In re Carleton

MILLER, Judge.

This is an appeal from the decision of the Patent and Trademark Office (“PTO”) Board of Appeals (“board”) sustaining the rejection under 35 U.S.C. § 103 of claims 9-13 and 15-16. We reverse.

BACKGROUND

The Invention

Appellant’s application¹ discloses a process for the substantially quantitative production of hydroquinone. Claims 9-10 and 15² are directed to the reaction of g-isopro-penylphenol and hydrogen peroxide in an inert solvent such as glacial acetic acid and in the presence of a catalytic amount of a strong mineral acid to produce hydroqui-none and acetone (see reaction step (3) below). Claims 11-13 and 16 ³ are directed to the semicontinuous production of hydroqui-none through the production of p-isoprope-nylphenol from phenol and acetone, according to the following reaction sequence:

The Rejections and Appellant’s Response

The examiner rejected claims 9-10 and 15 as unpatentable over a patent to Robert H. Saunders (“Saunders”),⁴ which discloses a process for the preparation of various phenols from , -unsaturated alkyl-substitut-ed aryl⁵ compounds. Hydroquinone is a phenol.⁶ His reaction occurs in a liquid solvent, such as acetic acid, with a strong acid catalyst and, preferably, with “low cost” hydrogen peroxide as the oxidizing agent. In example 2, -methylstyrene, which the examiner said is the closest in structure to the starting material of appellant’s claimed invention (p-isopropenylphe-nol), is reacted with t-butylhydroperoxide to produce phenol in about an 80% yield:

Although the production of hydroquinone as such is mentioned by Saunders, the only suggested process begins with tetramethyl-p-xylylene dialcohol as the starting material.⁷

The examiner’s rejection of claims 11-13 and 16 was based on Saunders in view of a patent to John L. Jones (“Jones”).⁸ Jones was cited for the disclosure of a method for the preparation of phenols having unsaturated substituents (such as p-isopropenyl-phenol — the starting material of claims 9-10 and 15) via a two-step reaction. In the first step, a phenol is condensed in the presence of an acid catalyst with a ketone to produce an intermediate “condensation product,” which is then treated with a strong alkali metal base to neutralize the acidic catalyst. The mixture undergoes pyrolysis producing the desired phenolic compound (with an unsubstituted side chain), phenol, and a polymeric residue. On appeal, appellant has not argued that the first two steps of his reaction sequence, in which p-isopropenylphenol is produced, would not have been obvious in light of the PTO’s citation to Jones.

In response to the rejections, appellant submitted several affidavits (including two by him) under 37 CFR 1.132. One by Sheng-Hong A. Dai described two tests. In the first test (“Dai I”), the procedure of example 2 of the Saunders patent was followed “exactly”⁹ (the starting material being -methylstyrene), except that hydrogen peroxide (instead of t-butylhydroperox-ide) was used as the oxidizing agent. A yield of only 15.8% phenol was reported; the major product was believed to be dimers of the -methylstyrene.¹⁰ In the second test (“Dai II”), the “exact” reaction conditions of example 2 of Saunders were followed except that p-isopropenylphenol (the starting material of appellant’s process) was used. A yield of only 47% hydro-quinone was obtained.

Appellant also directed the examiner’s attention to a published German application,¹¹ which discloses a process for the production of hydroquinone from 1,4-diisopropenylben-zene¹² in the presence of glacial acetic acid as the solvent, either mineral acids or Frie-del-Crafts type compounds as the catalyst, and hydrogen peroxide as the oxidizing agent. The pertinent examples indicate yields of hydroquinone of about 10-35%.¹³

The Decision Below

The board found that the cited references “make out a strong case of prima facie obviousness” against appellant’s claims, because (1) the starting material of the claimed process is encompassed in Saunders’ generic starting materials; (2) the oxidizing agent in the claimed process (hydrogen peroxide) is stated by Saunders to be the preferred peroxide (due to its low cost); and (3) Saunders and Jones disclose all of the other reaction conditions. It further found that the disclosure of hydroquinone production by the Saunders process, “albeit from a different starting material than the one recited in appellant’s claims, would lead one of ordinary skill to expect that the presence of the first formed hydroxyl group [on the benzene ring] would not interfere with the process of forming the second hydroxyl group.” The board said the affidavit evidence was unpersuasive and concluded that the Carleton I and II affidavits failed to follow the prior art closely enough. It particularly criticized the concentration of the reactants and catalysts and the short reaction times.

