August 13, 1969. Rehearing Denied November 24, 1969.
Edmund C. Rogers, Laurence C. Kingsland, John T. Rogers, St. Louis, Mo., Russell H. Clark, Chicago, Ill., Kingsland, Rogers, Ezell, Eilers Robbins, St. Louis, Mo., Hume, Clement, Hume Lee, Chicago, Ill., of counsel, for appellant.
William D. Hall, Elliott I. Pollack, Washington, D.C., Thomas A. Reynolds, Jr., Winston, Strawn, Smith Patterson, Chicago, Ill., Moore Hall, Washington, D.C., for appellee.
Before FAIRCHILD, CUMMINGS and KERNER, Circuit Judges.
KERNER, Circuit Judge.
Plaintiff (Deep Welding) sued defendant (Sciaky) charging infringement of plaintiff's United States Patent No. 2,987,610 (Steigerwald) covering a method of welding using a controlled electron beam. Defendant asserted as affirmative defenses the invalidity of the patent, laches, and unclean hands (allegedly resulting from the establishment of a limited monopoly in unpatented electron beam welding machines by a parent corporation of plaintiff through improper use of the patent). The trial court found the patent valid, rejected Sciaky's other defenses, found some infringement and ordered the case referred to a master for an accounting. Sciaky appealed asserting as error the finding of validity of the patent. On this ground, we reverse.
It is axiomatic that findings in patent cases are like those in any other case in that they are entitled to our respect and should not be set aside unless clearly erroneous under Federal Rule of Civil Procedure 52(a). Armour Research Foundation v. C.K. Williams Co., Inc., 280 F.2d 499, 503 (7th Cir. 1960), and O'Brien v. O'Brien, 202 F.2d 254, 256 (7th Cir. 1953). See also 41 Mod.Fed. Prac.Dig., Patents key nos. 324(5 5/8) (a)-(c) (1961 and supps.). The major exception to this axiom is in situations where the evidence is documentary or where it involves the actual examination of a product or a device in operation. In such situations, the Court of Appeals has the right to interpret such evidence for itself and is as equally competent as the trial court to do so. See, e.g., Southern States Equipment Corp. v. USCO Power Equipment Corp., 209 F.2d 111, 117-118 (5th Cir. 1953); Plax Corp. v. Elmer E. Mills Corp., 204 F.2d 302, 308 (7th Cir. 1953); Stewart-Warner Corp. v. Lone Star Gas Co., 195 F.2d 645, 647-648 (5th Cir. 1952); Falkenberg v. Golding, 195 F.2d 482, 486 (7th Cir. 1952); and Armour Pharmaceutical Co. v. Richardson-Merrell, Inc., 396 F.2d 70, 72 (3rd Cir. 1968). See generally 15 Mod.Fed.Prac.Dig., Courts key no. 406.3(12) (1960 and supps.). Careful analysis of the evidence below, both documentary and testimonial and by way of exhibits showing welds made by the process in question, convinces us that this case is determinable upon its documentary evidence. We find ourselves in the same position as did an earlier panel of this Court in Charles Peckat Mfg. Co. v. Jacobs, 178 F.2d 794, at 802 (7th Cir. 1949), cert. denied, 339 U.S. 915, 70 S.Ct. 575, 94 L.Ed. 1340 (1950), which held:
Findings of fact are not to be upset unless they are clearly erroneous; a court of review will not interfere with them, in the absence of a very obvious showing of error. Graver Tank Mfg. Co., Inc. v. Linde Air Products Co., 336 U.S. 271, 69 S.Ct. 535, [ 93 L. Ed. 672]. But the ultimate question of patentability is whether the device meets the requirement of the statute. 35 U.S.C.A. § 31. Here we have a finding of fact of anticipation because of existing prior art patents. Each of these documents was before the trial court and is before us. Their interpretation, in view of the statute, is as open to us as to the District Court. True, there was some parol testimony in the court below, but we find that it did not in any way throw light upon the question of anticipation by the prior art. Consequently we feel free to review the evidence bearing upon anticipation by the prior patents.
1. THE PROCESS IN ISSUE
The process here involved is one in which a particle accelerator (electron gun) utilizing electron beam currents ranging from approximately 9 to 250 milliamps supplies accelerating voltages to the electrons of as little as 10 and as high as 30 KeV. The beam thus generated passes through a focusing coil which narrows it at the impingement site on the pieces to be welded so as to put the greatest amount of power into the smallest possible area. The welding process is performed in a vacuum chamber reduced to a pressure on the order of 1 × 10-4 mm. Hg so as to minimize contamination and beam interference and scattering. Lastly, the machine is set so that either the gun or the work moves at a speed which insures that a weld will be achieved rather than a hole or a channel. Beam current and focus, accelerating potential, and welding speed are all variable and the correct settings for any given operation will depend on the characteristics of the metals to be joined and the thickness to be penetrated by the beam. The latter depends not only on the thickness of the metal pieces but also upon the type of weld to be made.
The process is basically an adaptation of the electron microscope to perform an energy transfer rather than its normal use for a reflection or interference visual display.
The process represents a considerable advance over older methods of welding which depended upon heat transfer by conduction. In older welding methods, e.g., metal inert gas welding, in the absence of bevelling, the heat of the torch flame melts a surface portion of the metal creating a molten puddle. Continued application of the flame causes the puddle to propagate itself by continuing to melt solid metal with the heat of the molten metal at the interface of the two. Such welds typically exhibit a hemispherical shape in which the ratio of width to depth is on the order of 2 to 1 though, in some cases, unity may be achieved. Width is measured at the widest part of the weld.
