Some other substance must be found. The use of carbon for arc lights and Edison's own experiments with carbon in his work on the telephone naturally suggested this substance as a possibility. It is said that this idea was brought forcibly to the inventor's attention by noticing the delicate spiral of vegetable carbon left in his hand after using a twisted bit of paper, one day, for lighting a cigar. This spiral of carbon was, of course, too fragile to be of use in its ordinary form. But it occurred to Edison that if a means of consolidating it could be found, there was reason to hope that it would answer the purpose. Experiments were begun at once, therefore, not only with processes of consolidation but also with various kinds of paper, and neither effort nor expense was spared to test every known variety of paper. Moreover, many new varieties of paper were manufactured at great expense from substances having peculiar fibres. One of these, made from a delicate cotton grown on some little islands off South Carolina, gave a carbon free from ash, and seemed to promise good results; but later it was found that the current of electricity did not circulate through this substance with sufficient regularity to get protracted and uniform effects. Nevertheless, since many things pointed to this fibre carbon as the ideal substance, Edison set about determining the cause of the irregularity in the circulation of the current in the filament, and a number of other experimenters soon became interested in the problem.
It was soon determined that the arrangement of the fibres themselves were directly responsible for the difficulty. In ordinary paper the fibres are pressed together without any special arrangement, like wool fibres in felting. In passing through such a substance, therefore, the current cannot travel along a continuous fibre, but must jump from fibre to fibre, "like a man crossing a brook on stepping-stones." Each piece of fibre constitutes a lamp or miniature voltaic arc, so that the current is no longer a continuous one; and the little interior sparks thus generated quickly destroy the filament. This discovery made it apparent that such an artificial, feltlike substance as paper could not be made to answer the purpose, and Edison set about searching for some natural substance having fibres sufficiently long to give the necessary homogeneity for the passage of the current.
For this purpose specimens of all the woods and fibre-substances of all countries were examined. Special agents were sent to India, China, Japan, South America, in quest of peculiar fibrous substances. The various woods thus secured were despatched to the Edison plant at Menlo Park and there carefully examined and tested. Without dwelling on the endless details of this tedious task, it may be said at once that only three substances out of all the mass withstood the tests reasonably well. Of these, a species of Japanese bamboo was found to answer the purpose best. Thus the practical incandescent lamp, which had cost so much time, ingenuity, and money, came into existence, fulfilling the expectation of the most sanguine dream of its inventor.
In using these bamboo carbon filaments the original spiral form of filament was abandoned, the now familiar elongated horseshoe being adopted, as the carbon could not be bent into the tortuous shapes possible with platinum. Later various modifications in the shape of the filament were made, usually as adaptations to changes in the shape of the bulbs.
At the same time that Edison was succeeding with his bamboo carbon filaments, J. W. Swan had been almost as successful with a filament formed by treating cotton thread with sulphuric acid, thus producing a "parchmentized thread," which was afterwards carbonized. A modification of this process eventually supplanted the Edison bamboo filament; and the filament now in common use—the successor of the "parchmentized thread"—is made of a form of soluble cellulose prepared by dissolving purified cotton wool in a solution of zinc chloride, and then pressing the material out into long threads by pressing it through a die.
The long thread so obtained is a semi-transparent substance, resembling catgut, which when carbonized at a high temperature forms a very elastic form of carbon filament. To prepare the filament the cellulose threads are cut into the proper lengths, bent into horseshoe shape, double loops, or any desired form, and then folded round carbon formers and immersed in plumbago crucibles. On heating these crucibles to a high temperature the organic matter of the filaments is destroyed, the carbon filaments remaining. These filaments are then ready for attachment to the platinum leading-in wires, which is accomplished either by means of a carbon cement or by a carbon-depositing process. They are then placed in the glass bulbs and the wires hermetically sealed, after which the bulbs are exhausted, tested, fitted with the familiar brass collars, and are ready for use.
The combined discoveries of all experimenters had made it evident that certain conditions were necessary to success, regardless of the structure of the carbon filament. It was essential that the vessel containing the filament should be entirely of glass; that the current should be conveyed in and out this by means of platinum wires hermetically sealed through the glass; and that the glass globe must be as thoroughly exhausted as possible. This last requirement proved a difficult one for a time, but by improved methods it finally became possible to produce almost a perfect vacuum in the bulbs, with a corresponding increase in the efficiency of the lamps.
THE TUNGSTEN LAMP
For twenty years the carbon-filament lamp stood without a rival. But meanwhile the science of chemistry was making rapid strides and putting at the disposal of practical inventors many substances hitherto unknown, or not available in commercial quantities. Among these were three metals, osmium, tantalum, and tungsten, and these metals soon menaced the apparently secure position of the highly satisfactory, although expensive, Edison lamp.
It will be recalled that the early experimenters had used two metals, platinum and iridium, for lamp filaments; and that these two, although unsatisfactory, were the only ones that had given even a promise of success. But in 1898 Dr. Auer von Welsbach took out patents, and in 1903 produced a lamp using an osmium filament. Its advent marked the beginning of the return to metal-filament lamps, although the lamp itself did not prove to be very satisfactory and was quickly displaced by a lamp invented by Messrs. Siemens and Halske, having a tantalum filament. On account of its ease to manufacture, its brilliant light, and relatively low consumption of power, this lamp gained great popularity at once, and for a single year was practically without a rival. Then, in 1904, patents were taken out by Just and Hanaman, Kuzel, and Welsbach, for lamps using filaments of tungsten, and the superiority of these lamps over the tantalum lamps gave them an immediate popularity never attained by either of the other metal-filament lamps.