US1197615A

US1197615A - Compound metal.

Info

Publication number: US1197615A
Application number: US790467A
Authority: US
Inventors: Byron E Eldred
Original assignee: General Electric Co
Current assignee: General Electric Co
Priority date: 1911-10-26
Filing date: 1913-09-18
Publication date: 1916-09-12
Anticipated expiration: 1933-09-12

Description

UNITED STATES PATENT OFFICE.

BYRON E. ELDRED, OF NEW YORK, N. Y., ASSIGNOR, BY MESNE ASSIGNMENTS, TO

GENERAL ELECTRIC COMPANY, A CORPORATION OF NEW YORK.

coMPoUNiJ METAL.

Specification of Letters Patent.

Patented sep 12, 1916.

Original application filed October 26, 1911, Serial No. 656,987. Divided and this application filed September 18, 1913. Serial No. 790,467.

To all whom it'- may concern Be it known that I, BYRON E. ELDRED, a citizen of the United States, residing at New York, in the county of New York and State of New York, have invented certain new and useful Improvements in Compound Metals, of which the following is a specification.

This invention relates to compound metals; and it comprises a composite wire or conductor of two or more metals having compensating rates of expansion and having a. core of metal of a low rate of expansion at moderate temperatures (said core advantageously being of nickel-steel), firmly and permanently united by a metallic union to a sheath or sheaths of one or more high melting metals having a substantially uniform and higher rate of expansion, such as copper, gold, silver and platinum, such composite wire, as a whole, having a low and approximately uniform'rate of expansion; as for example, a composite wire having a nickel steel core of low expansion, an annular layer of copper welded thereto and an annular layer of platinum welded to the copper, the composition of the core, and, consequently, its expansion, and the thicknesses of the sheath layer or layers being so correlated that the wire, as auwhole, has a low, substantially uniform rate of expansion, such rate of expansion being less than that of platinum or less than about .00000899 all as more fully hereinafter set forth and as claimed.

One of the most important uses of the novel wire or conductor of the present invention is for leading-in wires in electric lamps; such wires being sealed in orthrough the walls of the glass bulb of the lamp to afford communication between the filament or other incandescing member within the bulb and an outside source of current supply. The use of my novel conductor in this connection will therefore be referred to at some length in the following description. It is to be understood however that the novel wire or conductor per se is the essential subject matter of the present application, no claim being made herein to the combination 26, 1911, of which the present application is a division.

Electric lamps and, the like are generally provided with leading-in wires of platinum sealed in or through their walls; platinum having the property of making a gas-tight union with fused and softened glass. Platinum expands and contracts about .00000899 to .0000091 for each centigrade degree change in temperature, and as glass at temperatures below, say, 100 C. can be obtained of an expansion as high as 0.00000810, (rising to 0.0000087 or 0.0000088 at temperatures around 300 C.) a union can be formed with platinum at the softening point of such glass which in most cases will persist on cooling. The tendency of the platinum, which contracts more on cooling, to shrink away from the glass as its diameter lessens is resisted by the mechanical strength of the union formed with the softened glass. In the cooled lamp the layers of glass next thewire are however under'tension and this may produce cracks and air leaks.

It has not hitherto proved practicable to substitute any other metals for platinum since all the other high melting metals have a still greater rate of expansion and none of them tends to form the same physical union with glass. Certain alloys, such as nickel steel may be produced of various low rates of expansion; and it has been proposed idea. that the wire shall, as a whole, have substantially the same rate of expansion as platinum. Platinum, as stated, has a rate of expansion which is above that of the glasses, there being, as a matter of fact, no commercial glass with an expansion as high as that of platinum. The best glasses now in use for electric lamps show expansions at temperatures up to 100 C. of about .0000081. At temperatures above 100 (3., the rate of expansion increases somewhat and at 300 (3., it may be as high as .0000087 to .0000089; which is still a little below that of platinum.

The accompanying diagrammatic drawings will serve to make clearer the character of the present invention.

