formaldehyde and phenol were condensed in the presence of a mild acid catalyst and a completely volatile, non-gum-forming solvent selected from the class of monohydric aliphatic alcohols and mononuclear aromatic hydrocarbons. Ashland presented testimony, and argued before the district court and this court, that the use of butyl alcohol as the Rothrock solvent produced a butyl alcohol modified resin which was not a phenolic resin. The district court made a finding of fact to this effect. Ashland, however, has ignored the teaching of Rothrock that mononuclear aromatic hydrocarbons could have been used as the solvent, and that Example V taught the use of toluene, an aromatic hydrocarbon.

Although the process claims of the ‘797 patent do not have a specific limitation directed to a solvent, the process claims do require that the phenol and aldehyde be "in the liquid phase.” The '797 specification teaches that “[a]lthough it is not necessary to have an inert diluent present, it is generally preferred to conduct the reaction in the presence of one." The ‘797 specification further teaches that these solvents, when used, are non-polar organic solvents such as aliphatic, cycloaliphatic, aromatic and halogenated hydrocarbons. Toluene is set forth as a specific example of such a solvent.

26 . Contrary to the district court’s statement, Example VII of the Rothrock patent did not teach the removal of water. The only statement in Example VII with respect to water was that after the completion of the process, a clear solution was obtained "with traces of water on the sides of the flask."

27 . Mr. Kopac testified that an air condensor may be used, and depending upon its length, will condense certain vapors formed during the reaction in a manner so as to reintroduce these condensed vapors back into the reaction zone while unwanted vapors are transported outside of the reaction zone.

28 . Dr. Robins uncontroverted testimony was that reflux meant that water was being distilled from the reaction mixture, condensed, and returned to the reaction mixture.

29 . But see supra note 26.

30 . At one point in the specification the Japanese Patent taught that the alkaline catalyst could be selected from a group consisting of alkaline earth metal hydroxides, magnesium hydroxide, and alkoxyphenoxymethanes — the process disclosed therein was claimed in this manner— while at another point in the specification it was taught that both an alkaline catalyst and the alkoxyphenoxymethanes were used to produce the initial methylolphenol products.

31 . Delta argued that the Japanese Patent emphasized the importance of removing water in the process to produce the ortho-ortho orientation of the resultant phenolic resin. The Japanese Patent, however, taught that "in order to provide the ortho orientation to the resin, it is, of course, necessary to select the catalyst having the ortho orientation effect____” Although it was taught that both the pH and water content of the reaction system affected the final product, the Japanese Patent clearly indicated that the alkaline catalyst was the dominant factor in producing the ortho-ortho orientation. Since the Japanese Patent disclosed that it was the anhydrous conditions and the alkaline catalyst in combination which produced the ortho-ortho orientation in the resultant product, there was no suggestion or inference therein to one skilled in the art that the removal of water in the Rothrock process, which cannot utilize an alkaline catalyst, would have been advantageous in effecting an ortho-ortho orientation in the resulting resin. W.L. Gore, 721 F.2d at 1551, 220 USPQ at 311.

32 . See supra note 30.

33 . It appears that the district court used process claim 1 of the ‘797 patent as á blueprint, and abstracted individual teachings from the Roth-rock patent, the Japanese Patent, and the Fraser reference to create the process of claim 1, without due consideration for teachings in these references that would have led one skilled in the art to find it improper to combine these references. W.L. Gore, 721 F.2d at 1552, 220 USPQ at 312.

34 . The phenolic resin of claim 10 of the ‘797 patent is defined such that m and n are numbers the sum of which is at least 2 and the ratio of m to n is greater than 1. In contrast, the benzylic ether resin as claimed in the foundry mix of claim 17 of the '392 patent is defined such that m and n are numbers the sum of which is at least 2, and wherein m is at least 1.

35 . The specification described these highly reactive divalent materials as “[a]ny diisocyanate in which the isocyanate groups are the sole reactive groups...."

36 . Keeping in mind the admonishment that a disclosure should have been read in its entirety for all it divulged, W.L. Gore, 721 F.2d at 1550, 220 USPQ at 311, we have reservations about this finding of the district court. The specification of the British Patent stated that "[a]s far as [the inventors] are aware, soluble, fusible polymers which are the products of the reaction of a novolac resin and a highly reactive divalent material capable of condensing with phenolic hydroxyl groups have not heretofore been performed.” The specification further stated that "the reaction of a conventional novolac polymer with such divalent reactants results in the production of the expected insoluble, infusible thermoset materials." From these disclosures in the British patent we would not say it was a foregone determination that the use of phenolic urethanes, prepared by the reaction of a resin with an isocyanate, in foundry mixes was known to those of ordinary skill in the foundry art as of the critical date.

37 . On appeal, both Ashland and Delta have pointed out that this statement by the district court is technically incorrect in part. While polyisocyanates do react with hydroxyl groups, they do not react with benzylic ether bridges.

