Opinion
No. 80-2089.
March 25, 1982.
John W. Hofeldt, Haight, Hofeldt, Davis Jambor, Chicago, Ill. (Rolf O. Stadheim, Haight, Hofeldt, Davis Jambor, Chicago, Ill., and John E. Wilkinson, Topeka, Kan., with him on the brief), for plaintiff-appellant.
Warren N. Williams, Schmidt, Johnson, Hovey Williams, Kansas City, Mo. (John M. Collins, Schmidt, Johnson, Hovey Williams, Joseph A. Crites, Kansas City, Mo., and J. Donald Lysaught, Kansas City, Kan., with him on the brief), for defendant-appellees.
Appeal from the United States District Court for the District of Kansas.
Before HOLLOWAY and DOYLE, Circuit Judges, and KUNZIG, Judge.
Honorable Robert L. Kunzig, Judge of the United States Court of Claims, sitting by designation.
The Honorable Robert L. Kunzig heard the oral arguments in the above entitled case. Unfortunately, however, subsequent to the arguments and before the within opinion was submitted to the panel, the Honorable Robert L. Kunzig passed away, on February 21, 1982. Thus, he did not vote on or participate in the within opinion. The remaining members of the panel have proceeded in accordance with 28 U.S.C. § 46(d).
The appeal herein is from a judgment entered by the United States District Court Judge of the District of Kansas which dismissed a complaint which charged patent infringement. This action was brought by Swift Agricultural Chemicals Corp. against Farmland Industries, Inc., the appellee herein, and Farmers Chemical Co., a subsidiary of the Farmland corporation. The charge was that Farmland willfully infringed the patent of Swift No. 3,464,808, which was referred to as the 808 or Kearns patent. This is a process patent which pertains to the production of liquid ammonium polyphosphate fertilizer from a wet process phosphoric acid.
The trial court ruled that the patent was invalid. The crucial question in the case is whether the trial court erred in its conclusion that the invention contained in the Kearns letter patent was invalid. A further contention is that the district court was in error in adjudicating invalid the claims other than those which were in issue. (There is a dispute as to whether these additional claims were in issue.) A third issue is whether the court erred in holding that Farmland avoided infringement based on its argument that (a) it used superphosphoric acid and (b) the residence time in the melt in its process is more than one second.
In support of its claim that the trial court reached out and adjudicated claims that were not in issue, it is pointed out that the post-trial submissions of Swift charged Farmland with willful infringement in claims 1, 2 and 4 only.
Our main concern is with the validity of the Kearns patent which discloses a method for making liquid ammonium polyphosphate fertilizer (referred to as APP.) The fertilizer contains substantial amounts of nitrogen and phosphorous, elements essential to plant nutrients. The fertilizers are often sold in solid granulated form. However, liquid fertilizers appear to be preferred because of the ease of application in soluble form and its ready availability to plants. But the liquid is not free of problems, as we will discuss.
APP, as the substance is commonly called, is ordinarily made by the combination of phosphoric acid with ammonia. What is referred to as the ordinary "wet process" or orthophosphoric acid is low in phosphorus pentoxide. The formula for this is P[2]O[5]. It tends to contain substantial amounts of impurities which occur in the phosphate rock from which the acid is made. These impurities are ordinarily in the form of iron and aluminum compounds and they tend to precipitate out of solution when the orthophosphoric acid is combined with the neutralizer, ammonia. These are the substances which combine to make the product. Thus, the APP is inclined to have gelatinous sludge-like precipitates which renders its use difficult. It clogs the fertilizer application equipment.
The orthophosphoric acid form contains a significant amount of water hydrated to the acid. If the phosphoric acid is heated to drive off the water a point is reached where acyclic polymerization occurs. This form of the acid is best suited to holding impurities in solution.