On reconsideration, the board adhered to its original decision, saying that the affidavits merely indicate that it is possible to operate within the teachings of Saunders without obtaining a good yield. It also said that the Dai II result conflicts with the result in example 9 of appellant’s specification, illustrating that there can be considerable variation in yield from “minor modification in reaction conditions.”

OPINION

We do not agree with the board’s characterization of the showing made by the examiner as a “strong” prima facie case. In a 35 U.S.C. § 103 case, the burden of proof is on the PTO to establish a prima facie case of obviousness, In re Warner, 379 F.2d 1011, 54 CCPA 1628, 154 USPQ 173 (1967), cert. denied, 389 U.S. 1057, 88 S.Ct. 1201, 20 L.Ed.2d 101 (1968), and, once this has been accomplished, the burden of going forward with evidence to rebut that prima facie case is shifted to the applicant. In re Murch, 464 F.2d 1051, 59 CCPA 1277, 175 USPQ 89 (1972); In re Hyson, 453 F.2d 764, 59 CCPA 782, 172 USPQ 399 (1972). Whether a prima facie case is “strong” or “weak” is not material. If the applicant presents rebuttal evidence, the decisionmaker must consider all of the evidence of record (both that supporting and that rebutting the prima facie case) in determining whether the subject matter as a whole would have been obvious.¹⁴ In re Rinehart, 531 F.2d 1048, 189 USPQ 143 (Cust. & Pat.App.1976); In re Lewis, 443 F.2d 389, 58 CCPA 1270, 170 USPQ 84 (CCPA 1971). The correct procedure for considering rebuttal evidence was set forth by this court in In re Rinehart, supra, 189 USPQ at 147, 531 F.2d at 1052:

Though the burden of going forward to rebut the prima facie case remains with the applicant, the question of whether that burden has been successfully carried requires that the entire path to decision be retraced. An earlier decision should not, as it was here, be considered as set in concrete, and applicant’s rebuttal evidence then be evaluated only on its knockdown ability. Analytical fixation on an earlier decision can tend to provide that decision with an undeservedly broadened umbrella effect. Prima facie obviousness is a legal conclusion, not a fact. Facts established by rebuttal evidence must be evaluated along with the facts on which the earlier conclusion was reached, not against the conclusion itself. Though the tribunal must begin anew, a final finding of obviousness may of course be reached, but such finding will rest upon evaluation of all facts in evidence, uninfluenced by any earlier conclusion reached by an earlier board upon a different record.

The similarities between the processes of Saunders and those of appellant are evident, the only significant differences alleged by appellant being: (1) the starting material (Saunders uses -methyl-styrene, whereas appellant uses p-isopropenylphenol),¹⁵ and (2) the oxidizing agent (Saunders uses t-butylhydroperoxide, whereas appellant uses hydrogen peroxide).

Regarding (1), the board noted that the generic description of Saunders’ starting materials (“ , -unsaturáted alkyl-substi-tuted aryl compounds”) encompasses appellant’s starting material since the term “aryl” includes the phenyl radical. Assuming that a hydroxyl group on the benzene ring of Saunders’ starting material (which is the only difference between the starting material used by appellant and that used by Saunders) would not interfere with the reaction process disclosed by Saunders, the board said that it would have been obvious to make hydroquinone (which differs from phenol only in the hydroxyl group at the para-position) from a starting material such as that used by Saunders, in which a hy-droxyl group ¹⁶ is attached at the para-position.

Regarding (2), the board noted that, although example 2 of Saunders uses t-butyl-hydroperoxide as the oxidizing agent, Saunders makes the general observation that hydrogen peroxide is the preferred oxidizing agent “due to its low cost.” It concluded, therefore, that it would have been obvious to one skilled in the art to use hydrogen peroxide as the oxidizing agent in appellant’s process. In view of the foregoing, we are persuaded that the board properly found that the prior art cited by the examiner established a prima facie case of obviousness.¹⁷ However, we are also persuaded that the board erred in failing to give proper weight to appellant’s rebuttal affidavits, which directly attacked the premises on which the board based its determination of obviousness. Cf. In re Lewis, supra.

