Joseph Schimmel, Washington, D. C. (Raymond E. Martin, Washington, D. C., of counsel), for the Commissioner of Patents.

Before RICH, Acting Chief Judge, and MARTIN, SMITH, and ALMOND, Associate Judges.

RICH, Acting Chief Judge.

This appeal is from the unanimous decision of the Patent Office Board of Appeals1 affirming the examiner's rejection of process claims 1-3, 5, 7, and 8 in application serial No. 52,211, filed August 26, 1960, for "Process for Removing Carbonyl Sulfide from Normally Gaseous Hydrocarbons." No claim is allowed.

The invention relates to the removal of carbonyl sulfide (COS) from normally gaseous hydrocarbons, obtained from refining and cracking processes and employed in numerous conventional applications. COS is an impurity frequently found in such hydrocarbons and often has a deleterious effect on their use, as, for example, in the polymerization of propylene to polypropylene where the COS, if present even in small amounts, poisons the polymerization catalyst.

Appellant invented a method for lowering the COS concentration in such hydrocarbons to levels less than 0.5 ppm (parts per million) by treating the hydrocarbons in liquid phase with soda-lime, defined in the specification as "lime which has added thereto sodium hydroxide * * *." The term "lime" is said to include quicklime and hydrated lime, the former being the product of calcining limestone, i. e., calcium oxide and magnesium oxide primarily, the latter being a dry powder obtained by treating quicklime with enough water to satisfy its chemical affinity therefor. The hydrocarbons may be "scrubbed" with a monoethanolamine (MEA) solution, prior to contacting with soda-lime, to remove "the acid gases, H₂S, CO₂, etc. and part (from 20% to 90%) of the carbonyl sulfide."

1. Process for purify [sic] liquefied normally gaseous hydrocarbons containing carbonyl sulfide which comprises contacting the liquid hydrocarbons with soda-lime.

Claims 2 and 7 are dependent on claim 1 and define preferred embodiments of the invention there defined, claim 2 being limited to propylene and claim 7 being limited to hydrocarbons containing from 34-167 ppm COS. Claim 3 recites a polymerization process in which liquid propylene is first treated with soda-lime to remove COS and is then polymerized. Claim 5 recites a process in which a liquid mixture of normally gaseous hydrocarbons is first treated with MEA to remove some of the COS after which propylene is separated from the mixture and the separated propylene is then treated in liquid phase with soda-lime to remove the remaining COS therefrom. Claim 8 recites a process in which liquefied propylene having a specified COS content is treated with soda-lime to reduce the COS content to 0-5 ppm.

  Karchmer et al.    3,000,988     Sept. 19, 1961
  Fleming et al.     2,959,627     Nov.   8, 1960
  Schulze et al.     2,301,588     Nov.  10, 1942

The U. S. filing dates of the first two antedate appellant's filing date.

Karchmer et al. disclose the removal of COS from gases containing same by contacting the gas with soda-lime, which, according to the patentees, "is a very good absorption material." Light, olefinic gases obtained from the catalytic cracking of petroleum, particularly ethylene, are disclosed, and it is stated that the purified ethylene is such as to be usable in the production of polyethylene. Preferably the gas is pretreated to remove various acidic gases by scrubbing, for example with an aqueous solution of sodium hydroxide, and to remove moisture, for example by passing the gas through a bed of activated alumina. Karchmer et al. make no mention of using soda-lime to remove COS from materials in liquid phase.

Fleming et al. disclose an "improved method for purifying olefins [obtained from pyrolyzing refinery gas, kerosene, gas oil, etc.] which are contaminated with highly unsaturated compounds and compounds of sulfur." The patentees first selectively hydrogenate the acetylenic and diolefinic compounds in the gas mixtures without appreciably converting the monoolefins such as propylene and butylene, and then pass the treated gas over "an organic sulfur conversion catalyst," disclosed in the specification as comprising "the oxides or sulfides of iron or copper separately or in admixture with the oxides of the metals of the left hand side of the 6th group of the periodic table," namely, chromium, molybdenum, tungsten, etc., "to selectively convert the carbonyl sulfide to hydrogen sulfide" again without substantially destroying the low molecular weight monoolefins. The desirability of removing COS from both ethylene and propylene is discussed, it being also pointed out that:

The problem of removal of carbonyl sulfide was not serious in the purification of ethylene for the production of polyethylene since the carbonyl sulfide was removed with the C₃ fraction in the ethylene concentration procedure. Both carbonyl sulfide and propylene have a boiling point of about -50°C. which renders removal by fractionation essentially impossible. Thus propylene separated from pyrolysis mixtures by conventional procedures will contain high percentages of carbonyl sulfide which must be reduced to a concentration of about 10 p.p.m. to be suitable for use in the production of plastics.

