Alkyl benzene sulfonate ("ABS"), the detergent most widely used in the 1950's, created severe pollution problems because it resisted biodegradation. It was known that structural "branching" in one part of the ABS molecule, the alkyl substituent, was the source of the problem, and that an increase in the incidence of "straight chain" or "linear" alkyl substituents would improve biodegradability. The problem was (a) to start with a linear, rather than non-linear, source material for the alkyl substituent, and (b) to minimize or avoid loss of linearity during manufacture. The total elimination of non-linear alkyl substituents from the finished product was not expected and is still considered an economic impossibility. Continental's research chemists claim to have discovered that if non-linearity was eliminated in about 90% of the product, it would be "substantially completely biodegradable." The Patent Office rejected plaintiff's process claims, but allowed product claims which defined the patented subject matter by describing its method of manufacture.1 The district court, without appraising the differences between the claims and the prior art, held the patent valid and infringed. We reverse.

The essential facts are not in dispute. For the most part, they are recited in the patent specifications, in unchallenged findings,2 in a glossary3 and historical statement4 on which both parties rely, and in uncontradicted testimony.

The process includes four stages: (1) segregating or obtaining a supply of "normal paraffins" (i. e., a species of linear or straight chain hydrocarbons) from a suitable petroleum source, such as kerosene; (2) partial chlorination- between about 1/10 and 1/3 (preferably 20 mole percent)-of the normal paraffins yielding a mixture of unreacted linear paraffins and chlorinated paraffins; about 90% of the chlorinated portion are "monochlorinated" (which is desirable) and about 10% are di-, tri-, and polychlorinated (undesirable); (3) alkylating benzene (or another aryl compound) with the chlorinated paraffins in the presence of a catalyst and distilling the resulting mixture to remove unreacted benzene, normal paraffins (which can be recycled) and certain impurities, thereby yielding the "detergent alkylate" described in Claims 1 and 3 of the patent;5 and (4) sulfonation of the alkylate to produce a surfactant, the product described in Claim 2 of the patent.6 That product is described commercially as a "soft detergent," or as LAS (linear alkyl benzene sulfonate) (even though it is only about 90% linear), in contradistinction to ABS (alkyl benzene sulfonate), which is known as a "hard detergent" because it is "hard" to biodegrade.

*****

* * *

The patent issued in 1967 on an application filed in 1965, as a continuation of a parent application filed on August 4, 1961.7 The inventors were research chemists employed by the plaintiff. The problem which confronted their employer and other detergent manufacturers, and which their invention was intended to solve, was well described in their specifications:

"Throughout about the past two decades, synthetic detergents have been increasingly displacing the traditional soaps in domestic cleaning applications in the United States. At the present time, it is estimated that about 75 percent of the combined sales of household cleansing surfactants are of the synthetic detergent type.

"The household synthetic detergents presently used embrace a number of different products; however, the bulk of these detergents are of the alkyl benezene sulfonate type generally referred to as ABS. This type of detergent is made by alkylating benzene with a comparatively highly branched C12 and/or C15 olefin and then sulfonating the resultant alkylation product. While these materials are excellent cleansing agents, their use has posed a considerable problem. The specific problem involved is that the ABS detergents are not readily removed or decomposed in sewage treatment plants. These detergents will remain throughout the treatment step as such without being significantly decomposed by the bacteria present. This resistance to decomposition is not due to the lack of suitable strains of bacterial [sic] within the sewage effluent nor to the length of time that can be allotted for treatment, but is fundamentally due to the chemical structure of the surfactant which resolutely resists metabolic attack by any kind of bacteria. One of the adverse effects of the presence of ABS in sewage effluents is that voluminus [sic] foams are caused to be formed which result in difficult and hazardous working conditions at the disposal plants. Also, the persistent foam formed constitutes a health hazard since the bacteria contained in the foam is often as high as 20 times that contained within the liquid phase.

"In addition to causing foam problems in the sewage plants, the presence of non-decomposable ABS in sewage effluents, including septic tanks effluents, has already resulted in considerable contamination of the available drinking water supply in populous areas. While it is generally believed that the ABS-type surfactant is not toxic to man, a currently considered proposal for revising the Public Health Service Drinking Water Standards to include a limit on the amount of ABS in potable water most signifcantly foreshadows the necessity for detergent manufacturers to produce a soft type of detergent, that is, one which is completely and readily biodegraded.

"It is accordingly the primary object of this invention to provide detergent alkylate compositions, which upon sulfonation give effective detergent agents which at the same time are substantially completely biodegradable."8 (emphasis added).

The specifications then recite the fact that bacteriologists have noted that the branched configuration of the alkyl substituent in conventional ABS detergents impeded satisfactory biodegradation, and concurrently that detergents containing straight-chain or linear alkyl substituents were subject to complete biodegradation. The inventors' mission was to find an economical, feasible method of producing a detergent which was "substantially completely biodegradable."