The only difference between Dai II and example 2 of Saunders is the presence in Dai II’s starting material (the same as appellant’s starting material) of a hydroxyl group on the benzene ring in the para-position.¹⁸ The board assumed that this difference would not interfere with the reaction process disclosed by Saunders. However, the yield of hydroquinone was dramatically less when appellant’s starting material was reacted in Dai II according to Saunders’ process (47%) than the yield of phenol obtained with Saunders’ starting material (80%). Thus, the Dai II result contradicts both the board’s assumption and its conclusion that “one of ordinary skill . . . [would] expect that the presence of . [one] hydroxyl group would not interfere with the process of forming the second hy-droxyl group.” We are satisfied that one skilled in the art, viewing the Dai affidavit, would have concluded that the presence of such a hydroxyl group results in a significantly lower yield of hydroquinone than the amount that would be obtained with the Saunders’ process, and that appellant’s starting material could not be readily substituted in Saunders’ process to achieve similar results.¹⁹

Appellant also relies upon the published German application²⁰ for its teaching that, under the process conditions of the appealed claims, both isopropenyl groups of p-diiso-propenylbenzene are not readily converted to hydroxyl groups to yield hydroquinone.²¹ This teaching further supports appellant’s position and contradicts the board’s assumption that a first formed hydroxyl group would not interfere with the process of forming the second hydroxyl group; moreover, the low yields are in marked contrast to the substantially quantitative yields obtained by appellant.

Finally, appellant has shown by the Dai I affidavit that, despite Saunders’ observation that hydrogen peroxide is the preferred oxidizing agent because of its low cost, it would not have been obvious to one of ordinary skill in the art to use hydrogen peroxide with the starting material of appellant’s process. Dai I demonstrates that when hydrogen peroxide is used in Saunders’ example 2, a yield of only 15.8% phenol is obtained in contrast to the 80% yield in Saunders’ example 2. Accordingly, one skilled in the art would hardly have expected to obtain a substantially quantitative yield if hydrogen peroxide was the oxidizing agent of a different starting material from that used in Saunders’ example 2.

The board’s response to appellant’s rebuttal evidence was:

Appellant’s showing that it is possible to operate within the reference disclosure without obtaining a good yield of the desired product is insufficient to refute the teachings of the reference, since they have presented no evidence of making experiments and adaptations which one of ordinary skill in this art would make as a matter of course if he did not immediately obtain the desired results.

Although in appropriate cases such a response might be persuasive,²² it overlooks that the two variations from the closest prior art in the tests reported by the affidavits (different starting materials and different oxidizing agents) were the very ones that the board relied on in finding a prima facie case of obviousness. As related above, the results of those tests would negate any expectation of one skilled in the art that these variations in the Saunders’ process would result in a substantially quantitative production of hydroquinone. Although there is a vast amount of knowledge about general relationships in the chemical arts, chemistry is still largely empirical, and there is often great difficulty in predicting precisely how a given compound will behave. As the Second Circuit said in Scher-ing Corp. v. Gilbert, 153 F.2d 428, 433, 68 USPQ 84, 89 (2d Cir. 1946):

while analogy is at times useful, organic [as well as inorganic] chemistry is essentially an experimental science and results are often uncertain, unpredictable and unexpected.

Accordingly, we hold that the affidavits, when considered with all of the evidence, are sufficient as a matter of law to rebut the prima facie case of obviousness.

The rejection of claims 9-13 and 15-16 is reversed.

REVERSED.

1 . Serial No. 563,464, filed March 31, 1975, for “Process.”

2 . Claim 9, which is illustrative of this set of claims, reads:

9. A process which comprises reacting p-isopropenylphenol with an at least equimolar amount of hydrogen peroxide in the presence of glacial acetic acid and a catalytic amount of a strong acid selected from the group consisting of sulfuric, phosphoric, p-toluenesul-fonic, benzenesulfonic, methanesulfonic and ethanesulfonic acids, said reaction being carried out at a temperature not greater than 80 °C whereby there is obtained hydroqui-none and acetone.

