"A plan to replace platinum by aluminum, to the use of which metal objections based on rarity are not now ever likely to apply, has been suggested by Werner Bolton, who exhibited lamps made with the substitute at the recent meeting of the German Electro-chemical society in Berlin. The coefficient of expansion of aluminum is much larger than that of glass, and direct sealing in is impracticable. The joint between glass and metal is, however, made in the following ingenious manner:

The aluminum is heated in a glass tube until it fuses and the envelope of glass fuses round it. When the composite rod (having a glass envelope and an aluminum core) cools, the glass cracks; but this is of no moment, as the function of the glass is merely to prevent the aluminum melting when it comes to be sealed into the thick glass envelope which ultimately forms the base of the lamp. The arrangement is shown in Fig. 1, where the aluminum wires, Al (about three centimeters in length) are fused into the thick glass envelope Gl. The glass envelope carrying the wires is shown sealed into the lamp in Fig. 2, which needs no further explantion.

"The joint between the glass and the aluminum is made tight during the exhaustion of the lamp in the following manner: It is a well known fact that aluminum when amalgamated by contact with a solution containing mercury oxidizes rapidly, becoming covered with the oxide alumina, in the form of a dense powder. This incrustation of oxide, if caused to occur between the aluminum wires and the glass into which they are sealed, is said to make an air-tight joint. These properties are given effect by applying a drop of a strong solution of mercuric chloride to the outer ends of the wires during the exhaustion of the lamp. The liquid is sucked in between the wires and the glass, and causes the amalgamation of the surface of the metal and its consequent oxidation, the oxide being tightly enclosed and forming an air-tight lute. As soon as the joint is tight the ingress of air is, of course, prevented, and oxidation ceases. Should a small leak occur it cures itself by the action of the oxygen on the remaining aluminum amalgam, resulting in the production of a further supply of luting material (alumina) precisely at the point where it is needed. The aspect of the wires with their coating of alumina o o is shown in Fig. 3. The cuts are reproduced from the London Electrician.

Reference
1) "Aluminum Leading-in Wires for Incandescent Lamps," Western Electrician, Vol 15, No 25, Dec 22, 1894, pg 293.