As we understand the record, the four-stage process they developed differed from the conventional method of making hard detergents in the first and second stages, but was essentially the same in the third (alkylation)9 and fourth stages.10 The first step was different because the selected source material contained straight chained paraffins instead of branched hydrocarbons. Plaintiff does not claim to have discovered or improved the method of segregating or obtaining the first stage material.11

The second stage involved a significant change from the chlorination step in the production of hard detergents. It substituted partial chlorination (between 10 and 35 mole percent) for 100 mole percent chlorination of the hydrocarbon source material. The purpose of this change was to maximize the quantity of monochlorides and minimize the quantity of di-, tri- and polychlorides in the second stage product. The monochlorides were desirable because they would not cause isomerization of the paraffins in the alkylation step, whereas the polychlorides were more likely to isomerize.12 Thus, a relatively low quantity of monochlorides (and a correspondingly high quantity of polychlorides) would tend to destroy the linear characteristics which was needed to assure acceptable biodegradability.

Although the third stage alkylation step is admittedly old, plaintiff contends that the aluminum chloride catalyst was proved economically feasible in a manner not obvious under the prior art.13 It had been assumed that unless the unchlorinated paraffins, which constituted from 65 to 90% of the stage-two product, were removed by distillation prior to alkylation, they would be isomerized and therefore could not be recycled. The hydrocarbons and other impurities such as unreacted benzene are normally removed by distillation after alkylation. The fact that they did not isomerize during alkylation was, according to plaintiff, an unexpected discovery which makes it possible to recycle the unchlorinated linear paraffins without incurring the cost of removing them prior to alkylation of the chlorinated paraffins.

As with the first stage, plaintiff makes no claim of invention in the fourth stage (sulfonation). Indeed, Continental expressly states that "Each of the steps by itself was old." Brief for Appellee at 10.

Plaintiff's primary contention is that the "combination produced a new product with surprising and unpredictable properties." These properties relate to the effectiveness of the product. Continental claims it was not obvious that the product of the suit patent, which is 85-95 percent pure LAS, would biodegrade satisfactorily because hypothetical persons with ordinary skill in the art would have realized that the partial chlorination would produce at least 10% di-, tri- and polychlorides, which in turn would cause a non-linear configuration in a portion of the alkyl radicals in the alkylate. The ultimate fact that this amount of non-linearity would not defeat commercial success, plaintiff argues, was not obvious.

Defendant admits infringement. Its only defense is that the invention was obvious. In support of this position it points out that plaintiff's own executives concluded that the product "has no unique or patentable features";14 that promptly after the inventors were assigned their mission they made the correct theoretical calculations with respect to the partial chlorination stage and achieved success on their first experimental partial chlorination as well as their first experimental alkylation;15 and that competing manufacturers developed the same product in response to the same market stimuli, apparently without any assistance from plaintiff's work. These arguments tend to establish obviousness, but, in our opinion, are not in themselves sufficient to overcome the findings and conclusions of the district court or the presumption of validity.

We are, nevertheless, convinced that the subject matter of the allowed claims was obvious notwithstanding the unanticipated features stressed by plaintiff.

The defense of obviousness raises a question of law. Armour & Co. v. Wilson & Co., 274 F.2d 143, 151-157 (7th Cir. 1960). The statute, 35 U.S.C. Sec. 103, provides that a patent may not be obtained "if the differences between the subject matter sought to be patented and the prior art are such that the subject matter as a whole would have been obvious at the time the invention was made . . . ." It is, therefore, imperative that the differences between the new product, as a whole, and the prior art be identified and evaluated as of the date of the invention. As Mr. Justice Clark plainly stated in his opinion for the Court in Graham v. John Deere Co.:

"Under Sec. 103, the scope and content of the prior art are to be determined; differences between the prior art and the claims at issue are to be ascertained; and the level of ordinary skill in the pertinent art resolved." 383 U.S. 1, 17, 86 S. Ct. 684, 694, 15 L. Ed. 2d 545 (emphasis added).16

In this case, the findings of the district court describe the prior art in great detail, but omit any specific reference to the differences between the claims and the prior art. Since the findings contain no express delineation of such differences, they cannot shed any light on the trial court's evaluation of the difference or differences it regarded as sufficiently critical to demonstrate that the subject matter as a whole was not obvious before October, 1960, when the invention was complete.17 Without such delineation and evaluation of differences, the district court's findings of fact do not adequately support its ultimate conclusion of patent validity. We must, therefore, evaluate these differences and their significance.

The parties take issue on three allegedly novel elements of the invention. Those elements are (1) the selection of the 10-35 mole percent chlorination range in order to maximize the production of monochlorides; (2) the fact that alkylation may be catalyzed by aluminum chloride without causing isomerization of the excess, unchlorinated linear paraffins; and (3) the unanticipated and remarkable success of the product in affording a solution to the foam problem.

