US1670292A - Tungsten filament and method of making it - Google Patents

Description

F. BLAU ETAL TUNGSTEN FILAHENT AND METHOD OF MAKING IT Filed Nov. 29, 1924 Inventors Fritz Blau, Fritz Ker-e Kurt Moer-s,

b Their- Attorney Patented May 1928.

UNITED STATES PATENT OFFICE.

IBI'IZ BLAU, FRITZ KOBEF, AND KUR'I MOERS, OF' CHARLOTTENBURG, GERMANY, ASSIGIOBS T0 GENERAL ELECTRIC COMPANY, A CORPORATION OF NEW YORK.

TUNGSTEN IILAMENT AND METHOD OF MAKING IT.

Application filed November 29, 1924, Serial No. 752,968, and in Germany February 14, 1924;

Our invention relates to metal filaments and in particular to filaments such as tungsten filaments. The present invention contemplates a filament of the macro-crystalline form adapted for use as a light source for electric lamps. The filament of the present invention differs from filaments heretofore used in that whereas the filament is inherently of a character such that the macrocrystalline form is not readil assumed nevertheless this form is induced b ing or deforming the metal at a number of points along the filament, as for example,

y sharp bends of small critical radius of curvature and thereby providing nuclei of crystallization in the filament which promote such macro crystallization. The filament further differs in that the desired strain or deformation to produce along the filament many such nuclei of crystallization may be obtained by just coiling the filament into small coils so that the coiled filament may be coiled again and still preserve its form without sagging.

In the art of manufacturing incandescent electric lamps there are several reasons which make it desirable to obtain a filament of especially compact or compendious form and small enveloping length, for it will be understood that this makes for simplified anchoring and supporting, concentration of light, and increase in efliciency in the case of gas filled lamps. It has become customary for this reason, whenever the point is to reduce the length of the filament, as is true, for

example, of gas-filled lamps, to wind the filament in a spiral or helix fashion, and to choosetthe core of the helix as large as the stability of shape of the wire would permit. This reduction in length, ,i. e., the ratio between the stretched filament and the-envelope is important, e. g., in veryslender filament wires, since in this case the yield of heat to the ambient as assumes comparatively high values. he spirally helically wound tungsten filaments which consist of long non-bent crystals have been found to be particularly non-deformable.

However, even in these filaments the diameter of the mandrel or core on which the filament wire is coiled can be decreased only in a slight measure. This is particularly true of coiled filaments consisting of very thin wires, say, wires having a diameter less than about 0.025 mm., in the case of which y strain-" the production of Ion the core. diameter could heretofore be successfully diminished only to about five times the wire diameter since upon further de creasing the core, the coiled filaments become unstable and show a tendency to sag.

In the filaments according to the invention these drawbacks are obviated. It has beentfound that the cause underlying such trouble as has arisen was due to the fact'that the formation of long unbent crystals by the aid of the well known heat treatment does not occur with sufiicient reliability if the be produced by impact or pressure actions.

It is convenient to make these points of deformation by making sharp bends of great curvature, the latter alternating with places of less curvature. These points of deformation or of sharp curvature, in the heat treatment of the completely wound filament, act

as nuclei of crystallization and thus romote the production of macro-crysta ine growth, which begins at said nuclei and propagates to the neighboring non-deformed or insufiiciently deformed portions 'or the portions of less curvature. The provision of nuclei at adequately short intervals in-- sures macrorcrystallization of the metal and crystals throughout the entire filament, t e individual macrocrystals being joined without a joint or gap and mostly extending over a relatively great number of nuclei or a plurality of turns of the helix. a fact that can be confirmed by superficial etching and microscopic examination. In manufacturing filaments according t the invention, the most convenient plan is to start from a drawn tungsten wire, preferably one of the kind with an addition of diflicultly fusible oxides such as thoria, which is continuously heated until incipient loss of fibrous structure, and which is thereupon subjected to one or two passes through v are drawing dies which reduce its diameter slightly. In this starting wire, according to this invention, points of deformation or strain are produced at short intervals, for instance by passing a fluted roller over the wire. or b passing the wire throu h a small swaging machine. The wire is thereupon wrapped around a cylindrical mandrel or core having a diameter of, say, 8 to 10 times the wire diameter. However, the said points of deformation or strain can be produced in the wire also during or after the winding of the coil.

According to another very suitable form of construction the original wire is wound upon a core so shaped that in the turns of the wire between points of less curvature, points of greater curvature as deformation points are produced, for instance, in a way as illustrated in Figs. 1 and 2, upon a square or hexagonal core or mandrel. The wire is thus bent around the corners of the core with a sharp curvature, while at the surfaces of the core the wire becomes of less curvature or may even be entirely straight. But the core may also be made of a different shape; for instance, as depicted in Figs. 3 and 4, it may present a sharp curvature at onlyone place, while at all other parts it is practically semi-circular in shape. In lieu of cores of such cross-section, the core could.

also consist of two or more wires of like diameter, as in Figs. 5 and 7, or of different diameter as in Fig. 6. The wires, of different sizes, constltuting the core may be twisted or stranded together in a way as illustrated in Fig. 8, and. the filament wire then wound on this core, as illustrated in Fig. 6.

