General Electric Co. v. Alexander

MAYER, Circuit Judge.

The patents and the art here considered were fully discussed in previous opinions of the Circuit Court of Appeals and of this court. General Electric Co. v. Laco-Philips, 233 Fed. 96, 147 C. C. A. 166, affirmed 233 Fed. 96, 147 C. C. A. 166. General Flee. Co. v. Nitro-Tungsten Lamp Co. (D. C.) 261 Fed. 606, affirmed (C. C. A.) 266 Fed. 994.

The Just and Hanaman patent is now attacked upon the ground (a) of invalidity and (b) non-infringement. The Langmuir patent is attacked upon the ground of invalidity.

*291The bulk of the record is quite disproportionate to the simplicity of all but one of the questions involved. This extensive record is due in part to the fact that the court felt that the widest opportunity should be given to put forward almost any conceivable fact, experiment, or theory directed against validity or infringement as the case might be, in order that litigation, in any event, as to the validity of these patents should some day come to an end, one way or the other. If, in view of prior litigations, it he ultimately held in this case that the claims of these patents are valid, it is difficult to imagine what new matter can later be brought forth.

This introduction is regarded as desirable because otherwise the reason for the brevity of this opinion on certain points might not be understood. The three questions supra will be discussed in inverse order.

[1] 1. The Langmuir Patent. There is not a shred of merit to the present attack on validity. The previous case was ably tried for the defense by experienced counsel who did not neglect anything of consequence. The British Sinding-Larsen (Thompson) patent, No. 18,968 of 1899, was fully considered in the Nitro-Tungsten record.

Now it is claimed that the American Sinding-Larsen patent, No. 672,019, invalidates the Langmuir patent. The British and American patents are the same, except that the British patent mentions the pressure of four atmospheres, while the American patent does not specify the pressure, though it calls for a pressure above atmospheric.

No Sinding-Larsen lamps have ever been used, not because of the osmium lamp, but because the Sinding-Larsen theory was wholly wrong. As Langmuir testified:

“The point, is that the whole theory of the inventor is a false theory, but the theory shows very distinctly that he had in mind that an external pressure on the material of the filament would improve the life of the filament.
“As a matter of fact, we know to-day, and even Sinding-Larsen could have calculated it in liis time, that the pressure on the material of the filament, instead of on the vapor from the filament, would cause an increased rate of evaporation of the filament, and not a decrease. The whole theory is wrong, thermodynamically wrong. You can calculate through rigorous methods that the effect would be just the opposite of what he expeels. As a matter of fact, the penetration of a gas into the pores of a filament, if the phenomena'are dependent upon the evaporation, could not have any material effect upon the evaporation of the filament.”

Sinding-Larsen himself, in his British patent, No. 19,945 of 1900, realized the futility of his earlier patent.

Further, the material in the Sinding-Larsen lamp was carbon, and carbon will not do in the Langmuir co-ordination. The experiments or demonstrations on this score in court are not guides, because life is a basic factor in tests as to lamp efficiency. The point is, as Langmuir said:

“That producing a lamp * * * and running it for a couple of minutes at a high efficiency has no significance whatever as an indication of what a lamp is capable of doing.”

The Hop felt lamp (patent No. 29,285) consists of a small glass tube bent to the shape of an U, containing a drop of mercury and also *292containing, substantially in its center, a carbon filament, the whole surrounded by an external bulb like that of an ordinary lamp. The presence of the carbon filament is sufficient to exclude this patent from further consideration, and its theory as to the vapor of mercury which surrounds the filament after the drop of mercury evaporates, being raised to an intense white heat, is not Langmuir’s invention.

The same general observations apply to the extract from the Electrotechnische Zéitschrift. In brief, there is nothing of' theoretical or practical value in this Hopfelt patent, and, like other prior art, it was a failure.

Claims 4, 5, 12, and 13 are held valid and infringed. The remaining claims need not he considered.

[2] 2. Infringement of lust and Hanaman Patent. The opinion of this court as to this patent unanimously adopted by the Circuit Court of Appeals indicated that this patent disclosed invention of a high order in an art which performs a great service to mankind.