Schulze et al. disclose the removal of COS from hydrocarbon fluids such as the normally gaseous hydrocarbons by treating the fluid with an adsorbent material impregnated with an organic base such as monoethanolamine. In a preferred embodiment the adsorbent is also impregnated with a lead salt in addition to the monoethanolamine. The patentees say:

We usually prefer to treat in liquid phase since the volume of reagent required for nominal flow rates, say 0.5 to 5 volumes per hour per volume of reagent is not excessive. However, gas or vapor phase treating of normally gaseous hydrocarbons is satisfactory if provision is made in the size of the reagent bed to allow contact times corresponding to linear vapor velocities under five feet per minute.

The issue therefore is: would it be obvious to treat normally gaseous hydrocarbons in liquid phase with soda-lime to remove COS therefrom when the prior art shows the treatment of such hydrocarbons in vapor phase with soda-lime for the same purpose and the prior art also shows the treatment of such hydrocarbons in both liquid and vapor phase with a MEA type treating agent for the same purpose?

There appears to be agreement of the parties that essential to the proper resolution of this issue is a consideration of all the record evidence, including an affidavit filed under Rule 132. Such has been the law in this court for several years, and that regardless of whether any "doubt" as to patentability exists upon an examination of the prior art alone.2 The aforesaid affidavit purports to establish as fact that appellant's treatment of hydrocarbons in the liquid phase rather than the gas phase taught by Karchmer et al. gives "vastly superior results" in treatment of the same hydrocarbons. It is admitted that "essentially the entire case of appellant is based on the affidavit," and so it seems to us. We are concerned primarily with (1) whether these superior results have been shown and (2) if so, whether they would be unexpected.

What the affidavit shows is a comparison of two experiments, in each of which a charge stock consisting of propylene containing 250 ppm COS was passed through a reactor consisting of a two-inch diameter steel pipe, three feet in length, containing 3.1 pounds of commercial soda-lime. In "Run A — Vapor Phase Operation," the charge stock was passed through the reactor at a temperature of 15°F., a pressure of 50 psig, and a rate of 4.4 pounds propylene per hour per pound soda-lime; in "Run B — Liquid Phase Operation," the charge stock was at a temperature of 75°F., a pressure of 140 psig, and was passed through the reactor at a rate of 4.8 pounds propylene per hour per pound soda-lime. The effluents were periodically analyzed for COS content, it being explained in the affidavit that for polymerization of propylene to solid polymers the maximum tolerable COS content is 5 ppm — a so-called "break-through" value. This value was reached in Run A (gas phase) after a little more than one hour's operation and was reached in Run B (liquid phase) only after about 12 hours' operation.

Regarding this work, affiant Duff, a chemical engineer under whose supervision the experiments were performed said:

The data for these runs * * * clearly demonstrates the unexpectedly superior carbonyl sulfide removal efficiency of the liquid phase process when compared to the vapor phase process. In his opinion those skilled in the art would have been surprised to find that the liquid phase operation was significantly more efficient than the vapor phase operation, particularly, when the amount of propylene charged to the process was somewhat greater for the liquid phase run (4.8 #/hr./#) than for the vapor phase run (4.4 #/hr./#). The fact that applicant's liquid phase process is almost ten to twelve times more efficient than the prior art vapor phase process is striking proof that such liquid phase operation is not merely a matter of choice available to one skilled in the art.