In electron beam welding, two modes are available. In the first, beam strength is low and the weld is caused by the propagation of heat from a molten puddle formed at the point of beam impact. Welds thus achieved are substantially similar to metal inert gas welds.
In the second mode, electron beam welding is performed by the process at issue. The beam is set above a "threshold value" for the pieces to be joined and penetrates deeply into them diffusing its energy throughout the depth of penetration. The penetration of the beam into the material affords the electrons an opportunity to transfer energy and create heat throughout the depth of penetration thus creating a deep narrow molten zone which quickly solidifies. The resultant weld is deep, rather than shallow, narrow, rather than wide, and exhibits typical width to depth ratios from 1 to 3 through 1 to 30 and, in some instances, more. These welds are known as "deep welds."
Thus, when the beam is set above the "threshold value" for any given metals and the thickness to be penetrated, the electrons actually penetrate through the solid metal and, by energy transfer, almost instantaneously melt the metal throughout the depth of penetration. It is not important for our decision whether the energy transfer is effectuated through collisions with atoms of metal, through electron capture or through other excitation of the atoms, though the latter two seem the more probably frequent occurrence due to the small width of an atomic nucleus when compared with the larger volume of mass taken up by the various shells or energy levels of electrons surrounding the nucleus. It is important to note, however, that such energy transfer is not caused by the creation of a zone in which metal has been vaporized as this would result in drilling rather than welding. The long narrow "V" shaped welds which result show that there is an expected greater penetration at the surface nearest the gun (beam source), while the absence of a center line or core within such welds tends to cast doubt on the supposition that a vaporized channel is created which permits the electrons to pass through and which closes by cooling after the beam is turned off as the speed of cooling and solidification probably should show some interface or dislocation in micrographs of the weld structure. Thus, in "deep welding", the electrons penetrate the metal faster than heat conduction is able to create an old style melt puddle.
Drilling can be performed by the same machine by extending the time which the beam impinges on a particular area. Smaller holes can be drilled by causing the beam to pulse in microsecond pulses which vaporize small portions of surface metal sequentially so as to form a narrow precise hole. Cf. U.S. Patent No. 2,989,614 at col. 1, lines 58-69 (Steigerwald).
The crucial factor involved in the production of deep welds is the threshold value of the material to be welded. Unless this value is exceeded, the beam will not be of sufficient strength to penetrate the material faster than the heat created by surface impingement can form a molten puddle and the resulting weld will display the hemispherical welds achieved by older methods of heat propagation welding. This threshold value is expressed in terms of the amount of power brought to bear on a given area and is usually expressed in terms of kilowatts (KW) per square millimeter (mm²). The kilowatt power of the beam is determined by multiplying the accelerating potential or voltage of the electron measured in thousands of electron volts (KeV) by the beam current measured in milliamps (ma). Thus the value of the beam's power (P) on a square millimeter of metal surface with a constant resistance using the numerical unknowns "X" and "Y", can be expressed by the formula:
P = XKeV • Yma _____ _______________ mm² mm²
Expanding the terms to unity, the formula reads:
P = X • 103eV • Y • 10-3a _____ __________________________________________________________ mm2 mm2 or: P = XY W _____ ____ mm² mm²
Thus, either the wattage or the voltage and amperage per square millimeter defines a parameter of power per square millimeter. Since we have based our holding of invalidity, in large part, on our analysis of the prior art, it is important to realize that prior discussions of either type of parameter would, in the proper context, anticipate the patent in suit.
For example, a square butt deep weld can be produced in two pieces of Stainless Steel Type 304 with thickness of .500" using an accelerating potential of 30 KeV and a beam current of 250 ma. Substituting 30 for X and 250 for Y, the equation series in the text becomes:
P = 30 KeV · 250 ma _____ ______________________ mm² mm²
P = 3 · 10 · 10² eV · 2.5 · 10² · 10-² a _____ _____________________________________________________________ mm² mm²
P = 3 · 10 · 10² · 2.5 · 10² · 10-² W _____ _____________________________________________________________ mm² mm²
P = 3 · 2.5 · 10 · 10² W = 7.5KW _____ ____________________________________________________________ mm² mm² mm²
Since the accelerated velocity of the electrons in the beam is crucial, wattage per square millimeter would probably not define a deep weld for values less than 10 KeV.
2. STANDARDS OF PATENTABILITY
The date of invention in the United States is crucial as the state of the prior art is measured as of that date. The trial court was correct in determining that the date of invention was not the date of application for the United States patent (May 29, 1959), but the date the invention was first introduced into the United States and thereafter quickly reduced to practice. Since that determination here properly rested, at least in part, on oral testimony, we are content to adopt the lower court's findings on this issue. Deep Welding, Inc. v. Sciaky Bros., 155 U.S.P.Q. 561, 569-70 (N.D.Ill. 1967) (Findings of Fact 49-56). Accordingly, Steigerwald is entitled to a date of invention of June 28, 1958.
