US1140136A - Low-expansion wire. - Google Patents

Description

B. E. ELDRED. LOW EXPANSION WIRIE. APPLICATION FILED DEC 22, 1913.

D sTAsiEs ita-.fiiiniv'l.-` crimen.

BimoN n mmm, or NEW Yann, N. y., AssreNnnmo me .GMEERCIAL nnsnancn comrnmf, or NNW vomi, N. Y., A connommoN or ,NEW Yoan.

mwmNsIoN wm.

specmeammfmemrsatm.

amneamnaiedmmberaa, 1am. sami-m. ausm. f

T all whom tmay. concern:

Be it known :that I, BynoN E'. ELDRED, .a citizen of -thel United States, l.residing .at New York, in thecounty of .New Xork and 5 State of YNew York, have .invented certain new and useful Improvements .-in Low- Expansion Wires,"of which the following A1s a specification.

This invention relates especially 1to leadtion may be appliedalso too her uses, fand it includes as a new article of :manufacture a composite wire ,or Asheet having a core .or

from about 3008159 normaly temperature;`

such composite wire-.advantageously Ahaving sion and -an external sheath or layer of copper or silver; all as .more fully hereinafter described and as claimed.

The invention also includes ,a .method of producing wire and a Wire as 'fully .disclosed andclaimed hereinafter.

."In the manufacture of incandescent lamps it has heretofore beenthe practice ,to' use platinum yleading-in Wires .for conveying .the current from anexterior source to .the .lilament within the lamp tbulb. 'Platinum has been employed .for lthis l.purpose because .ci certain characteristics which ,render it particularly suited for the `purpose and which 46 have beemheretofore regardedaspractically indispensable in a leading-in Wire. .Amon these characteristics may-tbe mentioned (11 its relatively vl'ow coeilicient .of expansion, Awhich is nearer than .thato .aniv otherpure, 50 high melting metal to :the .coe cient-of exl pansion of glass; (2) :its `peculiar.superficial affinity for molten glass whereby -its'vsuna'ce is actually ywetted by molten ,.g'lass; .and (3l) its unoxidizability, which :insures `the 'main- 56 tenance of a clean metallic surface throughing-.in Wires for lamps, .although this .inmena core vof nickel-steel of proper low expanout .the heating operations Einvolved in maki'ngrflampfsils.

' he thermal ,coefficient of aexrpansion and contraction fezf 1platinunniis generally fstated to .be about 0.0000089 to 0.0000091 for each l Cent-Algra@ degree changein temperature. Curve A .in 1the accompanying drawmg `shavvs graphically the rate of .thermal @Patented May 1a, 1915'. r

rexpansion .of platinum, ,absciss representing ,temperatures in degrees gcentigrade, while ordinates Vrepresent ,in millimeters the corresponding .increases in length of a rod 1 meter .lang .at.0 iG. The thermal coeflicient.of glass isalways somewhat'below vthat- -of Aplatinium, :and with .many .types of .glass it .very imuch ,belaw that -of platinum.

' Glass .can :he prcduced .however having at temperatures, say, -belovv7 .100o C. an vexpansigm .as lhigh' Ias 0.0000081, Vrising to 0.0000087 v.to 0.0000088 at temperatures around .300 3C. ,A union canftherefore be formed with ,platinum .atA the softening point o f `such `glass which will in most cases persist. The tendency of .the platinum, which contractsmore Ithan thetglasson cooling, to shrink .away from the glass, is resisted by the mechanical strengtheofthe. union Iformed with *the softened ,glass In the 'cooledlamp the layers of glass T.next the -wireare however under itension and 'this may produce craclirs and ,air-leaks. Much, however, de- `ipeneilson the-thicknessofithe -vv-ire; thin wire beingfsafer .in ythis `respect-than thick.

All high meltlng metals other than platinumlliave .astilhgreater .thermal .rate of expansion .and consequently it .has not been' .feasible to einpnloysuch metalsalone as lead'- .ving-.11n wires, .furthe obvious reason that the .tensile -stlesses between ,such .wires and lthe ,glass ,Wouldbeiso great as to destroy .the-seal. Furthermore, ,the ltendency of metals, such as iron, lto .oxidize .readily especially under the .high `temperature conditions obtaining iin lamp manufacture .has lrendered it practically :impossible `to .keep the surface ofsuch 'leading-n .wires clean.- .With most metals,

poronsmidilayers are formed'and the Iforhas Zbeen prevented Further, sincefplatinum leadln --in .wires -in lractice must '-'be' .madeexcee ingly shim-:bot .for the sake'of economy .and to A'reduce 'fthe strain `in thevmation ofian .aimtight ,joint with theglass ductor.' In my v application No. 790,467 I 110 a glass article in which it is sealed.