In order to produce in the turns of the wire several places of different curvature, the starting wire could also be wound first upon a round mandrel and then the ensuin helix or coil is flattened with or .without t e aid of a mandrel thereby producing a number of spaced bends of much sharper curvature than the remainder of the wire.

After being thus formed, the filament is subjected to the well known heat treatment of heating it to the temperature of rapid grain growth far beyond 2000 C. whereupon the macro-crystalline growth occurs and the continuously joined long crystals roduced. As regards their crystallograp ic structure, the latter are unbent also at the previously deformed points since the said crystals grow without being influenced by the geometrical shape of the body. The heating of the filament can be effected either by an electric current passed therethrough or by externally applied heat, for instance, in a suitable tubular furnace.

The removal of the mandrel or core can take place either before the heat treatment, or else after completion of the same. In

this latter case, the mandrel, of course, must consist of a refractory material such as tung sten. Another scheme would be to heat the filament jointly with its core toaround 1900 or 2000 C., to thereupon remove the mandrel, and to finally heat the filament alone, without a core ormandrel, to the final temperature. The heating of the filament after removal of the core, if desired, could also take place after having mounted the filament upon the stem of the incandescent lamp.

If a section mandrel is removed prior to the heating of the filament, for instance, by the action of acid, in which case brass or steel wire is preferably employed. the coil will be caused to turn back somewhat as soon as it is no longer held by the core, with the result that the number of turns is small er after removal of the mandrel. For example, in the case of a coil whose core, as shown in Fig. 5, consisted of two contiguous brass wires (wire diameter 0.023 mm., a diameter of the core wires 0.09 mm.), it was observed that the number of turns was reduced one-fourteenth (1/14th), and that the points of curvature in each turn were displaced with reference to the corresponding points of curvature of the neighboring turn by 1/l4th of the circumference of the circle. However, this strange behavior is of no importance in the completion and the use of the filament.

Since in the making of the new filaments the diameter of the core or mandrel can be quite considerably increased without the crystallizing powers of the wire material being affected thereby, the length of the envelope of the filament can be diminished, for instance, to one-third of what has heretofore been customary, or even less.

What we claim as new and desire to secure by Letters Patent of the United States, is,-

1. The process of preparing a filament from tungsten that does not readily assume the macro-crystalline form when heated, which consists in producing at a plurality of )oints spaced along; the filament strains su cient to form nuclei of crystallization to promote macro-crystallization and then heat;- ing to cause macro-crystalline growth to set in.

2. The process of preparing a tungsten filament. which consists in straining the metal of the filament sufficiently at. a number of points spaced along the filament to create a number of nuclei of crystallization for starting macro-crystalline growth when the filament is heated and heating the filament to a temperature of over 1900 C. to cause macro-crystallization growth to set in at said nuclei.

3. The process of com'erting a tungsten wire into macro-crystalline form in a coil with turns havinga mean radius of curva- Ill produced in the wire and then causing macro-crystalline growth. to set in atsaid points by heating said wire to approximately 2OUO C.

4. The process of converting a tungsten wire into macro-crystalline form ina coil having turns of a mean radius of curvature.

so great that macro-crystallization does not readily set in, which consists in producing in the wire at points spaced along said turns strains sufliciently great to induce macroerystalline formation when said coiled wire is heated and then heating said wire to cause macro-crystalline growth to set in from said points. 4

5. The process of preparing a metal filament, which consists inbending the filament into a coil having of curvature considerably reater than that which causes macro crystalline growth to take place readily and also bending said filament at, a plurality of points spaced alongv its turns of a mean radiuslength into sharp bends of a radius of curvature sufliciently less than said mean radius of curvature to produce strains which induce macro-crystalline growth at each of said sharp bends when the filament is heated and thereafter heating the vfilament to cause macro-crystals to grow at each of said sharp bends and toward each other.

6. A coiled tungsten filament having a plurality of bends each with a radius of curvature considerably less than the mean radius of curvature of the turns of the coils and spaced along the length of the filament to induce macro-crystalline growth when heated.

7. The process'of converting a tungsten wire into macrocrystalline form which consists in bending it sharply on a curve ofsuch radius that strains are produced which-induce the glOWtll of macro crystals and then' heatingto the temperature of rapid grain growth.

In wltness whereof, we have hereunto set our hands this 12th day of November, 1924.

FRITZ BLAU." FRITZ 'KOREF. KURT MOERS.

Images