3 . Claim 12, which is illustrative of this set of claims, reads:

12. A semicontinuous process for the conversion of phenol to hydroquinone which comprises condensing phenol and acetone in the presence of acid to obtain Bisphenol A, subjecting said Bisphenol A to alkaline hydrolysis to yield a mixture of phenol and p-isopropenylphenol subjecting said mixture of phenol and p-isopropenylphenol, without separation, to reaction with at least an equi-molar amount, based on p-isopropenylphe-nol, of hydrogen peroxide in the presence of an inert solvent and a catalytic amount of a strong acid selected from the group consisting of sulfuric, phosphoric, p-toluenesulfonic, benzenesulfonic, methanesulfonic and ethan-esulfonic acids said reaction' being carried out at a temperature not greater than 80 °C, to obtain a mixture of hydroquinone, phenol, and acetone, recovering the hydroquinone therefrom, and recovering the phenol and acetone generated as by-products for re-use as starting materials in a subsequent cycle of the above steps.

4 . U.S. Patent No. 2,644,014, issued June 30, 1953, for “Phenol Production.”

5 . The patent defines the term “aryl” to include the “phenyl” radical.

6 . A phenol is a benzene ring to which a hydroxyl group (-OH) has been attached:

Other substituents (or groups) may be attached to the benzene ring at various points around the ring. Hydroquinone is within the class of phenols because it has an additional hydroxyl group attached to the benzene ring:

7 . The process would be represented as follows:

8 . U.S. patent No. 2,497,503, issued February 14, 1950, for “Preparation of Substituted Phenols.”

9 . A miscalculation in the original Dai affidavit was corrected before the board’s opinion on reconsideration, thereby alleviating the board’s concern in its original opinion that an excessive amount of sulfuric acid was used in Dai I.

10 . Although conditions were somewhat varied from those of the Saunders reference, the first affidavit of appellant Carleton (“Carleton I”) reports the result of a similar experiment. No phenol was detected, and the major products were a dimer of -methylstyrene and 1,1,3-tri-methyl-3-phenylindan.

11 . German patent application No. S33.841 (now patent No. 947,308), filed June 12, 1953, by Societe des Usines Chimiques Rhone-Pou-lenc and published on February 23, 1956.

12 . 1,4-diisopropenylbenzene has the following structure:

13 . The second affidavit of appellant (“Carleton II”) reports an experiment similar to those of the German application in which 1,4-diisopro-penylbenzene was also the starting material. A yield of about 28% of hydroquinone was obtained.

14 . The Solicitor’s brief states that “the issue is whether the decision of the Board of Appeals was clearly erroneous.” However, “obviousness is a legal conclusion based on factual evidence, Graham v. John Deere Co., [383 U.S. 1, 86 S.Ct. 684, 15 L.Ed.2d 545, 148 USPQ 459 (1966)] . . . and not a factual determination.” In re Warner, 54 CCPA 1628, 1634 n. 6, 379 F.2d 1011, 1016 n. 6, 154 USPQ 173, 177 n. 6 (1967). Therefore, the proper issue before us is whether the board erred, as a matter of law, in holding that the claims were properly rejected under 35 U.S.C. § 103. In deciding this issue the court will make “an independent determination as to the legal conclusions and inferences which should be drawn from . [the findings of fact].” See United States v. Mississippi Valley Generating Co., 364 U.S. 520, 526, 81 S.Ct. 294, 297, 5 L.Ed.2d 268 (1961).

15 . The structural difference is that appellant’s starting material has a hydroxyl group on the benzene ring in the para -position to the isopro-penyl group:

16 . Or, alternatively, a group, such as an isopro-penyl group, which can be readily converted to a hydroxyl group by Saunders’ process.

17 . As indicated earlier, appellant has not contended that Jones was improperly applied against claims 11-13 and 16.

18 . This difference of a hydroxyl group results in hydroquinone being the product in Dai II, rather than phenol, which is produced in Saunders’ example 2.

19 . Contrary to the board’s statement that the Dai II result conflicts with the result in example 9 of appellant’s specification, the results are consistent. Under different reaction conditions (e. g., temperature of 105°C in Dal II vs. 30°C in example 9), example 9 reports only a 32% yield of hydroquinone. This substantiates the conclusion that a significantly lower yield is achieved when appellant’s starting material is used in Saunders’ process.

20 . Although the teachings of the published German application were argued before the board, it made no mention of the German application in its opinion.

21 . The low yields of about 10-35% are obtained even though hydrogen peroxide is used as the oxidizing agent. The result of Carleton II (28%) confirms the teaching of the German application.

22 . Cf. In re Weber, 405 F.2d 1403, 56 CCPA900, 160 USPQ 549 (1969); In re Michalek, 162 F.2d 229, 34 CCPA 1124, 74 USPQ 107 (1947).