The standards concerning determinations of validity based on obviousness as based upon prior art were clearly stated in Graham v. John Deere Co., 383 U.S. 1, 86 S.Ct. 684, 15 L.Ed.2d 545 (1966). An excellent summary of the standards is set forth in Appleton Elec. Co. v. Effengee Elec. Supply Co., 412 F.2d 579, 581 (7th Cir. June 23, 1969):
In Graham v. John Deere Co., 383 U.S. 1, 86 S.Ct. 684, 15 L.Ed.2d 545 (1966), the Supreme Court construed § 103 of the Patent Act, 35 U.S.C. § 103, and set out the standards for determining the presence or absence of obviousness. After a scholarly analysis of the history of patent legislation and judicial interpretation thereof, Justice Clark, speaking for a unanimous Court, noted that § 103 was relatively unambiguous and that its test of "nonobviousness" was more definite than the "invention" language of Hotchkiss v. Greenwood, 11 How. 248, [ 13 L.Ed. 683] (1851), which had until then been the principle judicial authority on patentability. The Court, 383 U.S. at 14-15, 86 S.Ct. 684, quoted with approval the following language from the Congressional Report on § 103:
"An invention which has been made, and which is new in the sense that the same thing has not been made before, may still not be patentable if the difference between the new thing and what was known before is not considered sufficiently great to warrant a patent. That has been expressed in a large variety of ways in decisions of the courts and in writings. Section 103 states this requirement in the title. It refers to the difference between the subject matter sought to be patented and the prior art, meaning what was known before as described in section 102. If this difference is such that the subject matter as a whole would have been obvious at the time to a person skilled in the art, then the subject matter cannot be patented.5
5. "Sen. Rep. No. 1979, 82d Cong. 2d Sess. 6 (1952); H.R. Rep. No. 1923, 82d Cong., 2d Sess. 7 (1952)."
The Court then stated the procedure to be used in determining whether or not a particular article is obvious:
"Under § 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. Against this background, the obviousness or nonobviousness of the subject matter is determined. Such secondary considerations as commercial success, long felt but 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" 383 U.S. at 17-18, 86 S.Ct. at 694.
This Court recently noted that under Graham, "the ultimate question is whether a hypothetical person having ordinary skill in the art would have readily found the same solution when addressing himself to the same problem." Gass v. Montgomery Ward Co., 387 F.2d 129, 130 (7th Cir. 1967).
These standards are to be strictly applied. As the Supreme Court held in Graham, 383 U.S. at 18 and 19, 86 S.Ct. at 694.
We believe that strict observance of the requirements laid down here will result in that uniformity and definiteness which Congress called for in the 1952 Act.
* * * * * *
Although we conclude here that the inquiry which the Patent Office and the courts must make as to patentability must be beamed with greater intensity on the requirements of § 103, it bears repeating that we find no change in the general strictness with which the overall test is to be applied. We have been urged to find in § 103 a relaxed standard, supposedly a congressional reaction to the "increased standard" applied by this Court in its decisions over the last 20 or 30 years. The standard has remained invariable in this Court. Technology, however, has advanced — and with remarkable rapidity in the last 50 years. Moreover, the ambit of applicable art in given fields of science has widened by disciplines unheard of a half century ago. It is but an evenhanded application to require that those persons granted the benefit of a patent monopoly be charged with an awareness of these changed conditions. The same is true of the less technical, but still useful arts. He who seeks to build a better mousetrap today has a long path to tread before reaching the Patent Office. [Emphasis added.]
Thus, our duty is clear. We must examine the nature, content and scope of the prior art to see what it fairly taught one having ordinary skill in the art as the art existed on the date of invention. If the state of the art anticipated or made obvious the invention sought to be patented, 35 U.S.C. § 103 (as well as §§ 102(a), (b), (e) or (g)) requires a holding of invalidity. See Gass v. Montgomery Ward Co., 387 F.2d 129, 130 (7th Cir. 1967); Young Corp. v. Jenkins, 396 F.2d 893, 894 (9th Cir. 1968); Comm'r of Patents v. Deutsche Gold-und-Silber-Scheideanstalt Vormals Roessler, 130 U.S.App.D.C. 95, 397 F.2d 656, 659-660 (D.C. Cir. 1968); and MICR-Shield Co. v. First Nat'l Bk. of Miami, 404 F.2d 157, 161 (5th Cir. 1968).
In applying the standards, we are bound by certain rules which are applicable under the facts of this case. Firstly, we must determine the question of validity on the basis of the record before us and not as it has been litigated in other cases in domestic and foreign fora, which cases plaintiff would have us rely on here. University of Illinois Foundation v. Winegard Co., 402 F.2d 125, 126 (8th Cir. 1968). Secondly, even though the disclosures of the prior art may fall short of "complete anticipation," anticipation may be found where achieving complete anticipation only required that one of ordinary skill in the art merely exercised that skill to complete the work. Leishman v. General Motors Corp., 191 F.2d 522, 530-531 (9th Cir. 1951). Thirdly, while it is preferable that "all of the elements of the patented device or their equivalents * * be found in a single prior device," it is sufficient for anticipation "if the general aspects are the same and the difference in minor matters is only such as would suggest itself to one of ordinary skill in the art." Amphenol Corp. v. Gen'l Time Corp., 397 F.2d 431, 437-438 (7th Cir. 1968). Thus, rejection or invalidity may be properly based on a combination of references or the prior art taken as a whole. Gen'l Time Corp. v. Hansen Mfg. Co., 199 F.2d 259, 264 (7th Cir. 1952); Akron Brass Co. v. Elkhart Brass Mfg. Co., 353 F.2d 704, 706 (7th Cir. 1956); In re Young, 403 F.2d 754, 757-758 (Cust. Pat.App. 1968); and Koehring Co. v. E.D. Etnyre Co., 254 F. Supp. 334, 361 (N.D.Ill. 1966). In this respect, the prior art is cumulative. Amerline Corp. v. Cosmo Plastics Co., 407 F.2d 666, 669-670 (7th Cir. 1969), and In re Lemelson, 397 F.2d 1006, 1009 (Cust. Pat.App. 1968). Lastly, we must evaluate all prior art references, domestic and foreign, for what they fairly teach one skilled in the art including nonpreferred embodiments. In re Boe, 355 F.2d 961, 965, 53 CCPA 1079 (1966); In re Lemelson, ibid; and In re Umbricht, 404 F.2d 386, 390 (Cust. Pat. App. 1968).