In these drawings Figures 1 and 2 show expansion curves of certain lamp materials; Fig. 3 'is'a large scale cross-section of the new wire; Fig. 4 represents a mass of glass, such as a lamp stem, with'the new wire sealed therein; Fig. 5 represents a core or interior layer of the low expansion nickel iron alloy with an outer sheath or layer of platinum directly united thereto; and Fig. 6 represents a composite wire having an inner core of copper surrounded by an annular layer of a low expansion nickel iron alloy, this being again surrounded by an outer, sheath of platinum.

With the exception of glass, copper, nickel-steel, platinum, and the present composite wire, respectively, as the temperature increases from 0 to 300 0., longitudinal expansion being shown for the sake of convenience. Curve C, 'Fig. 1, shows the lengthening of a wire of the present invention having a length at 0 C. such that at 300 C. its length is equal to that of a rod of lamp glass also at 300 C. whose.

length at 0 C. is one meter.-

Ourve A (Fig. 1) shows the rates of expansion at different temperatures for glass such as is commonly used for incandescent lamps, while curve B (Figs. 1 and 2) is a similar curve for platinum. Lamp glasses of the type commonly used soften at temperatures varying from about 320 to 350 0., and while the expansion of glass above this temperature increases in rate, that is, of

. course, not material since it is only the rate of expansion of rigid, not fluent, glass which is important in the present connection. At

the hardening or setting temperature, say

320 to 350, the expansion of these glasses is sufficiently near that of platinum to insure a good joint, but in cooling to ordinary temperatures, the platinum tends to shrink away more or less from the glass, its rate of contraction being greater. This tendency to pull away is compensated to some extent, as stated, by the natural adhesion between glass and the platinum; but is responsible curve C, Fig. 1,- the curves show graphically the lengthening. of bars or wires one meter long composed of r for a certain number of lost lamps, since occasionally the junction between the glass and the platinum, which is under strain, yields and allows the platinum to pull away from the glass. In the use of heavy wire for series lamps. this diflicultyis experienced to a greater extent than in making lamps for smaller currents, because the absolute amount of shrinkage occurring between platinum and glass is greater for larger sizes of wire. Using the wire of the present invention, this difiiculty is avoided since compression or pinch is substituted for tension.

I have discovered that I can produce a composite wire which is particularly sui-tparticular glass to be employed; something .which is, of course, impossible with a platinum wire. I employ a composite wire havmg a core of nickel-steel .or other low-expansion high-melting alloy or metal, andan outer layer or sheath of platinum. The core itself may be composite as will appear hereinafter. Low-expansion high-melting ferrous alloys are particularly advantageous for the core of the present wire. In other words, the composite wire of the present invention may comprise an outer sheath of'a high-melting, substantially non-oxidizable metal, such as platinum, covering and internally supported by a high-melting metallic body of lower thermal expansion than the glass into which the wire is sealed.

Nickel-steel can be produced which, at temperatures under 100 0., has almost any rate of expansion desired; even to substantially no'expansion. For my present purposes I prefer nickel-steel which shall have an average rate of expansion, at temperatures up to 100 (1., of approximately 0.0000025; a rate of course very much less than that of platinum. Nickel-steel containing about 38 per cent. of nickel answers the purpose. very'well. The rate of expansion of alloys such as this is not constant however. WVhile there is a fairly constant rate of expansion from ordinary temperatures up to 160 or 170 C. yet-above this point, the rate of expansion rapidly increases;

I there is a rather pronounced upward turn in the curve of expansion, at about 200 0., as shown in curve D,Fig. 2, where, for convenience, he vertical scale is one-half that of Fig. 1. In other words, an alloy which, at

' copper.

ordinary temperatures would have the same rate of expansion as glass, would have forthe purposes of the present invention, too great an expansion at sealing temperatures. It is, therefore, necessary for the stated reason, and for other reasons later appearing,

ordinarily to choose an alloy having, at ormately as follows:

But as the seal produced at a high temperature must persist at all lower temperatures and since there must not be too great a. compressive stress or strain in the union of i the leading-in wire and the glass in the finished lamp, it is desirable to adjust the rate of expansion at lower temperatures. To this end, I may provide the nickel-steel with a metallically united expansion-regulating layer of copper, silver, gold or platinum, or their alloys. All these metals, with the exception of platinum, have a relatively high rate of expansion, but this rate of expansion is not far from uniform or rectilinear at all temperatures here important. Curve E, Fig. 2 is the expansion curve for The union between this annular sheath or layer and the nickel-steel core must be absolute, and in the nature of a weld union since otherwise, where such copper layer is employed, the wire would be useless for lamp purposes. Any seam or crevice between the metals will lead to leakage.

As silver and copper do not form a thoroughly satisfactory type 'of 'wetting union with molten or softened glass which is desirable in a sealed union, whether this be because of oxidation or because of lack of physical affinity, it is next best to provide the composite wire with a sheath of a metal with which glass will form such a wetting union. In practice, platinum is best adapted for this purpose. This outer layer of platinum must be united to the copper, with a true metallic union, free of all flaws, seams or defects; the two metals must be in absolute metallic union at all points between abutting surfaces. Of course, where the outer sheath of platinum is relied upon to straighten the expansion curve of the wire as a whole, the intermediate copper layer may be omitted; its use is preferred, however.

The low expansion core allows the Wire, as a whole, to have the desired low expansion; the intermediate sheathinglayer of copper or the like metal, where such intermediate layer is employed, corrects and makes substantially uniform the rate of expansion of the wire as a whole while materially raising it above that of the core alone; and the outside layer of platinum provides for the necessary physical union with the glass. Each layer has its own function in producing a wire of the desired characteristics; but the characteristics of each layer are modified by reason of the presence and integral union of the various layers. The expansion of the core must be low enough to compensate not only for the higher expansion of the expansion-regulat-' ing layer of copper, but also for that of the union-securing platinum. All three layers must be in absolute metallic union with each other, or welded; and in practice this can be secured by causing the intermediate layer of copper or the like to become fluid or molten during the process of manufacture, producing a wetting union. By

proceeding in this way a true metallic union or weld between the layers can be brought about without using pressure. All layers should be free of low-melting metals such as tin, lead and zinc to prevent injury to the platinum'in making the wire and in sealing it in place.

The curve of expansion of a typical sample of wire suitable for use for the present purposes is shown in Fig. 1, curve C and Fig. 2, curve C. It will be seen that the expansion of the wireis not far from uniform. The average coefficients of ex pansion per degree for this particular sample over certain temperature ranges are as follows:

It is possible to make this wire with a high-temperature coefficient as low as .0000050, but it is found inpractice that it is not advisable with ordinary glasses, and with the size. lamp stems now generally employed, to go much below .0000070. By using lamp stems sufficiently thick to withstand the'compressive strains set up be tween the glass and the sealed-in wire, a wire having a much lower rate of expansion than the glass of the stem can be used with entire success.

With glass having an expansion coefficient of.0000081 at temperatures up to 100 C., I find that-a wire having a coefficient between .0000068 and .0000072 at sealitn g temperatures gives particularly satisfactory re sults. With a wire of very low coefficient of expansion, there is such a strong contraction and compressionof the glass upon the wire that there is a tendency to produce cracked stems, though the comparalamp stem with its considerable elasticity is well able to withstand a reasonable strain. l/Vith heavy wires such as are used for series lamps, it is not ordinarily practicable to go even as low as .0000070 with the coefficient unless the mass of glass in the stem is increased.