The construction of claims of such a patent will not turn upon dictionary fine distinctions. The courts must always ascertain what the claims mean in the light- of the state of the art and the accomplishment of the inventor.

Practical experience demonstrated the desirability of counteracting what is known as “offsetting,” infra, and for that purpose one of the means or methods employed is to introduce the rare earth oxide known as “thoria.” It is contended that thoria does not cohere, and that the tungsten is not pure nor homogeneous, and hence that there is not infringement of the claims. These claims are as follows:

“(1) A filament for incandescent lights consisting of tungsten in a coherent metallic state and homogeneous throughout.
“(2) A filament for incandescent lights consisting throughout of substantially pure metallic tungsten of high fusing point and electrically conductive, the light emitting properties of the filament being due «to the coherent, homogeneous metallic nature of the tungsten.
“(3) A ‘filament for electric incandescent lights comprising dense, coherent tungsten metal, having its fusing point approximately 3,200“ C., and capable of incandescent efficiency at the rate of less than one watt per candle power and substantially free from perceptible disintegration at that efficiency.”

When Coolidge’s drawn wire tungsten filament (referred to in the Tungsten Lamp Case) was used, it was found that it differed from the squirted filament (of Just .and Hanaman)” in that, when used on an alternating current circuit, it was sometimes subject to what is known as “offsetting,” and this offsetting has to do only with the life of the filament. As'the drawn filament is operated in a lamp, the fibers of which it is composed tend to reform as crystals. This crystal formation may, on an A. C. circuit, produce crystal boundary planes at right angles to the length of the filament, which are planes of mechanical weakness at the temperature at which the filament runs. On these planes slipping may occur, causing one part of the filament to offset laterally from another part. This effect is disadvantageous, as a part of the filament (especially small vacuum lamp filaments) will grow too hot and burn out at the point affected.

Of several ways of preventing this, one is the introduction of thoria. This thoria is in practice usually introduced by a process de*293scribed in the Coolidge patent, which consists in wetting the tungsten powder with a solution of thorium nitrate. In one of the subsequent operations the powder is heated so that the nitrate breaks up, leaving miscroscopic particles of thoria sticking to the very fine particles of the tungsten powder, which latter particles themselves are so small as to be almost invisible to the naked eye.

As the filaments are drawn into wire, these little particles of thoria are drawn out into, microscopic rods which naturally are all parallel to the axis of the filament. As the filament crystallizes these microscopic rods break up into microscopic or submicroscopic particles, so arranged and oriented (Langmuir, pp. 795-798, X-Qs. 504—516; page 805, X-Q. 539) as to exert an influence on the character of the crystallization such as to diminish the tendency to create the planes of weakness.

There is much testimony as to tests and microscopic investigations, including microphotographs, all on the questions as to whether the introduction of the thoria destroys or impairs the significance of the words “pure,” “coherent,” and “homogeneous” as used m the claims.

The evidence is convincing that thoria has a mechanical action and performs a mechanical function and does not have any chemical effect. It does not perform any office in respect of the light-giving characteristics of tungsten, but is a mechanical bracer or resister which prevents the filament from being burned out sooner than might otherwise happen. Colloquially put, it is as if one put some kind of mechanical reinforcement in a coat to protect it from mechanical injury, but did not change the coat nor the wool which composed it.

There is some controversy as to the accuracy of the microphotographs, but, from my point of view, it is unimportant whether a photograph shows a break or hole.

When Just and Hanaman used the words of their claims, such terms as “pure” and “homogeneous” were employed in contrast with the impure, composite filaments of the prior art. It was as if they said:

“For lighting purposes, we shall not mix tungsten chemically with other materials and make a composite filament as did Bott.omc and de Lodyguine.”