It should be noted that, in 1958, the art was relatively new and still in a quasi-experimental stage. The development of the art of deep welding by electrons took place in the laboratories of the French Commissariat à l'Energie Atomique and the Zeiss Corporation in West Germany. This work was largely performed by men such as Stohr, Briola and Steigerwald who held D. Engr. degrees. Therefore, the level of "ordinary skill in the art" was unusually high and the prior art must be examined at that high level. Thus, the rule that the mere achievement of a better result than that taught by the prior art is insufficient to constitute patentable invention is especially applicable here. In re Welter, 45 CCPA 1034, 255 F.2d 944, 946 (1958).
3. STATE OF THE PRIOR ART
In its processing of the Steigerwald patent application, the U.S. Patent Office considered eight prior patents for their bearing on the application. One of those patents was No. 2,932,720 (Stohr), relied on by defendant here to show invalidity. We adhere to the salutary rule that, except in most unusual circumstances, determinations made by an agency relying on its expertise in a field in which it is uniquely qualified should not be disturbed. In the patent area this rule requires that a patent be presumed valid over any prior art shown to have been considered by a patent examiner in the file of the patent. 35 U.S.C. § 282.
However, defendant also relies on U.S. Patent No. 2,981,823 (Candidus) and numerous articles published in American and European sources, or distributed at conferences, which were not cited in the Steigerwald patent application file. The general rule of presumption of validity over cited prior art has a well-known exception which covers the precise situation before us. The exception, stated simply, is that the presumption of validity is largely, if not wholly, dissipated when pertinent prior art is shown not to have been considered during the processing of the patent application. California Research Corp. v. Ladd, 123 U.S.App. D.C. 60, 356 F.2d 813, 819 n. 14 (1966); B.F. Goodrich Co. v. Rubber Latex Prod's, Inc., 400 F.2d 401, 405 (6th Cir. 1968); Great Lakes Stamp Mfg. Co. v. Reese Finer Foods, Inc., 402 F.2d 346, 352 (7th Cir. 1968) (see esp. prior Seventh Circuit authorities cited id. at 352 n. 5); Appleton Elec. Co. v. Effengee Elec. Supply Co., 412 F.2d 579, 581 (7th Cir. June 23, 1969); and Koehring Co. v. E.D. Etnyre Co., 254 F. Supp. 334, 360 (N.D.Ill. 1966).
We place our primary reliance on the articles and conference papers published prior to the date of invention, June 28, 1958. The trial court rejected some of these papers on the ground that they were not publications. The items rejected were papers distributed at various conferences in Europe or read at the conference and summaries distributed. Where a summary contains the crucial material of the primary paper, we regard it as sufficient prior art.
The major issue remaining, then, is whether these items were public. The evidence clearly shows that the conferences involved were open and, in most cases, Americans not only attended the conferences but also read papers there. See, e.g., defendant's exhibit (hereafter DX) No. 44, the Abstracts of Communications at the First International Congress on Vacuum Technology at Namur, Belgium, June 10-13, 1958. Fifteen papers read at the Congress were given by Americans, several of these papers were in the same subject classification as that of Stohr and Briola describing deep welds. By way of contrast, earlier cases have held the following to be "printed publications" containing crucial prior art applied to determining the validity of American patents: a glass manufacturer's catalogues in Imperial Glass Co. v. A.H. Heisey Co., 294 F. 267 (6th Cir. 1923) (this case also held that it was no longer necessary to show full description in a "public work" under the Act of 1836); a printed catalogue of a German manufacturer written in French in Jockmus v. Leviton, 28 F.2d 812, 813 (2d Cir. 1928) (opinion by Judge Learned Hand); a thesis available only in the Iowa State University library in Hamilton Laboratories v. Massengill, 111 F.2d 584 (6th Cir. 1940); a typewritten thesis available only at a particular library in Gulliksen v. Halburg, 75 U.S.P.Q. 252 (Pat.Off.Bd.App. 1937); and an unpublished thesis containing some handwritten material available only at a particular library and only with prior permission in Ex parte Hershberger, 96 U.S.P.Q. 54 (Pat.Off.Bd.App. 1952). We agree with Judge Hand in Jockmus that a more widely circulated printed item is clearly to be preferred as a printed publication evidencing prior art than a thesis which is merely on deposit in some library and which has not been otherwise published. Thus, conference papers, such as here, which are distributed at public meetings of persons skilled in the art under consideration are proper evidence of prior art. See esp. Genesco, Inc. v. Knapp Bros. Shoe Mfg. Corp., 152 U.S.P.Q. 793, 800-801 (N.D.Ill. 1966, 1967). We hold that such papers are within the category "printed publications" as used in 35 U.S.C. § 102(a) and (b). Thus, the papers and articles here involved are properly before us for our consideration and we shall discuss them in chronological order.
Our citation of Hershberger is not meant to imply unquestioning acceptance of its rather extreme holding.
Conversely, we decline to consider DX No. 24, a paper dated March 1, 1957, which, while it does discuss deep welds ("* * * il s'agit d'une soudure pénétrée."), was not shown to be anything other than an internal paper of the Commissariat à l'Energie Atomique (hereafter CEA) which was marked on the top "Diffusion Restreinte" (a distribution restriction approximately equivalent to "classified information" or "secret").