have disclosedand claimed a compound wire having a platinum surface, a low-expansion nickel steel core and an intermediate linking layer of copper or silver, the wlre as a whole having an expansion below that of platlnum; and in my application No. 656,987 I have. claimed such a wire in connection 1s type of leading-in wire has proved .eminently satisfactory in practlce; but 1t 1s to be noted that it involves the use of a platinum surfaced leading-in wire. lt has been considered necessary to have platinumv in contact with glass in order to secure a perfectly satisfactory union or seal between the wire and the glass. 'lhe present invention is directed to animprovement over the leading-in -wre specifically claimed in the said applications, whereby the use of the expensive platinum sheath may be done away wlth if certain conditions, hereinafter to be more fully'described, are carefully observed.

In experimenting with leading-in wiresof the general type of said prior applications, that is, with leading-in wires having a regulated coeiicient of expansion below that of the glass into which they are to'be sealed, I have discovered that if the coeHicient of expansion' of the wire be carried sufficiently far below that ofthe glass employed at temv peratures involved in the sealing operation to cause absolute contact by compressive forces, it is not necessary to employ platinum for contacting with the glass, as has heretofore been generally considered indispensable.' Under the conditions just mentioned, the glass during the 4cooling down from the sealing-in temperature .exerts va strong positive pinch or compression on the sealed-in wire, this compression causing the glass to be in extremely intimate contactY with the surface of the leading-in wire, with the result that a tight seal is formed, amply sufficient to maintain the requisite degree of vacuum in the lamp bulb at all times. When the glass is hot, it is, of course, quite plastic" and can shrink into absolute conformity with the wire surface. By choosing ahigh melting, low expansion alloy, such as certain alloys of nickel and iron, a leading-in wire can be made to have a rate of expansion as much less than that of the particular glass in question as maybe desired. rllhese nickel-iron alloys'have av lower expansion than either iron or nickel alone. Although such alloys are of course more or less oxidizable, the 'formation of oxid to a reasonable v extent on a surface of the wire during the formation of the seal does not interfere with the production of an air-tight union with the glass, this being due to the fact thatthe v surface of the joint is so tightlycompressed by the surrounding glass during the sealingin operation as to cause a perfect joint. It

is of course feasible also to perform'the sealmatinee l ing-in operation in-.aninert atmosphere of hydrogen, nitrogenor the like if'it is deemed desirable to prevent oxidation altogether, in which event such a strong compression of the glass on the 'Wire is not required.

In the accompanying illustration in Figure l I have plotted the .expansions of variousmaterials as curves. Fig. 2 shows diagrammatically a compound wire'having a nickel-steel core and a copper sheath. Fig.

v3 illustrates a composite sheet having a layer of nickel-steel united. to a copper layer. Fig. 4 illustrates a modification cornprising an annular sheath of nickel-steel, surrounding a central body of copper.

Whileit is possible to obtain good seals with leading-in wires of nickel-iron alloys, suchv alloys when used alone have certain drawbacks, among which may be mentioned particularlytheir irregular rate of eXpansion through the range of temperatures involved in making lamp seals. This irreguflarity is well illustrated by curve C, from `which it appears-that the expansion curve for nickel-steel alloys of the low expansion types here in consideration is rather irregular and is by no means rectilinear between temperatures of `0" and 325.C., the latter temperature being approximately the highest temperature. involved in making lamp seals, that is, about the temperature at which lamp'glass sets to its hardened state from itsI softened or plastic condition. The particular nickel-steel alloy corresponding to `the curve contains about 38 per cent. nickel. Curve B shows the rate of eXpansion of a typical American lamp glass. ln

` alloy, although somewhat lessuniform of course than' the high expansion sheath. By selecting a nickel-iron alloy of sufiiciently low f' thermal expansion and combining a core of this alloy with a regulating sheath of high expansion 'metal' of the proper relative dimensions ,'the compound leading-in wire as a whole may be'given any combined or average coefficient of expans'on desired, which may be as much less thankthat of the particular type vof glass to be used as may be desired. l Various high melting metals, ezricluding platinum, `may be employed to give this forced or regulated rate of expanlull@ Ummm, I

sion through the range of temperatures involved in lamp making to .the compound leadin' -in wire. For most .pur osos 'however., Igind a regulating sheath o .copper to be most satisfactory .and this is especially desirable for leading-in wires, on account ofthe -high conductivity of copper. Curve D represents the rate of expansion of copper.; and it is to be noted that this rate, thou h relatively rather high, is much more neary uniform than that for nickel-steel. Such 'copper sheath may be'united :to the nickel-iron core integrally as by a Weldunion produced in accordance with the process described in the United States patent to Monnot 853,716. Better results however may be attained in another method of weld-uniting hereinafter more s ecifically described. Where the sheath an core are weld united in this manner, the regulating effect of the copper sheath -in straightening the expansion curve Aof the wire as a whole and forcin ity to approach rectilinearity is most eifectlve. Curve E shows the rate lof expansion of a copper-clad nickel-steel wire such as that just described, and illustrates clearly how the copper sheath renders the expansion rate of the compound wire. as a whole' much more uniform than' that of nickel-steel alone. l