Plaintiff properly emphasizes the importance of partial chlorination but implicitly acknowledges that the consequences of partial chlorination, as described in the patent, were predictable.18 Indeed, process claims encompassing partial chlorination were rejected by the Patent Office as obvious under the prior art.19 If the inventors' objective is assumed to have been the production of an 85-95% pure LAS detergent,20 the use of this established process as one step in the product's manufacture would therefore have been obvious to one skilled in the prior art. But there is an even more basic reason why this step in the process does not make the end product patentable. Even if we were to assume that this step was new-or indeed that the entire process was not only new but patentable as well-that assumption would not demonstrate that the product was new or nonobvious. For an old product made by a new process is not patentable; the product must be patentable in its own right.21 Therefore, the claim that the partial chlorination step was novel does not establish the patentability of plaintiff's product.

For comparable reasons the discovery that aluminum chloride will not cause isomerization of the unchlorinated paraffins does not establish the nonobviousness of plaintiff's product.

The argument that his discovery was unforeseeable relies heavily on the finding that aluminum chloride "was known to function as a catalyst for isomerization of hydrocarbon materials." (Finding 41, A. 44.) Since the term "hydrocarbon materials" encompasses an almost unlimited number of substances, that finding by no means indicates that a person with ordinary skill in the art would expect aluminum chloride to cause isomerization of linear paraffins.22 At most, we think the finding indicates that the consequences of using aluminum chloride as the catalyst were somewhat doubtful until after an experiment had been made. Since aluminum chloride had been so widely used as a catalyst in the production of ABS, at the very least it was obvious to experiment with this catalyst.23

Again, however, even if the discovery that aluminum chloride was an acceptable catalyst to use in the manufacture of LAS was nonobvious, that discovery related to the process rather than the character of the product. As a catalyst aluminum chloride simply promotes the alkylation reaction; it does not affect the qualities of the end product.24 The value of that discovery relates only to the fact that anticipated costs associated with the recycling of unchlorinated paraffins did not materialize.25 It seems to us that a discovery that certain procedures will result in a reduction of manufacturing costs cannot adequately support a product claim unless that discovery entitles the inventor to a valid claim on the process.26 If a product patent is to rest solely on the novelty of the processes by which it is made, then those processes must be patentable.27 In this case, since the Patent Office has determined that those processes are not patentable, we believe the product must be patentable in its own right.28

This brings us to the crucial issue- the obviousness of the product itself. In 1960, when the invention was made, it was perfectly obvious that a detergent containing 85-95% pure LAS would be considerably more biodegradable than the ABS, the hard detergent then on the market. Nevertheless, plaintiff contends that it was not predictable that the product would completely solve the foam problem, which it did. We believe that success was sufficiently obvious to bar patentability.

Plaintiff contends, and the district court found, that a detergent must biodegrade 90% in four to six hours in order to alleviate the foaming problem. Finding No. 8, A. 24. A detergent need not be fully decomposed because small amounts (less than .5 part per million parts of sewage) can be tolerated. The prior art taught that detergent molecules with linear alkyl substituents biodegraded promptly and presented no problem. Continental's patent claims a product which is 85-95% pure LAS.29 Since only 90% biodegradation was required, it was certainly obvious that the purest form of the product claimed (at least in the 90-95% range) would succeed. The LAS portion was expected to biodegrade; moreover, the remainder would also biodegrade to some extent.30 Clearly, most of the product range claimed by plaintiff could fairly be expected either to solve the foam problem completely or at least to be "substantially completely biodegradable."

Plaintiff argues, however, that it made the nonobvious discovery that certain impurities known as alkyl tetralins would readily biodegrade, making the product even better than expected.31 The patent itself does not disclose this discovery. Nor does Continental argue that this feature was of critical importance in solving the foam problem.32 Indeed, in the purest form claimed, the product would be a success even if the tetralins did not biodegrade at all. Regardless of whether or not the properties of the alkyl tetralins were nonobvious, the essential characteristics of the product as a whole were obvious.

It is true that the success of the product could not be determined with absolute certainty until it was field tested in an actual sewage plant. This fact, however, does not render the entire range of product described in the claims nonobvious.33 Even though the sewage plant is far removed from the laboratory, the prior art, based mostly on laboratory work, pointed directly toward plaintiff's work. While plaintiff should be commended for the commercial development of a socially desirable product, it has not earned the reward of a monopoly. As a matter of patent law, its work, must be regarded as routine experimentation and testing of an obvious product.34

The evidence supporting the conclusion that the subject matter of the invention as a whole was obvious is, in our opinion, plainly sufficient to overcome the presumption of validity resulting from the allowance of the three product claims after the application or its parent had been under study in the Patent Office for almost six years. The judgment is

Reversed.