It would, indeed, he small reward for a great advance in the art if “pure” and “homogeneous” were construed as meaning that a mechanical action such as this would avoid infringement and make that which was “pure” and “homogeneous” neither one nor the other, further, the broad proposition of law involved was practically disposed of in the previous case in relation to the Coolidge filament. So 'as to “coherent”; this word is used to point out the compacted nature of the metal in a true metallic filament as distinguished from powders and the like. Plaintiff is quite right in its contention that—

“Tliere had been so many fruitless attempts to produce filaments from refractory substances, obtainable only as powders eonsisling of individual small crystals, that when an inventor obtained, or thought that he had obtained, a truly metallic filament manufactured out of such powder, so consolidated as to form a solid structure that would really hang together, be naturally hastened to call attention to the fact that it was so' consolidated that it would hang together, or, in a word, that it was coherent.”

*294See Elizabeth v. Pavement Co., 97 U. S. 126, 137, 24 L. Ed. 1000; Hoyt v. Horne, 145 U. S. 302, 12 Sup. Ct. 922, 36 L. Ed. 713; General El. v. Hoskins Mfg. Co., 224 Fed. 464, 140 C. C. A. 150; Edison v. Waring (C. C.) 59 Fed. 358.

It is therefore held that the claims of the Just and Hanaman patent are infringed.

[3] The Question of Invalidity of the Just and Hanaman Patent. This defense rests solely on the proposition that there was a German patent prior to the American patent here in suit, and that the American patent is invalid because it falls within the prohibition of R. S. 4887, as that section now reads. Originally section 4887 was as follows :

“No person shall be debarred from receiving a patent for his invention or discovery, nor shall any patent be declared invalid, by reason of its having been first patented or caused to be patented in a foreign country, unless the same has been introduced into public use in the United States for more than two years prior to the application. But every patent granted for an invention which has been previously patented in a foreign country shall be so limited as to expire at the same time with the foreign patent, or, if there be more than one,.at the same time with the one having the shortest term, and in no case shall be in force more than seventeen years.”

By the act of March 3, 1897 (chapter 391, 29 Stat. 692), which applied to all patents granted on and after January 1, 1898, on applications filed after such date section 4887 was amended to read as follows:

“No person otherwise entitled thereto shall be debarred from receiving á patent for his invention or discovery, nor shall any patent be declared invalid, • by reason of its having been first patented or caused to be patented by the inventor or his legal representatives or assigns in a foreign country, unless the. application for said foreign patent was filed more than seven months prior to the filing of the application in this country, in which case no patent shall be granted in this country.”

It is said that the original form of the section led to certain abuses, in that it made it possible for an inventor to file an application for an American patent many years after the foreign patent was granted, and at any time within two years after an American manufacturer had in good faith commenced the commercial exploitation of the invention.

To correct this situation Congress reduced the period allowed to the foreign patentee for filing his American application, and, since the application had to be filed rather promptly, removed the limitation on the life of the patent.

In the case at bar there was a German patent issued to Just and Hanaman under date of September 8, 1904 (patent in the German Empire from April 15, 1903), while the application for the patent in suit was filed July 6, 1905, or, in other words, “the application for said foreign patent was filed more than seven months prior to the filing of the application in this country.”

The original and amended section 4887 are identical in referring to “by reason of its having been first patented or caused to be patented * • * * in a foreign country.”

*295The point is that both the original and the amended section 4887 were dealing at all times with protection in respect of a patent, the same abroad as here, and thus cases on this point which arose under the old section may be appropriately cited in connection with the present or amended section 4887.

Obviously in the case at bar this is a defense which defendant is called upon to prove. As Judge Coxe said in considering a case which came up under section 4887 in its original form:

“A patent should never he declared Invalid because of the expiration of a foreign patent if there is doubt about the inventions being the same. The burden is upon the defendants, and the doubt'should be resolved in favor oi the patents.” Brush Electric Co. v. Accumulator Co. (C. C.) 47 Fed. 48.

See, also, Brush Electric Co. v. Julien Electric (C. C.) 41 Fed. 679; Westing house v. Stanley, 138 F. 823, 71 C. C. A. 189.

The inquiry, then, is to ascertain what it was that the German patent covered. In what we would call the specification the statement is made:

“The present invention relates to the production of incandescent filaments of pure tungsten or molybdenum metal.”