While .the integral union between the cop- I per sheath and its supporting core is a disnot absolutely indispensable,

tinct vadvantage for the reasons specified, I do not desire to be limited to the use of such a bimetallic wire. A union effected by' solderingI or hammering, `hot swaging, etc., may be used but is more liable to be defective and such defects may only be disclosed when the finished lamp is tested. Since however a. leading-in wire under the present invention may be made so as t0 be strongly compressed by the glass into which it -is sealed, during the sealing-in operation the weld union between core and sheath is although it is much to be preferred.

In addition to straighteningv out and rendering. more nearly uniform the expansion rate of the leading-in wire, the copper sheath has the additional function of materially increasing its conductivity. .Nickeliron alloys have a conductivitv relatively low as compared with copper. This second function of the copper sheath is therefore an extremely important. and advantageous one. In this compound wire the copper gives conductivity and athe nickelsteel strength and low expansion; and, in a way, the wire may be regarded as a reinforced, low-expansion copper wire.

Instead of using copper for the sheath; either silver or gold may-be employed, both metals being good conductors and the rates of expansion of both of'these metals being .much more uniform than that of nickel-.iron

coefficient of expansion up alloys, although considerably higher than thatof glass -.or of platinum. Nickel, iron, or :any other high melting metal havinga relatively high but. sufficiently uniform coeiiicxent fedi expansion may also be used for the 1 sheath under some conditions.

class, copper 'gold and silver are however better. Thenickel-steel of the core is not highly conductive and pure metals of the iron class .are not much .better. Use of metals 4of the copper class as one layer is, therefore, much more advantageous.

With glass having a coefficient of expansion of, say, 0.0000087 from common tem- 'peratures'up to 300 C., it is advantageous narily employed in present practice. By4

increasing the lsize of the lamp stem its strength (to withstand stress) may be cor-` respondingly increased, permitting the use of leading-in wires having .much lower coeiicients of expansion than that above mentione'd'; and under some circumstances this may be desirable. Of course where glasses of lower expansion coeiiicients are to be used, the leadinglin wire should have a correspondingly lower coeiiicient of expansion. In this connection, it is tobe understood that by suitably varying the proportions of nickel and iron' in a nickel-iron alloy, an alloy having practically any desired average rate of expansion is obtainable, the upper limits being those of the individual'inetals while the lower limit may be carried down very low and may be made substantiallyzero for temperatures not over 100 C.

In a typical embodiment of the present invention in its most advantageous form, I may use a nickel-iron alloy containing, say, 38 per cent. `nickel and having an average to 100o C. of about 0.0000025. A billet of this alloy may be provided by any suitable method with .an outer layer or sheath of copper. Thish .layer may then be turned down in a lathe to the exact thickness required to correct alloy may be 0.892 inches in diameter, and

after the copper layer is attached` thereto,

Athe assemblage may be turned down to a The highly -conductive metals Aof the copper cient of expansion lying between the expansion curve of the nickel-iron alloy cylinder of 1 inch diameter. In the ln'- ished wire resulting from. drawing down the billet, the total diameter may be, say, v

` sions are to be understood to be illustrative l glass sets only and as capable of considerable' variation.

In producing the article, I find it advantageous to use a vertical bar or core of low expansion nickel steel alloy w1th-a surrounding layer of molten copper in a carbon or graphite mold, causlng the molten copper to solidifyand weld-unite to the core. By this process (which is more specifically described in ,my application No. 539,215, filed Jan. 21, 1910), not only 1s a firm and permanent weld secured, but the article gives a wire which unites better with glass than does a copper-surfaced wlre secured in other ways.

As further explaining curves of the drawing illustrating the above embodiment of the invention, the following data are useful.