Affiant also states that the process claimed in the subject application is used in a propylene purification plant having a capacity of 120,000,000 pounds of purified propylene per year and that, in his opinion, "similar superior results would be found when charging other normally gaseous hydrocarbons in liquid phase to applicant's process." It is against this background that the examiner, in his answer to appellant's brief before the board, said:

Applicant argues the unexpected result using liquid phase contact by referring to an affidavit under Rule 132 (Paper No. 16) submitted to show a comparison between liquid and vapor phase contact. This data is in part reproduced on pages 13 and 14 of applicant's brief. Such a showing is not convincing of the patentability of liquid over vapor phase contact. It merely discloses that over an extended period of time, liquid phase contact is more efficient than vapor phase contact. However, applicant's claims are silent as to the length of time of contact so that any advantage of using liquid over vapor phase does not appear in the claims.

It is further noted that the comparative results are merely a comparison of a single temperature and pressure with another single temperature and pressure. To conclude from such a showing that alleged unexpected results are dependent on a difference in phase, which includes many different temperature and pressure variations, would appear to be a stretch of inductive conclusions. Upon viewing the Karchmer et al. data (columns 3 and 4) it is seen that C.O.S. concentrations are obtained in the gas phase which vary from 0.2 p.p.m. to 74.8 p.p.m., the former being less than that in either phase of applicant's affidavit tests.

Notwithstanding the fact that applicant may obtain an improved efficiency using particular conditions such a showing cannot lend patentability to the present claims for a more basic reason. Schulze et al. disclose the fact that liquid phase contact is known in carbonyl sulfide removal systems. Accordingly, it would be obvious to try liquid phase contact in the Karchmer et al. carbonyl sulfide removal process, so that no matter what result was obtained thereby, such a feature would not be patentable. It is noted that by using liquid as opposed to vapor phase, applicant merely obtained the expected result, namely, removal of carbonyl sulfide. To obtain even an improved efficiency is a matter of degree not a difference in kind.

While we disagree with the Examiner and do consider that the Rule 132 affidavits [sic: affidavit], on the basis presented, demonstrate an apparent improved result, this improvement is not an unexpected one in view of the differences to be anticipated between vapor and liquid-phase operations at the same through-put. As is evidenced by Schulze et al. (page 3, column 1, last paragraph) linear velocities and apparatus must be altered to obtain equivalent results in vapor-phase operation as compared with liquid-phase operation.

We agree with the Examiner's position as expressed in his Answer except for his interpretation of the affidavit results noted above, and possibly with the choice of language in the expression "obvious * * * to try". In our opinion it would be obvious to substitute, in the Karchmer et al. vapor-phase contact of olefinic hydrocarbons with soda-lime, the known engineering alternative of liquid-phase operation. This substitution with the requisite alteration of apparatus and process parameters to obtain equivalent results is clearly taught in the above-indicated paragraph in Schulze et al.

We think a reasonable interpretation of the board's position is that appellant's affidavit does establish his process to be more efficient than that of the prior art, but that such an improved efficiency would have been expected from a reading of the Schulze et al. disclosure and therefore does not put appellant in a more favorable position on the question of patentability.

While much theoretical discussion has been presented, for the first time in the three briefs filed here, about the true nature of the phenomenon occurring within the "reactor"; about the effect or effects thereon resulting from changes in temperature, pressure, through-put, etc.; about what this phenomenon has been labeled in appellant's specification, in both its original and amended form; about the nature of the phenomena occurring in the prior art processes; about the necessity, or lack of it, for the number of variations in conditions between Run A and Run B reported in the affidavit; and about the significance of variations in conditions between Run B (the liquid phase operation) and those reported in an example in the specification, we find the arguments on most of these points of little help in drawing conclusions about the issues there discussed. What we have been able to conclude from this discussion, however, is that what appellant attempted to compare in his affidavit was the effect on COS removal of liquid vs. vapor phase, and that what he had hoped to do was to maintain temperature, through-put, and all other parameters except pressure constant, the variation in pressure being necessary and responsible for the difference in phase. Instead, because of "unforeseen mechanical limitations in the equipment in which the experiments were performed," appellant says, temperature and through-put were not held constant. Whether the effect of these variations was to decrease rather than increase the efficiency of liquid phase operation, as contended by appellant, is a point we feel it unnecessary to decide since we have not been persuaded by the Patent Office arguments that "an apparent improved result," to quote the board, does not in fact exist. Thus we, like the board, believe the important consideration is whether such an improvement in efficiency was to be expected. We think it was not.