It should be noted that the intermediate layer of copper in the specific embodiment of the invention above described has a three fold function. Mechanically it enables a weld to be made between the platinum and the nickel steel; it also functions in partially straightening the curve of expansion of the nickel steel and giving a wire of a more nearly constant rate of expansion; and, finally, it increases the conductivity of the wire as a whole. In this specific embodiment of the present invention the wire has a central core of low expansion metal such as nickel-steel, a surrounding expansion-correcting layer of copper or the like, and an exterior layer ofplatinum. The wire as a whole must have a rate of expansion less than that of platinum where the usual lamp glass is used, or, in general, less thanthat of the particular glass which is to be used for the lamps. Advantageously, this expansion will be quite substantially less than that of the glass to be used. The copper and the platinum should, each form a layer of substantial thickness as compared with the diameter of the wire in question. In practice the platinum layer may be thinner than the copper layer. With sufficiently thick layers of the two metals, a substantially uniform expansion of the wire as a whole, is forced; and this expansion will be substantially the same for a given change of temperature not only below 100 C. but above 100 C. up to 325 C. With a wire of this typehaving an expansion less than that of glass, a novel-type of union is produced in sealing-in. As explained, with a platinum leadlng-in wire, the surrounding cohering layers of glass on cooling are under tension. WVith the present wire, the hot fluent glass adjusts itself to the platinum in scaling in, and as the assemblage cools, the wire contracts less rapidly than the glass so that the glass around the wire is, so to speak, shrunk on, forcing an intimate contact in the first stages of cooling while .the glass is stillplastic, and later producing compression. A double advantage is thus secured.

The platinum surface of the composite wire gives the ideal union with glass characteristic of platinum. But while in the case of an all-platinum wire the adhesive union with glass in the finished lamp is, as before pointed out, opposed by the tensile strain due to the higher expansion and contraction coefficient of platinum, with the present composite platinum-surfaced wire sealed into glass, on the contrary, the platininn-glas's adtice that the percentage of imperfect lamps is reduced considerably. But as before stated, for ordinary purposes the composite wire should'not be designed for such low'expansion coefiicient as to give an excessive compression between the. glass stem and the wire, unless a sufiiciently large stem be used,

as before explained. Extremely low coefficient wire may be used for special purposes of course, as for example in the manufacture of mercury lamps, vapor rectifiers and the like which are preferably made of an extremely tough glass of very low expansion coefficient. As additional evidence of the importance of having a coefficient of expansion for a leading-in wire lower than that of glass the fact may be mentioned that the.

composite copper-nickel-steel core of the present wire has, without any coating of platinum, been drawn into lamp wire and has afforded seals. With such a copper-clad ferrous alloy wire, the sealing-in-operation is best carried on under non-oxidizing conditions, as for example in an atmosphere of nitrogen or other inert gas. Three lamps out of five have given perfect seals where the usual volume of glass is employed in the stem; that is, the size stem customarily used with platinum seals. With larger stems av much larger proportion of perfect seals can be obtained, failures being reduced to a negligible minimum. Here the character of the seal depends, not so much upon any adhesion of the glass for the copper, as the.

surface of the latter metal may oxidize considerably in the process of sealing-in, but mostly upon the strong pinch or contraction of the glass upon the wire. For the present purposes however a welded-on outer sheathing of platinum is generally desirable and preferable.

In a typical embodiment of my invention, a nickel-steel billet of any desired composition, say one which will have an expansion of about 0.0000025 may be provided-with a welded-on copper layer by any suitable method, as by' that described in patent 853,716 or in my co-pending application Serial No. 539,245, filed Jan. 21, 1910, and this layer turned down in a lathe to the exact relative thickness required to correct the expansion curve and then inserted into a closely fitting platinum tube or thimble. On

now heating to the melting point of copper,

the three layers become metallically united.

The billet may next be drawn or swaged to wire and annealed. In practice the nickelsteel rod may be 0.892 inches in diameter, and after the copper layer is welded on, the assemblage may be turned down to a oneinch cylinder. The platinum thimble em.- ployed may have a thickness of 0.096 inch. In a finished wire of say 0.006 to 0.007 inch diameter, the thickness of the platinum may advantageously be about .0005 inch. The foregoing dimensions are to be understood as illustrative only and capable of consider able variation.

