The incandescent lamp has always required a conductor through which power can be supplied from the outside world into the interior in such a way that an airtight "seal" is obtained. For the first 30 or so years of lamp manufacture platinum served this role very well but substitutes were frequently sought due to the high cost of that metal. This subject is of interest to delve into and a cursory survey of the literature is presented here.

First, the requirements for a good seal should be known. These can be obtained from E.B. Shand's book, Glass Engineering Handbook. They can be found on Shand's pg 119. These are listed below.

1) The glass should adhere to the metal as well as "wet" it. The degree of wetting is indicated by the contact angle between the surfaces. For example, a bead of water on a waxed kitchen floor has a high angle and doesn't wet well. On the other hand, the contact angle is less when a bead of water rests on an unwaxed floor—it wets better.

2) To avoid cracking, the linear expansion of the metal in the seal should closely match the type of glass being used. There is an exception, however, which will be commented on later, in which it is possible to use very thin metal sections that don't result in glass cracking.

3) The metal used cannot have different structural forms in which it can exist in the temperature range of use.

4) The metal should not have occluded gases in it and the glass should not reach the vaporization stage.

The common leading-in wire in incandescent lamps today consists of "Dumet" in the sealing region. Dumet was a development of Colin Garfield Fink in 1912. He was issued U.S. Patent No 1,498,908 on Jun 24, 1924. A description of his composite wire will be given later.

Articles appeared in the technical journals of the day that described materials that were used in the hope of replacing platinum. Some of these are mentioned below.

1894—Silver Films, Edward Pollard Lamp; The Electrical Engineer, Vol XVII, Jan 17, 1894, pg 55.

1894—Aluminum, Werner Bolton (The Electrical Engineer, Vol XVIII, No 346, Dec 19, 1894, pg 503 and Electrical Review, Vol 26, No 7, Feb 13, 1895).

1898—Iron-Nickel Alloys (The Electrical Engineer, Vol XXV, No 522, May 5, 1898, pg 491).

1899—Nickel-Iron Alloy (Electrical World and Engineer, Aug 19, 1899, pg 279).

1905-Copper, John H. Guest

1907-Copper, G. Calvert

1907-Copper or Iron, W.R. Whitney, Electrical World, Vol XLIX, No 12, pg 583, 1907 and Electrical World, Vol LII, No 19, pg 989, 1908).

The use of iron-nickel alloys failed because of the formation of detrimental oxide films. The copper used by Guest was interesting in that the wire was sealed into the glass in circular shape. No explanation was given in this particular article (American Electrician, Vol XVII, No 4, Apr 1905, pg 219) for that arrangement. Apparently this seal met with some success. See Hammer lamp, 1907-Guest.

Another seal that probably was successful was the one described by Calvert in 1907 (Electrical World, Vol XLIX, No 23, Jun 8, 1907, pg 1166). Calvert used copper from the lamp base to the filament with a portion of the round wire flattened where it was sealed into the glass. It was found that if the thickness of the copper was 0.075 mm no cracking of the seal occurred. Thus, the procedure of increasing the proportion of surface area to mass reduced the strain in the seal area. This technique is used today in tungsten-halogen lamps where the high quartz bulb temperatures require the use of molybdenum as the conductor.

In his book, The Electric-Lamp Industry: Technological Change and Economic Development from 1800 to 1947, Arthur A. Bright, Jr. also reviewed some of the attempts to replace platinum. The review can be found on pages 205-207.

So—what is it that Fink did that resulted in a leading-in wire that is still used today? This is perhaps best told by quoting from Fink's patent. He said:

"In order to make a seal...which will be gas-tight, the leading-in conductor used should preferably have a coefficient of expansion which does not differ greatly from that of the material of the envelope. It is also necessary that there shall be a certain affinity between the material of the envelope, when in a plastic or partially liquid condition, and the surface of the conductor, by reason of which there will be a tendency for the two to stick together. This property is usually spoken of as a "wetting" of the conductor by the material of the envelope. When the envelope consists of ordinary glass, platinum has been found especially suitable for a leading-in conductor because of the fact that it is practically non-oxidizable and has a coefficient of expansion which is nearer to that of the glass than any other metal, and also because platinum is "wet" by glass. Because of the great cost of platinum, however, many efforts have been made to secure a suitable substitute. It has been found that if nickel and iron are alloyed in the proper proportions, a conductor may be obtained which has the same coefficient of expansion as glass. In attempting to use such a conductor in commercial practice, however, difficulty has been experienced because of the large number of seals which are defective. This is probably due to the fact that when subjected to the heat of the flame necessary for the sealing-in operation an oxide coating is formed which is more or less porous and allows air to gradually leak through between the glass and conductor. It has also ben found difficult to prevent the formation of bubbles in the seal due to the liberation of gas from the conductor or the formation of gas by chemical reaction during the sealing-in operation.

"I have discovered, however, that there are certain metals whose oxides are readily soluble in glass at lamp seal making temperatures. Among these are copper and cobalt and some of their alloys. Ordinary lead glass, such as commonly used for incandescent lamps and rectifiers, has a coefficient of expansion of about 8.8x10^(-6) while copper and cobalt have coefficients of expansion of about 17x10^(-6) respectively. Hence, solid conductors composed entirely of either of these metals are not commercially suitable for leading-in conductors because of the great difference between their coefficients of expansion and that of glass. By making a composite conductor, however, having a sheath of copper or cobalt and a core of metal having a lower coefficient of expansion than that of glass, and by properly proportioning the relative thicknesses of the core and sheath, it is possible to provide a conductor, the total radial coefficient of expansion of which is substantially the same as glass with which it is to be used. I have found that if such a conductor is used that there is an even better union in the cold state between the conductor and the glass than in the case of platinum. In actual practice it has been found that by using my invention the number of defective seals has been reduced over 75% from the number commonly found when platinum is employed."