As we recently said in In re Neely, 361 F.2d 255, 53 CCPA ___:

The appealed claims are all method claims, but, so far as the issue of obviousness is concerned, the considerations applicable are no different from [those pertaining to] claims for physical structures or arrangements.

The improved results achieved in the soda-lime treatment of normally gaseous hydrocarbons in liquid phase as opposed to vapor phase is a property of appellant's process which is to be considered in determining the obviousness of appellant's claimed invention. In re Papesch, 315 F.2d 381, 50 CCPA 1084.

In support of its position that appellant's improvement would be expected in view of "the differences to be anticipated between vapor and liquid-phase operations at the same through-put," the board pointed to the passage in Schulze et al. which we have set out above in discussing that reference. This is all the evidence the board provides, and, after a diligent search of the record and briefs, we find it is all the examiner and solicitor have provided.

While the Schulze et al. disclosure undoubtedly says that linear velocity and apparatus must be altered in going from liquid to vapor phase operation, and presumably vice versa, to obtain equivalent results, we think that is all that it says. We find no disclosure or suggestion that liquid phase operation would give more efficient COS removal. That, however, is what appellant claims to have found. While a teaching that another, albeit related, process can be carried out in both gas and liquid phases may suggest the possibility of carrying out the Karchmer et al. process in liquid phase, by making certain alterations, this is not tantamount to a disclosure of how to make those alterations and, more importantly, that a substantially improved efficiency will result.

We have considered the solicitor's reliance upon In re Leum, 158 F.2d 311, 34 CCPA 762; and Ex parte Appeal No. 13,2313 for the proposition that because "the use of liquid phase in a known vapor phase process is well known to the chemical engineer," a disclosure of one type of process renders the other unpatentable. However, we find those arguments unpersuasive. Here the invention is more than a mere change from the vapor to the liquid phase; it also includes the discovery that by so doing, a much greater efficiency is obtained.

We also find unpersuasive the solicitor's contention that the affidavit "is entitled to no weight whatsoever" because it is not directed to an unexpected property "within the ambit of the original disclosure," quoting from In re Herr, 304 F.2d 906, 907, 50 CCPA 705, which opinion had in turn quoted the board. We think Herr not to be support for an affirmance for two reasons: (1) what was said there was not that the affidavit was entitled to no weight, but only that if the advantage is not disclosed in the application appellant is "not in a favorable position to urge it as a basis for the allowance of claims" (emphasis ours), citing In re Lundberg, 253 F.2d 244, 45 CCPA 838; (2) here we think the affidavit is directed to that which "would inherently flow" from what was originally disclosed, which is an interpretation we put on Herr in In re Zenitz, 333 F.2d 924, 52 CCPA 746.

Since we find the improvement appellant found has not been suggested by the prior art, we hold the claims to be patentable. In re Carabateas, 357 F.2d 998, 53 CCPA ___.

The decision of the board is reversed.

Reversed.

SMITH, Judge (concurring).

The legal conclusion of obviousness required by 35 U.S.C. § 103 must be predicated on factual considerations. Graham v. John Deere Co., 383 U.S. 1, 86 S. Ct. 684. Included among such considerations are the results which are achieved by the invention in issue. United States v. Adams, 383 U.S. 39, 86 S. Ct. 708. The further command of section 103 is that we consider the invention as a whole. This requires a factual assessment of such factors as the existing problem in the art towards which the invention in issue is directed; the respective contributions of the art and the applicant towards the solution of that problem; a consideration of the differences between the prior art and the invention in issue; and, finally, a determination of what would have been obvious to one of ordinary skill in the art at the time of appellant's invention.