For a long time efforts have been made to prevent the creation of this gelatinous sludge from forming during the APP manufacturing process. The common solution is to combine purified furnace grade or highly concentrated wet process superphosphoric acid. This is created by heating wet process acid until enough molecular dehydration has taken place. This is characterized by high P[2]O[5] contents, 70% or more, and the presence of a substantial percentage of phosphates in the polymerized form. Those phosphates sequester impurities in solution and form stable compounds that do not precipitate out of solution when the acid is neutralized with ammonia. One of the weaknesses of this approach is its high cost. Ordinarily wet process orthophosphoric acid is cheaper and so most efforts have been directed at the development of a commercial process capable of producing APP from such acid. The Kearns patent, which dates back to 1969, at which time it was granted to Tommy Carter Kearns, a former Swift employee, undertakes to disclose one such process.
There is no dearth of prior art. All of the patents describe a method which seeks to use the ingredients so as to avoid the sludge problem.
Discussion of Kearns Patent
The Kearns patent, which is the patent in suit, describes the use of a pipe shaped jet reactor for the purpose of mixing or combining the wet process orthophosphoric acid with ammonia at a high temperature, 450 degrees to 650 degrees F. The period of time for combining is less than one second, followed by a quenching with water to form a liquid APP fertilizer. Thus, the orthophosphoric acid utilized in this process contains 54%-68% P[2]O[5]. The process utilizes heat energy provided during the reaction for the purpose of simultaneously neutralizing and molecularly dehydrating the acid. At least 20% of the acids orthophosphates are converted to polyphosphates during the reaction and the resulting product has self-sequestered properties. The impurities present in the acid remain in solution in the APP instead of precipitating out in the form of gelatinous sludge.
This process has been employed by Swift since 1967. Its commercial success has not been extensive. The problems with the product limit its usefulness. Nevertheless, Swift has prevailed in at least one infringement action involving the Kearns patent. Its adversary was Usamex, Inc. See Swift Chemical Co. v. Usamex Fertilizers Co., Inc., 197 U.S.Q. 10 (E.D.La. 1977) (Usamex I); Swift Chemical Co. v. Usamex Fertilizers Co., Inc., 490 F. Supp. 1343 (E.D.La. 1980) (Usamex II), aff'd, 646 F.2d 1121 (5th Cir. 1981).
The time span in which the chemicals come together is disputed. Swift employs a steam residence calculation for determining the mixing time of the acid and ammonia and, as noted, calculates residence time at less than one second. Farmland measures the time that the APP melt itself remains in the reactor, resulting in a melt residence time greater than three seconds.
For an explanation of the steam residence calculation see infra pp. 1357-1358.
The "melt" is the product formed where the ammonia and phosphoric acid combine.
The length of the trial was approximately eight days. At its completion the trial court ruled in favor of the defendants, concluding that the Kearns patent was neither valid nor infringed by Farmland. The opinion of the district judge is reported at 499 F. Supp. 1295 (D.Kan. 1980).
On the subject of validity, the ruling of the trial court was that the patent was anticipated by prior art and that it was obvious. Before issuance of the Kearns patent several other inventors had patented or applied for patents in processes similar to the 808 patent. The court focused in particular on methods developed by Young, Getsinger and Lutz and Rubio.
The trial court also ruled that even if the Kearns patent were valid, the Farmland process did not infringe upon it because the process used by Farmland utilized superphosphoric acid (68% P[2]O[5]) instead of the wet process orthophosphoric acid 54%-68% P[2]O[5] which was called for by the Kearns patent. Further, the court continued, the doctrine of file wrapper estoppel was capable of preventing Swift from claiming that a process involving superphosphoric acid infringed upon the Kearns process.
During patent prosecution Kearns had represented to the patent office that the process was distinguishable from the prior art because it utilized ordinary phosphoric rather than superphosphoric acid. To be noted is the fact that the Farmland process entails a residence time of more than one second. The court found Farmland's melt residence method for calculating residence time scientifically more valid than Swift's steam residence formula. Based on this difference there was no showing of infringement.
Swift has advanced five prior art processes which it seeks to distinguish from the Kearns method; these include the following: 1) Young, U.S. patent No. 3,044,851; 2) Lutz and Rubio, Patent Application Serial Number 352,764; 3) Hignett, U.S. Patent No. 3,171,733; 4) Bookey, U.S. Patent No. 3,375,063; and 5) Getsinger, U.S. Patent No. 3,382,059. These are briefly described below.