1) DX No. 25, PX No. 120-21, paper by Stohr presented at the Fuel Elements Conference in Paris, France, November 18-23, 1957, printed in French (PX No. 121) in the CEA Bulletin d' Informations, Scientifiques et Techniques No. 15 (March 1958) and in English (PX No. 120) in the Proceedings, Book I, Fuel Elements Conference by the U.S. Atomic Energy Commission (March 1958). The teachings of this paper are rather limited. It describes the electron beam welding machine and briefly discusses closure of the nuclear fuel element cans by circumferential welds. It teaches welding in areas (point d'impact) on the order of 0.1 mm². However, it does not mention deep welds. The parameters of impact power given are
50 KW to 50 KW _____ _____ 6 cm² 8 cm²
or approximately 83.3 to 62.5 W/mm². This is even less than the lowest value shown for a deep weld in thin beryllium in PX No. 14 which is 110 W/mm² (10 KeV • 11 ma/mm²). While it is alleged that there was a typographical error and 6 — 8 cm² should be 6 — 8 mm², thus showing P/mm² values of 8.3 to 6.25 KW/mm², we believe that the publication must be read as distributed and may not be found to teach more than it fairly states on its face.
Given the infant state of the art at the time of publication, we are not convinced that the alleged error would have been immediately recognized and that those who read it would have automatically multiplied by 100.
2) DX No. 39A is a communication by Stohr presented at the Vacuum Conference on February 4, 1958, and published in French in Le Vide, No. 75 at 163 (May-June 1958). This paper is an expansion of the earlier one and most importantly, corrects the alleged error showing impact energy in thick pieces
of 5 · 104 W which do definitely yield deep ________________ 6 to 8mm²
welds when, as also stated, "la tension de soudure utilisée est comprise entre 10000 et 14000 Volts." Id. at 166. Compare PX No. 14 which shows voltages ranging from 10,000 to 30 eV resulting in impact powers ranging from 110W/mm² to 7.5 KW/mm². Le Vide's parameters of 8.3 to 6.25 KW/mm² are not only within but even exceed the energy levels used presently. While the paper discusses deepening welds by the use of a pulsing device ("un hacheur de courant à thyratrons," ibid), thus leading plaintiffs to correctly argue that this was not a true deep weld but merely a method of deep heat propagation welding, plaintiffs ignored the later crucial discussion of the fuel elements which were made with "a larger penetration of weld" ("une plus grande pénétration de soudure." id. at 168). This discussion did not depend on the thyratron pulsator and is Stohr's first clear statement as to deep welding, also teaching the parameters required.
This was possibly due to the difficulty in focusing the beam down to less than 6 mm². Thus, it was necessary to use large impact power with an accelerating voltage at least in excess of approximately 9 KeV so as to overcome the normal tendency of the electrons in the beam to mutual repulsion. While the Pierce gun used by Stohr had its problems, we are not persuaded that plaintiff has shown that it was incapable of deep welding. In fact, DX No. 81 is a piece of plaintiff's advertising literature which admits that the Pierce gun is capable of such welds but merely argues that the Steigerwald gun is better. We agree. However, this still does not impair the fact that one could perform the patented process with a Pierce gun, albeit not with as sophisticated results as the Steigerwald gun, or with accelerating voltages from 100 KeV and up.
3) DX No. 34 is an excerpt from a paper by Stohr, published in French by CÉGOS, which was presented at a public information session of the French Départment d'Orientation Technique et d'Information Scientifique entitled Développements Récents dans la Technique du Vide et ses Applications held March 26-29, 1958. This paper cites back to the prior papers as to the process. It also teaches a power parameter of 10 KW/mm² in a vacuum of 10-6 mm Hg, a parameter sufficient for most deep welding and in fact, much greater than any of those listed in PX No. 14.
4) DX No. 40 is a paper by Stohr and Briola which was first presented as a communication to the Société des Ingenieurs Soudeurs (Welding Engineers Association) on April 24, 1958, presented later to a meeting of the Institut International de la Soudure (International Welding Institute) in Vienna on June 30, 1958, and published in French in 12 Soudages 165 (May-June 1958). It refers back to the first two items above and sets forth all which was taught by them. It teaches the power parameters of 50 KW/6 to 8 mm² and notes the use of voltages as high as 40 to 50 KeV for some applications ("* * * que l'intensité des rayonnements dépend de la structure du métal mis en œuvre," id. at 172). It teaches the variability of voltage and current for obtaining differentiated results. It also states that narrow and precise welds are obtainable by virtue of concentrating large amounts of energy on small surface areas.
The paper states, id. at 171:
La possibilité de concentrer sur des surfaces de très faibles dimensions une énergie considérable permet de réaliser des soudures fines et précises impossibles à obtenir par tout autre procédé assurant la fusion du métal. [Emphasis added.]
That the word "fines," as used here, means "narrow" rather than (as alleged below) "excellent" is apparent not only from the context, but also from Petit Larousse (11th ed. 1962) which, at 434, defines it in its usual meaning as "Qui est d'une extrème petitesse", "mince", "leger", and "très aigu."
5) DX No. 36 is virtually identical to the third item above (the CEGOS article) and states that it is based on it. It is a paper by Stohr published in English in 3 Nuclear Power 272 (London, June 1958) which was received in the John Crerar Library in Chicago on May 22, 1958. It, therefore, evidences the availability of Stohr's invention in the United States in the English language prior to the date of invention.
6) DX Nos. 20, 43 and 44 are a translation, an original and a summary of a paper by Stohr and Briola given at the First International Congress on Vacuum Technology in Namur, Belgium on June 10-13, 1958. This is the paper referred to previously in discussing the large number of Americans present at the Congress. The paper contains an excellent description of the process teaching the variation of beam focus, current and voltage to produce deep welds in the aluminum EL3 fuel elements for the Saclay high flux nuclear reactor. It also points out the perfect homogeneity of the weld. The crucial part of the paper states:
(2) Variation de la puissance appliquée
Suivant la focalisation du faisceau d'électrons, la surface d'impact varie considérablement. La puissance dissipée sur cette surface d'impact peut également prendre des valeurs très différentes. Il suffit pour cela de modifier: soit le courant de chauffage (nombre d'électrons), soit la tension d'accélération (énergie cinétique).
