An incandescent lamp must be hermetically sealed and yet must have current led through its walls to the filament. From the beginning this has been effected by sealing two platinum wires through ,the red-hot glass, all other methods having been quickly discarded. Lately, however, sundry substitutes for all-platinum seals invented by Byron E. Eldred have, to all intents and purposes, displaced the use of platinum. These seals have proved not only as good as platinum, but better.
In sealing a wire through glass two things come into play, one being the adhesion of the metal and the glass, or what is termed the " wetting " of the metat by the glass, and the other the relative expansion and contraction of the metal and the glass. Softened or fluid glass " wets " platinum readily. This may be in part due to the specific physical affinity of the molecules of glass for the molecules of platinum, and may in part be due to the fact that the platinum maintains a metallic surface during the sealing operation.
The expansion of glass, however, is somewhat more than platinum, even with the soft glasses which are often used for lamps. The difference is not great, but it exists. The net result in cooling a platinum glass seal from a high tempera.ture to a lower temperature is that the platinum tends to shrink away from the glass. This shrinkage is not great, but it is responsible for a little strain--a strain which is resisted by the adhesion of the glass and the platinum. Mr. Eldred conceived the idea of doing away with this condition of tension by making a wire whose expansion was a little less, but not much less, than that of the glass to which it was to be sealed. With a wire of this kind, on sealing and cooling, ,the glass shrinks down on the wire and there is a little compression in the seal. The amount of this compression must not be great, since otherwise dangerous strains might exist, but a little compression there should be. He devised a type.of wire having a core of nickel-steel of very low rate of expansion, a jacket of copper on the core, and a further jacket of platinum on this copper sheath. The composition of the nickel-steel was so chosen that its own expansion, averaged with that of the copper and the platinum, gave the wire, as a whole, a little less expansion than that of the glass, so that, in sealing, the desired shrink-on effect or compression seal could be attained. The function of the copper in this combination was not only to give a greater electric conductivity-something much needed in these small-gauge leading-in wires--but also to make more regular the expansion of the nickel-steel. While nickel-iron alloys can be made to have any expansion within a certain range that may be desired, this expansion is not regular through the range of temperature incident to the sealing-in. The copper serves to make this curve of expansion more regular.