The two most relevant inventions to our inquiry are the Young patent and the Lutz and Rubio application. The latter failed to receive a patent; the patent office declared an interference and awarded the patent to Bookey. Lutz v. Bookey, 170 U.S.P.Q. 594 (Patent Board of Interference 1971).
The Young Patent, filed 6/9/61, patented 7/17/62
In February 1957 Dr. Young, a chemist with a Ph.D, was employed by the Union Oil Company. He had entered with a good many inventions to his credit. He personally made a sketch of APP reactor apparatus, which was in fact built and operated. This sketch depicts a continuous tubular pipe apparatus and method wherein ammonia and preheated merchant grade phosphoric acid are simultaneously injected into the pipe and produce APP by a simultaneous neutralization and molecular dehydration reaction; self-sequestering APP melt is then quenched with water to yield a self-sequestering liquid fertilizer product. In the Young process wet phosphoric acid is first heated in a pressurized tank to dehydrate and concentrate it. The acid is then cooled and reacted with ammonia to produce a self-sequestering liquid fertilizer.
The trial court found and concluded that the Young method anticipated the current patent even with Young's recital of two steps instead of only one, because the nature of Young's method was fundamentally the same as that of Kearns. The lower court found further that using Swift's method of calculating residence time the Young process established a residence time of less than one second.
The Young patent heats the wet process acid (35-55% P[2]O[5]) prior to the addition of ammonia. The introduction of ammonia can be brought about either during or after the heating process. If ammonia is added during the heating, it is a one step process because the exothermic heat of reaction further concentrates the acid; if added after, it is two steps. The heated acid is concentrated so that up to 40% of the acid is in the acyclic polyphosphate form. This heating drives off water and volatile impurities.
The patent was summarized by the trial court as follows:
"5. The Young patent, 3,044,851, discloses a process that is materially identical to that described in the Kearns patent in suit, No. 3,464,808, in that it discloses a continuous reaction between phosphoric acid and ammonia, in a tubular reactor, at a temperature of between 450 degrees F. and 650 degrees F., with a residence time of less than one second, involving simultaneous neutralization and molecular dehydration, and producing an ammonium polyphosphate fertilizer bearing self-sequestering properties."
Getsinger filed 7/2/64, patented 5/7/68.
The trial court considered the Getsinger process to be "materially identical" to Kearns. Getsinger uses wet process acid with a P[2]O[5] content of between 30 54%. The acid is put in a scrubber vessel and partially neutralized with ammonia at a temperature between 250-350 degrees F. Then the partially neutralized product is led into a reactor where more ammonia is added to form the ammonium polyphosphate melt. In the reactor the temperature is between 400-500 degrees F. The court found simultaneous neutralization and molecular dehydration of the acid occurs in a reaction vessel (No. 6) of the Getsinger patent where the appropriate pressures are present. The trial judge did not consider it a significant difference that the Getsinger patent requires a tank instead of a reactor pipe since the method and not the equipment is the crux of a process patent. The specifications recite that the Getsinger process involves "the ammoniation of wet process phosphoric acid."
Lutz and Rubio Patent Application, filed 6/9/61.
The court made a factual finding that the Lutz and Rubio process describes a continuous reaction between phosphoric acid and ammonia in a tubular reactor, at a temperature of between 450 and 650 degrees F., with a residence time of less than one second, involving simultaneous neutralization and molecular dehydration and producing an ammonium polyphosphate fertilizer bearing self-sequestering properties. The patent application states that approximately 40% non-ortho APP can be produced by injecting respective streams of phosphoric acid (50%-58%) P[2]O[5] and ammonia into the pipe and causing a simultaneous neutralization and molecular dehydration reaction to occur at temperatures of 400-500 degrees F. The contact time for the Lutz process was determined to be well below one second. Lutz had testified that the residence time of his invention, calculated by the use of the Swift formula, was on the order of .01 seconds.