The claim is for a process, as follows:

“Process for producing incandescent bodies of tungsten or molybdenum, characterized thereby, that a carbon filament is raised to a high temperature by means of current passed therethrough in the vapor of oxyhalogen compounds of tungsten or molybdenum in the presence of a little free hydrogen, whereby the carbon is replaced entirely by tungsten or molybdenum.”

The Just and Hanaman patent, 1,018,502, here in controversy, is for the article itself; i. e., the filament of tungsten in a coherent metallic state and homogeneous throughout. Plaintiff insists that a patent for a process is different from a patent for an apparatus or product. As a general proposition, of course, this is true. Under this statute, however, the real inquiry is noi to be embarrassed by fine points of procedure. In Fireball v. Commercial, 239 U. S. 156, 36 Sup. Ct. 86, 60 L. Ed. 191, Mr. Justice McKenna points out:

“That a process may be independent of the instruments employed or designed to perform it. They may be independent or they may he related.”

I am inclined to think that this case upon which plaintiff' places much emphasis does not go so far as to say that a patent for a process and a patent for an apparatus necessarily represent different inventions in the sense of section 4887.

Where the process is used as the means of creating or developing the product, it is difficult to see how there is any real distinction between the patents for the purposes of section 4887, where one patent claims a process and the other claims a product, but both patents clearly describe the same invention.

If so narrow a construction were placed on section 4887, the result might be the very disadvantage to American business against which this section 4887 was evidently intended to safeguard.

*296What the controversy in this respect comes down to is whether in point of fact, the same invention is patented abroad as is said to be patented in the United States.

In the German patent it will be noted from the quotation supra that the invention related to the production of a pure tungsten filament, and the specification then went on to show how this result can be accomplished. In the American patent the specification also shows how to produce the pure tungsten filament. There is no difference between the aims of each patent. The only difference is that according to our technique the- German patent claims a process while the American patent claims a product or article.

This view, therefore, necessitates further inquiry, namely, whether, as matter of fact, the German patent and the American patent are the same.

[4] Here there must be actual identity. There cannot be introduced into the German patent any after-acquired knowledge or any deviation. The obligation rests upon defendant of clearly satisfying the court that by strictly following the German patent a tungsten filament will be prodüced which will'be tiré same as the tungsten filament disclosed and taught to the art by the American patent.

In the Tungsten Lamp Case Dr. Liebmann, for plaintiff there, testified that the German patent was inoperative. Pro.f. Main, on behalf of defendant in that case, attempted to show the operativeness of the German patent, but the distinguished counsel for defendant LacoPhilips Company, the late Thomas B. Kerr, did not press the point. The foregoing, however, is not binding on this defendant, and must be regarded only as argument.

In this case plaintiff produced Dr. Langmuir as its expert. His high qualifications have already been referred to by this court. 261 Fed. 606.

Defendant produce^ Dr. Scholl, heretofore unknown to this court, but a man of much practical experience in the art, of marked technical equipment, and of infinite patience and industrious application.

Dr. Scholl produced certain lamps, known as Defendants’- Exhibits V-l, W-l, and X-l, and certain unmounted filaments Exhibits G-l and H-l. He asserted that these filaments were made in accordance with the teaching of the German patent and possessed the characteristics of the filaments of the patent in suit.

Dr. Fonda, for plaintiff, who worked in co-operation with Dr. Langmuir, testified that, after repeated attempts to follow the process of the German patent, he could produce only a composite filament; i. e., not the tungsten filament of the patent in suit. The court entertains no doubt of the integrity of Dr. Fonda’s experiments. In this case, as in the Nitro-Tungsten Case, 261 Fed. 606, Dr. Fonda impressed me as thoroughly conscientious and fully equipped to do the work which was assigned to him.- Indeed, one need but see and hear Langmuir and Fonda to be satisfied that no employment nor retainer would deflect either of them from stating what they believed to be the truth.

*297The result of Dr. Fonda’s testimony is illustrated by Plaintiff’s Exhibit No. 36 here inserted.

