Semiconductor Lasers

Early work on semiconductors focused on silicon, but silicon itself cannot emit laser light. The invention of the transistor at Bell Laboratories in 1948 by William Schockley, Walter Brattain, and John Bardeen stimulated research on other semiconductors. It also provided the conceptual framework that would eventually lead to an understanding of light emission in semiconductors. In 1952, Heinrich Welker, at Siemens in Germany, described semiconductors from elements found in column III and V of the periodic table as potentially useful for electronic devices. One of these, gallium arsenide, or GaAs, was to feature prominently in the search for an efficient communication laser. Necessary precursors to its full exploitation were basic studies of layer-by-layer growth of high-purity crystals, research into defects and dopants (impurities added to a pure substance to change its properties), and analysis of the effects of heat on the stability of compounds. With these advances, research groups at General Electric, IBM, and the Lincoln Laboratory at the Massachusetts Institute of Technology developed working GaAs lasers in 1962.

But an old problem persisted: overheating. Lasers that used a single semiconductor, usually GaAs, were not very efficient. They still required so much electricity to initiate laser action that, at normal room temperature, they quickly overheated; again, only pulsed operation was possible, which was not practical for communication. Physicists tried various methods for removing the heat--such as placing lasers atop other materials that were good heat conductors, but were unsuccessful. Then, in 1963, Herbert Kroemer at the University of Colorado proposed a different approach--to build a laser consisting of a semiconductor sandwich, with a thin active layer set between two slabs of different material. Confining the laser action to the thin active layer would require very little current and would keep the heat output at manageable levels.

In 1967, Bell Laboratories researchers Morton Panish and Izuo Hayashi suggested the possibility of creating a suitable multilayered crystal of GaAs and AlGaAs. AlGaAs is actually a crystalline solid solution in which some of the atoms of Ga in GaAs are replaced by Al. The needed variation in electrical and optical properties in the multilayered crystal required that a large fraction of the Ga be replaced, typically as much as thirty percent. At all ratios of Al to Ga in AlGaAs crystals the atomic spacings differed from those of GaAs by less than 1 part in a 1,000 so that the interfaces between the layers would not have defects caused by poor size matching of the crystal lattices. When grown on either side of a thin layer of GaAs, AlGaAs would restrict all laser action to the GaAs layer. Several years of work lay ahead of them, but the path to "heterostructure" lasers--miniature semiconductor devices operating continuously at room temperature--now lay open.

One hurdle remained: how to transmit light signals across long distances. Long-wavelength radio waves travel freely through the air, piercing fog and heavy rain with ease. But short-wavelength laser light bounces off atmospheric water vapor and other particles to such a degree that it is either scattered or blocked. A foggy day could shut down a laser communication link. Light therefore needed a conduit, analogous to telephone lines.