The Lutz application states "tests were in a pipe reactor under various conditions whereby concentrated wet process orthophosphoric acid and anhydrous ammonia were contacted directly for the purpose of producing a concentrated ammonium polyphosphate product."
The trial court concluded that the Lutz process was virtually identical to the Kearns process, citing additional evidence that the residence time was less than one second. The trial court also held that even if the prior art did not literally anticipate the Kearns patent, that the patent was nevertheless invalid because the process it disclosed was obvious to anyone with ordinary skill in the art. The court also found that the prior art included descriptions of all of the elements important to the Kearns process, such as the use of a pipe or jet reactor, high temperature ranges and short residence times. The difference between the prior art and the Kearns patent according to the court amounted only to minor distinctions about recommended equipment. Also, the end product in the Lutz effort was dry material, which, of course, is of no consequence since a soluble APP melt can be produced by the introduction of water at the appropriate stage.
Hignett, filed 9/8/61, patented 3/2/65.
72%-85% or "super H[3]PO[4] (electric furnace acid) or wet process of 65%-75% is used. Apparently, the process is marked by no pre-heating. This has been called a three-step process. First, there is employed a reactor vessel, then a solidification and granulation chamber is used, then cooling equipment and finally through a screen. The Hignett process has an agitator in the reactor vessel. Though the Hignett process has general similarity to the Swift process, it is much less similar than the methods of Young, Getsinger and Lutz and Rubio.
Bookey filed 2/23/64, patented 3/26/68.
The Bookey patent begins with wet process acid with a P[2]O[5] content between 45-60%. A vertical column or tank type reactor is employed. There is simultaneous concentration of the orthophosphoric acid to the nonortho form along with dehydration and ammoniation. Temperature in the reaction is between 338 degrees to 482 degrees F. When a vessel is used agitation is to occur. A residence time of one hour was utilized in one of the patent's examples.
The Issue of Unpatentability Due to Obviousness.
The view that we take, namely that this patent was unpatentable because of obviousness, renders it unnecessary for us to consider in detail the issue of infringement. We approve the district court's ruling on this question. The court's findings are not clearly erroneous. We agree as well with its conclusion as to anticipation in the prior art. Our view of the case is that the patent is invalid; furthermore we are of the opinion that the patent is void due to obviousness. This latter issue we consider to be entitled to the emphasis in this opinion. Accordingly, the patent is contrary to 35 U.S.C. § 102 and is obvious within the terms of 35 U.S.C. § 103.
The Supreme Court noted in its landmark decision in Graham v. John Deere Co., 383 U.S. 1, 17, 86 S.Ct. 684, 693, 15 L.Ed.2d 545 (1966) that "manifestly, the validity of each of these patents turns on the facts". Id. at 5, 86 S.Ct. at 687. The Court in Graham observed that § 103 and matters contained in it had been recognized by the cases long prior to the adoption of the section and that § 103 was a codification of the requirement of non-obviousness. The 1851 decision of the Supreme Court in Hotchkiss v. Greenwood, 11 How. 248, 13 L.Ed. 683, recognized this when it said that there had to be more ingenuity and skill than that possessed by an ordinary mechanic acquainted with the business. The decision recognized the distinction between new and useful innovations capable of sustaining a patent and those that failed to meet this standard. When § 103 was enacted nonobviousness became an added statutory condition of patentability. The Court's opinion in Graham termed nonobviousness a more practical test of patentability. The procedure under § 103 of the Act for determining patentability was considered by the Supreme Court in Graham, supra, 383 U.S. at page 17, 86 S.Ct. at page 693:
Under § 103, the scope and content of the prior art has 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. Against this background, the obviousness or nonobviousness of the subject matter is determined. Such secondary considerations as commercial success, long felt and unsolved needs, failure of others, etc. might be utilized to give light to the circumstances surrounding the origin of the subject matter sought to be patented. As indicia of obviousness or nonobviousness, these inquiries may have relevancy. See Note, Subtests of Nonobviousness: A Non-technical Approach to Patent Validity, 112 U. of Pa.L.Rev. 1169 (1964).