A composite filament is worthless, as is fully demonstrated in this record. It was the very thing which denoted failure in the prior art and which Just and Hanaman were trying to get away from. There was a great deal of controversy as to the filament Exhibit H-l, the details of which may be found in the record. Finally, the court, in response to plaintiff’s urgent demand, ordered a test, i. e., that a small piece of the filament H-l should be removed and mounted so as to produce a longitudinal section. The- result of this test was to show that the filament H-l was hollow and not solid. Plaintiff’s Exhibits 72 and 73; P. Ex. p. 178; Records, p. 662, Q. 171.

Dr. Scholl had testified that the filament H-l was pure, homogeneous, and solid in the sense, inter alia, that it did not contain any carbon or carbide. This testimony was based on two tests: (1) The potassium nil rite test; and (2) the microscopic test. I am satisfied from Dr. Langmuir’s testimony that the first test is inconclusive and the second negative. There were no life tests, and, in this art, that is Hamlet without the Dane. One of the exhibit lamps, known as lamp No. 10 of Exhibit V—1, was carefully examined by Dr, Langmuir, and microphotographs of a portion of the filament were made by Dr. Jeffries. Plaintiff’s Exhibit No. 80, Figs. B and C. The result of the examination and of comparison with the tungsten filament of the patent in suit is thus summed up by Dr. Langmuir:

“I then compared the characteristics of the lamp before it was broken open with those of pure tungsten filaments of the same geometrical sizes—that is, length and diameter. I did this in two ways: First, by comparison of the characteristics with the published data of the characteristics, which was ob*298tained in our laboratory, as a result of several years’ work, and also compared tbe characteristics of this lamp with a tungsten filament lamp, made up with drawn tungsten wire, having a filament of approximately the same dimensions as the lamp marked No. 10.
“As a result of all these measurements I found that the ratio of hot to cold resistance at approximately one watt per candle was 71 per cent, as great as that of a pure tungsten filament under the same, conditions. The specific resistance of the filament at about one watt per candle was about 130 per cent, of that of a pure tungsten filament under the same conditions, and the total energy radiated in watts per square centimeter of surface from the filament at approximately one watt per candle was 111 per cent, of that of a pure tungsten filament under the same conditions. Finally, the cold specific resistance—that is, figured over to cubit centimeter basis—was 182 per cent, of that of a pure tungsten filament at the same temperature. This temperature was about 28 degrees Centigrade. These results showed that the filament was very far from being a tungsten filament or far from being a substantially pure tungsten filament. Differences in the characteristics between this filament and those of a pure tungsten filament were greater than those of any tungsten lamp that I had ever heard of before, and were such that it would be impossible to rate the lamp or class it as a tungsten filament lamp.”

It is impossible, in view of the large amount of testimony and of the many exhibits, to do more than outline the controversy; but a careful reading of this branch of the case must lead to one of two conclusions, i. e.: (1) That the process of the German patent cannot, as matter of fact, produce the tungsten filament of the patent in suit; or (2) that defendant has failed to prove that it can.

But, even if it were concluded that what has just been stated is erroneous, there is still an insuperable barrier to the success of the defendant’s case. The patent in suit discloses two different processes for the production of the tungsten filament. In each process a composite filament of tungsten and carbon is produced, and the carbon is removed by glowing the filament at a high temperature in an atmosphere of moist hydrogen. This is a step not disclosed in the German patent.

The first process described in the patent in suit involves forming a pasty mixture of carbonaceous material and tungsten, or carbonaceous material and a tungsten compound, into a thread or filament, drying and baking it, and decarbonizing it at high temperature by the use of moist hydrogen. The filament may then be equalized by the deposition process. This first process is the process by which most of the tungsten filaments were made prior to the Coolidge invention of drawn wire.

The second process of the patent in suit (lines 77-103) begins with the formation of a composite filament having a carbon core, just such a filament as results from the process of the German patent except that there appears in the patent in suit, and not in the German patent, the idea of using exceedingly small carbon filaments; filaments which are much smaller than those used in commercial work.