Ces possibilités de réglage permettent des soudures trés délicates a réaliser par les moyens classiques.
(a) Soudure d'acier inoxydable très mince. — De nombreuses soudures ont été réalisées avec succès sur des tubes d'acier inoxydable d'épaisseur comprise entre 0,07 et 0,15 mm. Pour ce type de soudure, la surface d'impact est très faible: inférieure a 1 mm².
(b) Soudures profondes sur aluminum. — Dans le cas d'une soudure sur bords relevés (par exemple, soudure d'un tube avec un bouchon en forme de coupelle), en concentrant une forte puissance sur une petite surface bien centrée sur la ligne de séparation des deux bords, de fortes pénétrations de soudure sont obtenues. Il faut alors opérer en deux passes: la première passe soude en profondeur, la seconde passe (avec spot plus large éventuellement) donne I'aspect arrondi a la ligne de soudure.
Les cartouches du réacteur EL3 de Saclay (Figs. 2a et b) sont gainées en aluminium de 1 mm d'épaisseur. Le bouchon a une épaisseur de 2 mm. En utilisant le procédé de soudure par bombardement électronique, nous avons obtenu des profondeurs de penetration deux fois plus grandes qu'avec le procédé argon-arc utilisé au cours des essais.
A cet avantage s'ajoute celui, deja cité, de fermer la cartouche sous vide.
The translation of this portion reads as follows:
2. Variation of the applied power
Depending on the focalization of the electron beam, the impact area varies considerably. The power dissipated on this impact area can also have very different values. For such purpose, it is sufficient to modify either the beam current (number of electrons) or the accelerating voltage (kinetic energy).
The possibilities of adjustment permit to obtain welds which are very difficult to achieve by the classical methods of welding.
a. Welding of very thin stainless steel
Numerous successful welds have been made on stainless steel tubes of thicknesses between .07 and .15 mm. For this kind of weld, the impact surface is very small: smaller then 1 mm 2.
b. Deep welds on aluminum
In the case of a weld made on raised edges (such as a tube with cup-shaped plug), by concentrating a high power on a small area, well-centered on the line separating the two edges, deep welding penetrations are obtained. In this case, it is then necessary to work in two stages: the first one to weld in depth, and the second one (with a larger spot, possibly) to give a rounded appearance to the weld seam.
The cartridges for the reactor EL 3 in Saclay (Figs. 2a and b) are sheated in 1 mm thick aluminum. The plug has a thickness of 2 mm. With the electron beam welding process we obtained penetration depths twice as great as with the argon-arc process used during the tests. As already mentioned, another advantage is the possibility to seal the case in vacuum.
To argue as plaintiff did below, that "profondeur" means "profundity" rather than "depth", is to completely ignore the context in which the word was used. Moreover, Petit Larousse (11th ed. 1962) defines "profondeur" as "distance du fond par rapport à la surface, à l'ouverture" and, more importantly from a conceptual point of view, "L'une des trois dimensions d'un solide."
The abstract published in conjunction with the Congress includes the extent of vacuum used (10-4 to 16-6 mm. Hg) and states clearly that one of the advantages of the process is the obtaining of welds with considerable depth ("Advantages de ce procédé: * * *, (3) obtention de profondeurs de soudure considerables").
7) DX No. 42 is virtually identical with the Namur paper. It is credited only to Briola and was presented at the Second United Nations International Conference on the Peaceful Uses of Atomic Energy on June 13, 1958, and distributed in English and French versions. Its only additional teaching over the Namur paper is that the voltage setting used for the welding of the EL3 fuel element is 15 KeV. Comparing this figure with parameters given in prior papers, it is apparent by simple division that the necessary current strength to be used was on the order of 100 to 500 ma with a target area approximately 4 to 6 mm².
It is thus apparent from these seven papers that the process in issue was discussed at length and fully described prior to Steigerwald's date of invention. These papers teach welding in depth by beam penetration using an electron gun with variable focus, voltage and current. They also specify the parameters for voltage, impact power, impact area and vacuum strength. Nothing else remained to be taught. The only thing which remained to be done was to determine the particular settings needed for particular types and sizes of metals. Steigerwald's patent did not perform that function as it, too, offers no formula for calculating the needed values. The patent only teaches that the operator experiments with various settings until he finds one which produces a deep weld. Thus it is clear that Steigerwald's patent application was wholly anticipated prior to his date of invention and, by some papers, more than one year prior to the date on which the application was filed with the United States Patent Office.
In fact, the patent only teaches one setting using 100 KeV for welding 4 mm. thick zirconium plates. It does not even say whether this is for a butt, seam, lap or Tee weld thus leaving the reader unaware of whether the actual depth involved is 4 mm. or 8 mm. Also, apparently Steigerwald had not yet learned that it was more desirable to use higher beam currents so as to diminish the 100 KeV value which produces dangerous X-rays and requires more shielding between the gun and the operator than Stohr's 10 to 25 KeV.
Accordingly, we hold that United States Patent No. 2,987,610 issued June 6, 1961, to Karl Heinz Steigerwald is invalid as being obvious and wholly anticipated by published prior art. It is invalid under 35 U.S.C. § 102(a) because articles nos. 2-7 supra fully described the process in printed publications prior to the date of invention available to Steigerwald here. It is invalid under 35 U.S.C. § 102(b) because articles nos. 2-4 supra sufficiently described the process more than one year prior to the May 29, 1959, date of application.