Thus, when a process or device is not identically disclosed in the prior art, but would be obvious to a worker of ordinary skill in the field at the time the invention was made, then § 103 is not satisfied. True Temper Corp. v. C.F. I Steel Corp., 601 F.2d 495 (10th Cir. 1979). In Graham the challenged device had a spring clamp which allowed plow shanks to be pushed upward when an obstruction was encountered in the soil. Thus, the upward push was reversed and the shank returned to its normal position. It was held that it failed to pass the test of non-obviousness. The device was held by the Supreme Court to be invalid because it did not meet the tests set forth in § 103.
Further Analysis of Kearns in Light of Prior Art
We turn now to consideration of the Kearns patent and a comparison of that with the relevant prior art.
The district court found the Kearns process to be a combination patent, synthesizing known elements to get an improved result.
Creation of liquid ammonium polyphosphate fertilizer from phosphoric acid and ammonia is not a new idea. The novelty that Kearns claims is the use of merchant grade wet process acid, a cheaper acid less pure than the "electric furnace" acid which contains virtually no metal impurities, in combination with liquid or gaseous ammonia to simultaneously concentrate and dehydrate the acid, while ammoniating it to produce the APP melt that becomes the aqueous fertilizer. Although we have examined a number of variables in each of the prior processes in the Kearns patent, the crucial elements of comparison are 1) the type of acid used, 2) whether there is a simultaneous concentration of the acid creating a substantial non-orthophosphate formation along with dehydration and ammoniation, 3) the temperature of the reaction, 4) the mixing time in the reactor (residence time) and 5) the percentage of non-orthophosphate in the end product melt.
The type of phosphoric acid which Kearns begins with was known to the prior art. In fact, of the five processes we have mentioned, only the Hignett method advocates starting with a "superphosphoric" acid made by the electric furnace method that has a phosphorus content, expressed in terms of phosphorus pentoxide (P[2]O[5]), between 72% and 85%. This is highly concentrated acid; it is already dehydrated to such an extent that a substantial amount of the acid is in the polyphosphate form before contacting the ammonia. The remaining methods teach that wet process acid with P[2] O[5] percentages between 30 and 60% may be used. Even the Hignett method permits the use of wet process acid if it is concentrated between 65 and 75% P[2]O[5].
The Swift Company in using Kearns' process frequently makes use of wet process acid nearer the low end of the Kearns P[2]O[5] range of 54%-68%. The crucial step in the Kearns process of reacting the acid ammonia to produce the concentrated polyphosphate acid without first dehydrating the acid alone is not unique in the Kearns patent. Of the prior art, only Hignett and Young start with sufficiently concentrated acid to obviate the need for simultaneously ammoniating and creating a substantial amount of polyphosphates. As to Young, however, his process can be performed so that the ammonia is added during or immediately after the heating of the acid, allowing the exothermic heat of reaction to further concentrate the acid.
Kearns requires that the ammonia and phosphoric acid combine at a temperature of between 450 degrees and 650 degrees F. Since no heat is added to the reaction, apparently the temperature is controlled by the pressure, volume and molarity of the reactants.
The Lutz and Rubio application prefers a reaction temperature of 400-500 degrees F. Getsinger, although he teaches a two-step approach, combines the partially neutralized acid with more ammonia at a temperature between 300-600 degrees F.
The mixing time or residence time among the various methods was emphasized by the trial court. There appears to be a significant amount of confusion over its definition and its value in determining validity and infringement.
Kearns gives a residence time of less than one second in claim one of the patent. Kearns equates residence time with the time that the steam is in residence. In a deposition entered into evidence at trial Kearns explained that residence time is a mathematical calculation to estimate the amount of mixing time that transpires during the reactants' high speed exothermic propulsion through the reactor pipe. Farmland calculates a longer residence time. However, Young and Lutz-Rubio have a residence time comparable to that of Kearns using the Kearns calculation method.