Average coeficiente of expansion. Curve B. Glass- Up to 200 C 0.0000030 -Curve E. Compound lamp wire (copperclad nickel-steel wire)- Up ,to 200 C 0.0000057 Up to 300 C 0.0000063 It is to be pnderstood, of course, that the important temperature is that at which the from its plastic condition. This temperature varies somewhat according to the nature of the glass and in general is above 300 C. l

` In a modified form of the invention, I may use a wire comprising an annular sheath of nickel-iron alloy or other suitable high-melting ferrous alloy .surrounding a central body of copper, but for most purposes the reverse arrangement is better. I may also use a leading-in wire having more than two layers of metal. For example, the wire may have a core of' a high-melting ferrous alloy, a sheath of copper, and an exterior layer or coating of silver or gold incasing the compound core of the ferrous alloy and copper. lIt is to be understood therefore that the term core as4 herein employed may refer not merely to a single core of a single metal or alloy, but may refer also to a composite core such as has just been described. The present invention innorman memes'.

cludes ait'wo-layer wire, the core-being one layer and the sheath another; but either layer may be composite. It is also to be noted that a wire may be made with a highmelting alloy either coating or centrali ml coring a body of nickel-steel. The coe -cient of expansion of such alloy is not likely, as a rule, to be as uniform as that of copper, for example; but it isonly necessary that its expansion' be suiciently uni- 75 form to exert a substantial corrective eect on that of the nickel-steel and thus to force a more nearly uniform rate of expansion in the finished wire.n

In referring to nickel-iron alloys, it is to so be understood that this term covers all alloys containing nickel and iron which are suitable for present purposes by reason of their low expansion as compared with either nickel or iron alone and is used as synonye5 mous with the term, nickel-steel.

By reason of the high compressive strains between glass and the like, and a metallic "conductor sealed therein, which may be obtained according to the present invention, seals of this character are in many instances superior to those obtained by using leadingin wires of solid platinum. 'Ihe tensile stresses between platinum and glass, due to the considerably higher rate of expansion and contraction of the platinum, may be entirely eliminated by the present invention,

as has been clearly pointed out. Cn this account, it is now possible to substitute for platinum, which is so expensive and which is not wholly satisfactory, leading-in conductors having base metal or alloy surfaces;

or having cores =of base metals or alloys pro-` vided with external sheaths, of silver, gold or other non-platinum metal, much cheaper' 105 than platinum itself, the wire as a-whole also being a better conductor than platinum and better suited tomodern high capacity lamps.

While 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 experimental error, especially where, as in the present instance, heat measurements are involved. 'Ihe curves are therefore to be considered as merely indicative of the probable comparative conduct of the materials in question, and of typical examples examined, and not as absolutely accurate.

While, as stated, I may use other combinations of metals, the best embodiment of the invention comprises a copper-coated, nickel steel. The copper notl only has the function ofy correcting the c urve of expan- 125 sion but the greater further advantage of given a relatively high conductivity to the composite wire. Gold-surfaced and silversurfaced wires are good conductors and less oxidizable; but the copper-surfaced wires 13 tures such as lead, tin,

course precluded by the heat necessary 1nsuited for my of alloys of metals melting at low temperaantimony, etc., is of sealing wire through glass.

While advantageously the expansion of the wire vis as above pointed out, less than that of platinum or of the glass with which it is to be used, yetv in another aspect the present invention may be said to include a4 method of reducing the expansion of highly conductive metals by uniting them with alloys of-low 'expansion to produce a compounded reduced expansion; in the provision of a highly conductive wire of reduced,

expansion.

What I claim is 1. A composite low expansion wire comprising, a core of nickel steel and an external copper sheath welded thereto, said wire as a whole having less'expansion than platinum.

2. A composite low Aexpansion wire comprising a core ofnickel steel and an external sheath of a metal of the copper class welded thereto, said wire as a whole having less expansion than platinum.

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

4. As a new article of manufacture, a composite wire having a surface of base lOther low-expanding alloys and metals may be used in lieu of nickel steel, but the latter is best. The use metal and havinga rate of expansion as a whole below that of platinum.

As a new article of manufacture, a wire having a surface of metal of the copper class and a rate ofexpansion as a whole below that of platinum.

6. A 2-layer -composite wire, one v'such y layer being of low-expansion nickel steel and the other layer of high-expansion highmelting metal, the wire as a whole'having an expansion less than that of platinum. Asa lamp wire, reinforced copper surface Wire, said reinforcement consisting of an interior layer of low expansion nickeliron alloy.

v8. As a lamp wire, arcomposite wire composed of a layer of copper and'another layer of low expansion nickel-steel. in suflicient amount to reduce lthe total-expansion sufficiently to secure a seal with lamp glass.

9. A leading-in wire comprising an outer sheath of high-melting, high-conductive material united to a core of nickel-steel having an average coefficient of expansion distinctly below that of the sheath, the core being under compression by the peratures to which the'wire is subjected in lamp-making use.

In testimony whereof, I afiix'my signature in the presence of two subscribing witnesses.

NED J. WHELAN, W. .-S. HOWELL.

sheath at all tem-v

Images