Here appellant has asserted the factual aspects of the problem facing the art; has disclosed a factual solution for this problem and has submitted evidence of facts from which it appears that the differences between the prior art and the invention claimed produced improved results of a magnitude that raises serious doubts as to the factual basis for the examiner's finding of obviousness. Considering the record as a whole it seems to me the examiner's position as affirmed by the board must fail for want of factual support. Viewed with hindsight after appellant's disclosure it seems obvious indeed to substitute the claimed liquid phase operation for the vapor phase operation of the prior art. The appealed claims, however, are claims to a process and thus necessarily involve more than the mere substitution of a liquid phase for a vapor phase operation in an otherwise old process. Operational differences over the prior art and new parameters of operation are clearly required before appellant's liquid phase process is operational.

A process is a unitary concept which does not lend itself to the type of dissectional analysis here employed by the examiner and the board. In Cochrane v. Deener, 94 U.S. 780, 788, 24 L. Ed. 139 (1876), Mr. Justice Bradley pointed out:

A process is a mode of treatment of certain materials to produce a given result. It is an act, or series of acts, performed upon the subject matter to be transferred and reduced to a different state or thing.

The unitary nature of the claimed process does not appear to have been considered here by the examiner and the board. Instead, the rationale of the appealed decision appears to be simply that it would have been obvious for one of ordinary skill in this art to substitute a liquid phase reaction for the vapor phase reaction of the prior art. The error here in so considering appellant's process is the old error of considering the separate steps of the process rather than considering the process as a whole.

To here give effect to the mandate of 35 U.S.C. § 103, we must start with a unitary process, which is the invention as a whole, and then consider the factual significance of the claimed portions thereof as they contribute to the "effect" by which the obviousness of the process is to be measured. Thus when appellant submitted affidavit proof of this "effect" and related it to the differences between the claimed process and the prior art processes, I think much more is required of the examiner than a reiteration of his subjective opinion concerning the obviousness of the invention even though stated in terms of criticism of the averments in the affidavit. At this level both the applicant and the examiner should be concerned with development of facts on which the ultimate legal conclusion of obviousness can be grounded. Here the office position fails because the examiner did not come forward with factual support for his position after appellant had factually supported the contrary position. If the office position is to be sustained, when challenged it must be supported by factual data from the technical literature or references which would have made obvious to one of ordinary skill in this art, at the time of appellant's invention, that the change from vapor phase to liquid phase operation would produce the results appellant has shown to flow therefrom.

I do not consider that the examiner has here discharged the obligations imposed on the Patent Office by the Deere case. The legal conclusion of obviousness under 35 U.S.C. § 103 must be based on factual considerations which I find to be missing from the present record.

MARTIN, Judge (dissenting).

The key question here is whether the improvement in efficiency, which I assume arguendo is shown by the affidavit, was to be expected.1 The majority is of the view that it was not; I think it was expected from the teaching of Schulze et al. that:

* * * gas or vapor phase treating of normally gaseous hydrocarbons is satisfactory if provision is made in the size of the reagent bed to allow contact times corresponding to linear vapor velocities under five feet per minute. [Emphasis added]

The improved result alleged by appellant here is one of efficiency: i. e., a savings of time, or as Schulze et al. say, "contact times". In my view, the above quoted passage of Schulze et al. teaches that one must adjust the size of the COS reagent bed so that the linear vapor velocity through it is not over 5 feet per minute. This is in contrast to the liquid phase requirement of .5 to 5 volumes of liquid propylene per hour per volume of reagent. Regardless of the exact figures that one of ordinary skill in the art could determine on the basis of the relationships explicitly taught by Schulze et al., it is clear that in going from say .5 volumes liquid propylene per hour per 1 volume of reagent to the volume of gas corresponding to the 15 volume of liquid at a rate such that the gas passes through a five foot long bed in no less than one minute, more reagent bed will be required. That means if the same amount of reagent, "1 liquid volume" in the example above, is to be used, the times will be longer, and correspondingly the efficiency less.

Thus, while I do not view the exact efficiency alleged to be shown, 12 hrs. vs 1 hr., as being necessarily predictable, I do not think such absolute predictability to be required by the law, where an improvement in efficiency is to be expected from the teachings in the art.

I note also that there is no convincing argument that the resultant ppm value of COS left in the propylene is any improvement in view of the Karchmer et al. (gas process) data showing .2 ppm. The problem of removal of COS from propylene was recognized by Fleming.

Thus I would affirm.