. The German patent contemplates a replacement in situ followed by an equalization. No one knew at that time what the resistance of a tungsten filament would be. There was, therefore, no inducement to use a specially small carbon filament, particularly as the inventors thought that they had found a way of turning that carbon filament *299into a tungsten filament. As is contended by plaintiff, it is therefore obvious that one skilled in the art, in practicing the process, would start with an ordinary carbon filament, would attempt to replace the carbon by tungsten, and, if he thought that he had succeeded in doing it, would deposit more tungsten, to strengthen the filament and equalize it, hut only to the extent to which he thought it needed strengthening and equalization.

• Defendant has started with a carbon filament about Viooo of an inch in diameter, smaller even than the .04 millimeter filament referred to in the patent in suit (.04 mm. equals .0016 inches-—Rec. p. 859); “special filaments”; filaments so small that they cannot be procured in this country, but must be obtained from one Jean Planchón in Paris. “These filaments,” said Dr. Scholl, “here were out of the ordinary as far as diameter was concerned.” Rec. p. 402, X-Qs. 506, 507. They then proceeded, in accordance with the second process of the patent in suit, to build up around this almost microscopic thread a heavy coating of tungsten having a volume 20 times that of the carbon.

In addition to the foregoing, there was the use of a brilliant white heat, of wet hydrogen, and the raising of the temperature of the filament at the point at which the instruction of the German patent is to lower the heat—all departures from the German patent.

Plaintiff’s counsel have concisely summarized the situation. I cannot do better than to adopt, in substance, their resume:

(1) The process of the German patent is inoperative to produce the product of the American patent.

(a) The German patent, starting with a carbon filament, prescribes a tungsten replacement process, followed by a tungsten deposition process.

(b) The replacement does not occur, because the replacement process is entirely inoperative.

(c) Since the carbon is not replaced by tungsten, but remains unaltered, the final result is a carbon filament surrounded by a shell of tungsten.

(d) Such a composite filament is worthless.

(2) The process practiced by Dr. Scholl in producing the filaments which defendant put in evidence involved vital departures from the German patent, such as:

(a) The use of a brilliant white heat, though the German patent specifically directs the use of a “bright red” heat followed by a “red heat.”

(b) The raising of the temperature of the filament at the very point at which the Gennadi patent directs that the temperature be lowered.

(c) The use of wet hydrogen introduced by the “indicating botttle,” the use of a vacuum in the treating bottle, the use, as noted supra, of an abnormally fine carbon filament, covered with tungsten having 20 times the volume of the carbon, though the German patent obviously does not specify the use of any unusual carbon filament and prescribes tho deposition process as a mere means of evening the filament, and the use of an abnormally rapid rate of vaporization of the oxychloride.

*300But again, if I am in error in being so fully convincedin plaintiff’s-.favor in respect to conclusions where the subject-matter is so much a matter of expert controversy at every step and of examinations under the microscope, at least it is true that there are so many doubts which attach to defendant’s case that' it cannot he said that it has proved, as it must, the identity of the patents. That these doubts should exist is natural enough, because there is usually as much human nature in a patent case as in any other litigated cause. Plaintiff has made a great success with its lamps. The exploitation of these patents has vastly widened the commercial field, and the temptation to invade so lucrative a territory is not strange.

Many years have passed since Just and Hanaman made their great contribution in the invention disclosed in their American patent. During all that time no one practiced the process described in the German patent. It is idle not to recognize that after-acquired knowledge is a subtle and subconscious agent of inaccuracy and departure.- The phrase “any one skilled in the art would have known” how to do then what is known how to do now is the danger signal which admonishes courts to be cautious, and, in a case under this statute, to insist that there must be undoubted compliance with the disclosure of the foreign patent before domestic rights may be destroyed. ,

In no case has fuller opportunity been accorded to a defendant and has more time been allowed to make experiments and prepare defenses.; and, yet, notwithstanding the ingenuity and ability with which the Just and Hanaman American patent has been attacked, it remains where it was and where it should be—the embodiment of an impregnable invention of the highest order. The court having on other grounds dismissed the bill as to Fabian, plaintiff may have the usual decree, with costs against the remaining defendants.

Fabian individually is entitled to his costs against plaintiff.

Submit decree on five days’ notice.