38 . The district court made a finding that the British Patent was not prior art. Therefore, the district court could only have utilized the British Patent in its "analysis” to the extent that the British Patent showed the general level of skill in the art as of the critical date. Cf. In re Farrenkopf, 713 F.2d 714, 219 USPQ 1 (Fed.Cir.1983). But see supra note 36.

39 . Based upon the discussion infra, it is not necessary to review the factual findings made by the district court with respect to tertiary amines and curing catalysts having a pKb value in the range of about 7 to about 11.

40 . See supra note 38.

41 . Claim 10 of the 797 patent requires that the phenolic resin claimed therein must have a ratio of m to n of greater than 1. To satisfy this constraint, m must be greater than n such that the number of benzylic ether bridges is always greater than 50 percent. The benzylic ether bridges are defined by the chemical formula CH2-O-CH2, and as such contain oxygen. In contrast, the specification of the ‘571 patent teaches that the preferred phenolic condensate must have at least 60 percent ortho-ortho alkylidene bridges, and that these bridges do not contain oxygen.

42 . Case law requires that a nexus be established between the merits of the claimed invention and the evidence proferred on secondary considerations, if the evidence on secondary considerations is to be given substantial weight in the calculus of obviousness/nonobviousness. Simmons Fastener Corp. v. Illinois Tool Works, Inc., 739 F.2d 1573, 1575, 222 USPQ 744, 746 (Fed.Cir.1984), cer t. denied, - U.S. -, 105 S.Ct. 2138, 85 L.Ed.2d 496 (1985); Stratoflex, 713 F.2d at 1539, 218 USPQ at 879.

Ashland argued that the merits of its phenolic urethane foundry binder mixes — Isocure and Pep Set — were due to the characteristics of the phenol-formaldehyde resin, i.e., the Pep resin, substantially as claimed in claim 10 of the 797 patent. See supra note 34 and accompanying text. The district court did not make any explicit finding as to the nexus between the merits of the claimed invention and the proferred secondary considerations.

The district court at one point in the opinion stated that the Pep resin is an ingredient in Ashland’s Isocure and Pep Set foundry binder mixes. The court stated that the main advantage of these products was the speed and timing of the cure, i.e., the "S” or "Z" cure curve, which increases foundry productivity.

Delta had argued before the district court that Ashland’s Isocure and Pep Set foundry products were not covered by the ‘579 and '392 patents, respectively, because the Isocure and Pep Set foundry products have only an average of 2.5 phenol rings per molecule, whereas the claims of the ‘579 and ‘392 patents require an average of 3 or more phenol rings per molecule. The district court stated that since it had found the patents in suit to be invalid for obviousness under § 103, there was no need to determine whether the patents in suit covered Ashland's products. This was error as a matter of law. For secondary considerations to have probative value, the decision maker must determine whether there is a nexus between the merits of the claimed invention and the secondary considerations. Simmons Fastener Corp., 739 F.2d at 1575, 222 USPQ at 746; Stratoflex, 713 F.2d at 1539, 218 USPQ at 879. Under the circumstances of this case, Ashland’s proferred evidence of secondary considerations cannot properly be considered in reaching a conclusion on obviousness/nonobviousness unless the decision maker first determines that these secondary considerations are relevant to the subject matter as claimed. For example, had the decision maker made this determination in this case, and determined that the Isocure and Pep Set products were not covered by the ‘579 and ‘392 patents, respectively, then the secondary considerations would not have had any relevance to the obviousness/nonobviousness determination.

43 . For support for this proposition the district court cited to: Sakraida v. Ag Pro, Inc., 425 U.S. 273, 283, 96 S.Ct. 1532, 1538, 47 L.Ed.2d 784, 189 USPQ 449, 453, reh’g denied, 426 U.S. 955, 96 S.Ct. 3182, 49 L.Ed.2d 1194 (1976) ("[Producing a desired result in a cheaper and faster way, and enjoying commercial success ‘without invention will not make patentability.’ ”); Great Atlantic & Pacific Tea Co. v. Supermarket Equipment Corp., 340 U.S. 147, 153, 71 S.Ct. 127, 130, 95 L.Ed. 162, 87 USPQ 303, 306 (1950), reh’g denied, 340 U.S. 918, 71 S.Ct. 349, 95 L.Ed. 663 (1951) ("But commercial success without invention will not make patentability.”); and Eltra Corp. v. Basic, Inc., 599 F.2d 745, 756, 202 USPQ 630, 640 (6th Cir.), cert. denied, 444 U.S. 942, 100 S.Ct. 297, 62 L.Ed.2d 308 (1979) ("Of course, commercial success and satisfaction of long-felt needs are alone not sufficient to establish that the product is the result of invention.”).

44 . Graham v. John Deere Co., 383 U.S. 1, 17, 86 S.Ct. 684, 693, 15 L.Ed.2d 525, 148 USPQ 459, 467 (1966).