It is also invalid under 35 U.S.C. § 103 because the sole difference between the application and the patent as issued as compared to the prior art was the partial explanation of the natural phenomenon which caused the process to succeed. Thus Steigerwald's discussion in the patent of why the beam penetrates and his mention of threshold value are mere surplusage in view of Stohr and Briola's articles. It is unnecessary that Stohr and Briola knew why they produced deep welds. It suffices that they did produce them and described the process to the public so it could be duplicated. "An inventor is only bound to show how his invention works, not why." 4 Deller's Walker on Patents § 238 (2d ed. 1965), quoted in Chemithon Corp. v. Procter Gamble Co., 287 F. Supp. 291, 303 (D.Md. 1968). The obvious corollary to this rule is that once a process is demonstrated in the prior art by its inventor, a patent cannot issue for the explanation by someone else of the scientific principle behind the process. See also Smith v. Hall, 301 U.S. 216, 226-227, 57 S.Ct. 711, 81 L.Ed. 1049 (1937), citing DeForest Radio Co. v. General Elec. Co., 283 U.S. 664, 686, 51 S.Ct. 563, 75 L.Ed. 1339 (1931); and Davison Chemical Corp. v. Joliet Chemicals, 179 F.2d 793, 794 (7th Cir. 1950).
Nor could the asserted validity of the patent be based on Steigerwald's use of a better electron gun. As noted above, fn. 11 supra, while the Steigerwald gun is clearly superior to the Pierce gun used by Stohr, the Pierce gun was capable of performing deep welds. An apparatus must go a substantial step beyond the prior art in order to be patentable. Dorr Co. v. Yabucoa Sugar Co., 119 F.2d 521, 528 (1st Cir. 1941). The improved gun is a relatively small step over the prior art, is not essential to the performance of the process, and was a change which naturally flowed from the prior art. Cf. In re Holmberg, 242 F.2d 791, 792, 44 CCPA 865 (1957).
As previously observed, defendant also claimed that the patent in issue was invalid when read in the light of U.S. Patents Nos. 2,932,720 (Stohr) and 2,981,823 (Candidus). Since '720 (Stohr) was cited in the file of '610 (Steigerwald), the presumption of validity is available against it. Also, '720 (Stohr) only teaches varying the parameters in an electron beam welding operation and makes no reference to beam penetration or depth of weld nor to the parameters necessary to achieve a deep weld.
On the other hand, '823 (Candidus) was not cited in the '610 (Steigerwald) file even though it was filed 1 year and 24 days earlier (May 5, 1958). This patent teaches welding in a vacuum of 10-5 mm. Hg using voltages of 20 KeV and a highly precise focusing mechanism. It states that "By adjustment of the beam focus, the weld can be made as narrow or wide as desired by the operator" (Patent No. 2,981,823 (Candidus) at col. 3, lines 4-6). It further states that the invention makes it possible to control "the size of the melt at the weld to achieve any desired depth or width of melting" ( id. at lines 13-14) and it "permits the welding of irregularly-shaped articles with ease since the electron beam can be made to have a very substantial depth of focus and the focal plane of the beam can be varied * * * by electrical or magnetic means" ( id. at 44-48). Reading these lines in context, we believe that '823 (Candidus) was probably not teaching deep welding by beam penetration but merely teaching the use of very high heat levels in a narrow area to produce deeper than usual heat-propagated welds in a relatively shorter period of time so as to reduce problems of impurities in the welds. Combining '823 (Candidus) with '720 (Stohr) teaches nothing more than using a machine with variable settings to perform such a welding operation. While the machine and process thus resulting would be capable of producing deep welds, there is no fair indication in this record that such was ever done nor that it was intended to be done.
However, the combination of '823 (Candidus) with any of the Stohr or Stohr and Briola articles teaching deep welding and/or deep welding parameters, does result in anticipation and obviousness. Accordingly, we hold that '823 (Candidus) was a part of the relevant prior art, that it was not considered by the Patent Office in its relevance to the patent at issue, and that its combination with the articles and papers discussed earlier is clear evidence of obviousness and anticipation such as would render U.S. Patent No. 2,987,610 (Steigerwald) invalid under 35 U.S.C. § 102 (a), 102(b) and 103.
The above holdings as to obviousness and anticipation strike down the crucial claims of the patent (claims 1, 4 and 7) which repeat prior art teachings of beam penetration or deep welding by an adjustable electron gun. The other claims in the patent and, indeed, claims 4 and 7 also, are merely the adaptation of the process to long-known welding configurations known to the most rudimentary practitioners in the art far below the level of what was ordinary skill in the art in 1958. Thus, claims 1-3 teach seam and spot welding by the process, claims 4-6 teach Tee and lap welding and claims 7-11 teach seam welding in metals having equal or unequal thermal or preparation characteristics. Based on all of the foregoing, we hold all the claims (claims 1-11) invalid.
The Steigerwald patent is also defective and invalid because of a fatal indefiniteness. While the patent purports to teach welding, it nowhere specifies the avoidance of drilling, a result which is the antithesis of the one sought to be obtained. In fact, the parameter given in col. 3, lines 32-38 of the patent is on the order of 100 KW/mm², a parameter which could produce a drilled hole in only a fraction of a second more time than would be required to weld. In later patent applications filed in other countries, the parameters recited amounted to approximately one-tenth of the power here taught. However, these later parameters are of no avail in this case as it is a firm rule that a patent's validity is determined by what is recited on its face and not what is asserted elsewhere. Graham v. John Deere Co., 383 U.S. 1, 34, 86 S.Ct. 684, 15 L.Ed.2d 545 (1966); Fife Mfg. Co. v. Stanford Engrng. Co., 299 F.2d 223, 226 (7th Cir. 1962); and Ohio Citizens Trust Co. v. Lear Jet Corp., 403 F.2d 956, 958 (10th Cir. 1968).