Final Analysis
From a consideration of the Kearns patent in relation to the others, we have concluded that Kearns' use of residence time is no more than a descriptive analysis of what goes on when the ammonia and acid are allowed to mix at one end of the pipe to create the melt and steam, and then uninhibitedly flow out the other end of the reactor along with any unreacted ammonia. The parties and the lower court considered very significant the different methods of measuring residence time, depending on whether the assumption was made that the melt separates from the steam within the pipe, and moves at a different speed, or instead, the APP melt is entrained in the steam generated by the reaction and both melt and steam move at the same speed. Kearns assumed that the melt was entrained in the steam, both moving at the same speed. He calculated the residence time by adding the total volume per unit of time of steam, free ammonia and melt and then dividing by the volume of the reactor. Kearns did not measure the mixing time empirically. There seems to be some evidence that the Kearns hypothesis is incorrect. However, it doesn't matter in considering the present problem. How the process works is of importance. Whether the inventor knows the theory is unimportant.
Deposition evidence at trial was that, based on Kearns' method of calculating residence time, Young, Getsinger and Lutz came up with a figure not greatly dissimilar to the Kearns calculation. Apart from this retrospective use of Kearns' residence time formula, the record contains a memo dated February 11, 1964 prepared at Dorr-Oliver, Inc., Lutz's employer, where a pipe reactor test for APP melt occurred using a residence time of a fraction of a second. Because the pipe reactor of the Kearns process involves a "plug flow" wherein the reactants are headed in the same direction, and the Kearns patent does not teach any mechanically induced mixing, agitation or hindering the melt from leaving the pipe of its own accord, then, at least as to the Young and Lutz processes, their residence measurements appear credible, for both of these prior methods use similar pipe reactors, temperature ranges, and types of acid.
We finally consider the ability to sequester or hold in suspension, the impurities present in the wet process acid. This is a crucial feature to the successful storage and application of APP fertilizer. The sequestration is best accomplished when the phosphoric acid is dehydrated to the point where acyclic polyphosphate formation occurs. The wet process acid is almost exclusively orthophosphate when manufactured. Kearns claimed that his process would achieve at least a twenty percent conversion of the phosphate in the melt product to the non-orthophosphate form. The prior art evidenced by Hignett, Getsinger, Young and Bookey contains this desired sequestration ability and the required polyphosphate formation is well known. The Young patent and the Lutz and Rubio application, the most similar processes to the Kearns method, teach that their processes will yield as much as forty percent orthophosphate conversion.
Under the Graham discussion, we have examined the content of the prior art and the Kearns patent and made the crucial comparisons. We find, as did the district court, that considering the overlap among the prior art, the gaps where Kearns has altered the prior art in some fresh way are few and insignificant to the end product of self-sequestering APP fertilizer. We agree with the district court that because of the similarity between the prior art and the Kearns patent and the extensive research in this area, the level of ordinary skill in this area was sufficient so that the particulars of the Kearns process were obvious at the time of its creation.
Our mode of analysis here should not call into question the clearly erroneous standard of review involved. We have not tried this case de novo on appeal, for such is not our mission. Halliburton v. Dow Chemical Co., 514 F.2d 377 (10th Cir. 1975). We have, however, looked at the evidence and have found nothing clearly erroneous in the lower court's application of the Graham approach to the nonobviousness issue.
Swift complains that the district court erred in holding the entire Kearns patent invalid and it should have limited its ruling to claims one, two and four, those that were dealt with in the post-trial submissions of the parties. This issue is without merit. Farmland's answer brought up the defense of the patent's invalidity. There is nothing to show that the validity of the entire patent was not at issue, therefore. Moreover, the other four claims in the patent, 3, 5, 6 and 7 are dependent on the others, and particularly on claims 1 and 2, and could not stand alone as patentable processes.
Our conclusion differs from that of Judge Sear in the two Usamex Fertilizer opinions. The Lutz and Rubio application was not before Judge Sear. Furthermore, our function in reviewing the non-obvious determination of the lower court is to apply the extremely limited clearly erroneous standard. We have found that the district court acted without error. We have previously noted our approval of the trial court's decisions on the invalidity based on prior anticipation together with the infringement issue. We have expressed our preference to emphasize the obviousness issue and at the same time we have expressed our approval of the infringement and anticipation issues as they were decided by the district court.
The judgment of the district court is affirmed.