In a modification, I may use a composite core comprising an annular sheath of nickelsteel or the like surrounding a central body of copper. But for most purposes the reverse arrangement is better. It is to be noted that the term core as used in the present specification and claims is intended to cover either a simple or a composite core; the core being that portion of the wire inclosed by an outeror surface sheath.

\Vhile the expansion curves shown in the drawings represent data obtained by careful experiments, it is to be understood that such data are always subject to more or less .ex perimental error, especially where, as in the present instance, heat measurements are involved. The curves are therefore to be considered as merely indicative of the probable comparative conduct of the materials in question, and of typical samples examined, and not as absolutely accurate.

\Vhat I claim is 1. .As a new material a composite wire comprising a core of nickel steel, a layer of another metal of high expansion welded thereto and an exterior sheath of platinum welded to said layer, said wire as a whole having a lower coefficient of expansion than platinum.

2. As a new material a composite wire comprising a core of nickel-steel, a layer comprising copper welded thereto and an exterior sheath of platinum welded to said layer, said wire as a whole having a lower coefiicient of expansion than platinum.

3. As a new article of manufacture, a platinum surfaced wire having a rate of expansion below that of platinum.

4:. As a new article of manufacture, a wire comprising a shell of platinum and a core of high-melting metal united thereto by a seamless metallic union at all points therebetween, said wire as a whole having a lower rate of expansion than platinum.

5. As a new article of manufacture, a wire comprising a shell ofplatinum and a core of nickel-steel united thereto by a seamless metallic union at all points therebetween, said wire as a whole having a lower rate of expansion than platinum.

6. As a new article of manufacture, a wire comprising a core of low-expansion high melting alloy, an annular linking layer of another high-melting high-expansion metal weld-united thereto and an exterior sheath layer of platinum weld-united to the linking layer, said wire as a whole having a lower rate of expansion than platinum.

7. As a new article of manufacture, a Wire comprising a core of low-expansion nickelsteel, an annular linking layer of copper weld-united thereto and an exterior sheath layer of platinum weld-united to the copper, said wire as a whole having a lower rate of expansion than platinum.

8. As a new article, a composite wire comprising a core of a metal of low expansion and of a varying rate of expansion at different temperatures, and at least one sheath of metal of higher and substantially uniform rate of expansion, said sheath being metallically united to said core and of sufficient thickness to force a substantially uniform rate of expansion for the wire as a whole, said rate of expansion being less than that of platinum.

9. As a new article of manufacture, a wire comprising a core of a low-expansion highmelting ferrous alloy, a platinum sheath, and a linking annular layer of a high-expansion non-ferrous metal of high-melting point uniting said core and said sheath, said wire as a whole having a coeflicient of expansion below that of platinum.

10. As a new article of manufacture, a wire comprising a core of a low-expansion highmelting ferrous alloy, a platinum sheath, and a linking annular layer of copper uniting said core and said sheath, said wire as a whole having a coefficient of expansion below that 'of platinum.

11. As a new article of manufacture, a composite wire having a platinum surface layer integral with a high melting metallic core, said wire as a whole having a rate of expansion below that of platinum.

12. As a new article of manufacture, a'

composite wire comprising a compound core of nickel-steel and copper, and an inclosing sheath of platinum, said wire as a whole having a lower rate of expansion than platinum.

13. As anew article of manufacture, a composite wire comprising a compound core of a high melting metal of low expansion and a high melting metal of higher expansion, and an inclosing sheath of platinum,

compound metal article comprising a core ture in the presence of two subscribing Witof felrrous nlletall1 and a. Eheath of a noble nesses. meta metal ion y unite thereto at al w 1 w points, said compound metal article as a BYRON LLDRLD' 5 whole having a lower rate of expansion than \Vitnesses:

platinum. K. P. McELnoY,

In testimony whereof, I aflix my signa- W. S. HOWELL.