The prior art use of electron beams for welding, machining and drilling metals is an old and crowded one. As long ago as 1907, U.S. Patent No. 848,600 (von Pirani) was granted teaching the use of electron beams ("cathode rays") to melt metals in a vacuum. In 1936, U.S. Patent No. 2,128,581 (Gardner) taught the use of an electron beam to drill a hole in a diaphragm which would then be used to control beam size at a lower energy than that used for drilling. In 1941, U.S. Patent No. 2,267,752 (Ruska) taught substantially the same method of using increased beam intensity for producing ultra-filters. In 1948, U.S. Patent No. 2,435,273 (Hatfield) taught the use of an electron beam for heating two materials so as to fuse a coating material to the material to be coated. In 1956, U.S. Patents Nos. 2,746,420 and 2,771,568 (Steigerwald) (same as the inventor of the patent in issue) taught the use of high energy electron beams for surface evaporation and vaporization of metals including etching by vaporization which is a partial drilling. In 1957, U.S. Patent No. 2,778,926 (Schneider) taught the adaptation of the electron gun to welding. In so doing, he also taught etching and drilling ( id. at col. 2, lines 46-71). Also issued in 1957 were U.S. Patents Nos. 2,793,281 (Steigerwald, a comprehensive patent covering many phases of drilling by use of a high energy electron beam), 2,793,282 (Steigerwald, a companion patent covering the use of the beam for rounding off corners of extremely small cubes of metal so as to form miniature ball bearings), and 2,809,905 ( Davis et al., the use of the beam for melting and refining metals). Later U.S. Patents, all with applications filed prior to the date of invention of the patent before us, included:
2,819,380 (Eaton) — use of high intensity electron beam to penetrate a metal mask for use in color television;
2,844,706 (Lorenz) — similar use of high intensity beam for making minute holes for spinneret apertures;
2,883,544 (Robinson) — use of 50 KeV beam to etch 1/10mm. cavities in germanium to make transistors;
2,899,556 (Schopper et al.) — use of 200 KeV beam to machine or cause chemical reactions on metals;
2,902,583 (Steigerwald) — use of an electron beam which is pulsated for focus and defocus as a method of heating materials so as to shape them (including drilling);
2,932,720 (Stohr) — the electron beam welding patent discussed before;
2,968,723 (Steigerwald) — use of the beam to create tips or points on semi-conductor or transistor material by heating sufficiently to cause crystal expansion and
2,981,823 (Candidus) — discussed earlier.
'581 (Gardner), '752 (Ruska), '568 (Steigerwald), '281 (Steigerwald), '282 Steigerwald), '544 (Robinson), and '823 (Candidus) were not cited in the file of the patent in issue. If they had been, the patent examiner might have perceived that while the application taught welding, it failed to teach the avoidance of melting, drilling, cutting and etching. Nowhere in the patent or its file is there any mention of avoiding these nor of time or power parameters which would avoid them.
It is axiomatic that process patents in crowded arts must have a preciseness of description. Universal Oil Prods. Co. v. Globe Oil Refining Co., 322 U.S. 471, 484, 64 S.Ct. 1110, 88 L.Ed. 1399 (1944). The language of the patent must tell the operator of a machine who desires to perform the process exactly how to proceed so as to successfully perform the patented process. Halliburton Oil Well Cementing Co. v. Walker, 329 U.S. 1, 8-9, 67 S.Ct. 6, 91 L.Ed. 3 (1946). The precision required must also be such as would put on notice all those working in the crowded art so that they do not infringe on the patented process. Pennsylvania Crusher Co. v. Bethlehem Steel Co., 193 F.2d 445, 448 (3d Cir. 1951).
The patent here is too broad as by not teaching the avoidance of drilling, melting, etching and cutting it can be read to cover valid claims in other patents recent in the prior art. In failing to teach with sufficient precision the duration and depth of beam penetration, it is analogous to the situation in Kaiser Industries Corp. v. McLouth Steel Corp., 400 F.2d 36, 48-51 (6th Cir. 1968), in which a patent for steel manufacture (also a crowded art) failed to describe with sufficient precision the depth of penetration of a jet of oxygen into the molten metal.
Moreover, as held in Standard Oil Co. of California v. Tide Water Assoc. Oil Co., 154 F.2d 579, 582-586 (3d Cir. 1946), the patent must be sufficiently specific so as not to compel independent experimentation to determine the bounds of the inventor's claims. Here, experimentation is necessary not only to avoid the undesired results of drilling, etching, cutting and melting, but also to avoid infringement. As correctly observed by defendant, innocent infringement could occur when an operator would perform a normal heat propagation weld in .500" tungsten and then put a piece of .100" aluminum in the machine and start welding without changing the parameters. As noted in PX No. 14, while parameters of 8.3 KW/mm² will be required for some deep welding, parameters only 1/70 of that level will suffice for other metals. Thus, also, a machine set for electron beam cutting could also infringe by producing deep welds in other metals without changing the parameters.
This was recognized by plaintiff when, in intra-company correspondence to its salesmen it noted "Even Electron Beam Cutters are also capable of deep welding." DX Nos. 99-102, answer to Question 3 on DX Nos. 100 and 102.
Therefore, we hold that U.S. Patent No. 2,987,610 (Steigerwald) is not only invalid for anticipation and obviousness (as discussed supra), but is also invalid for being improperly and unpatentably vague and indefinite under 35 U.S.C. § 112 which requires the specification to "conclude with one or more claims particularly pointing out and distinctly claiming the subject matter which the applicant regards as his invention."
Wherefore, the judgment below is reversed.