Epitaxial Semiconductor Structures for the Infrared: Where Are We Now?
- PDF / 142,843 Bytes
- 2 Pages / 420.48 x 639 pts Page_size
- 26 Downloads / 283 Views
Epitaxial Semiconductor Structures for the Infrared: Where Are We Now?
Morton B. Panish AT&T Bell Laboratories Murray Hill, New Jersey 07974 EXTENDED ABSTRACT The use of epitaxial structures for the generation, detection, and modulation of light has had as its strongest driving force the development of fiber optic communications systems. Partly for that reason, the semiconductor materials systems that can be most readily optimized for wavelengths (800 - 1600nm) transmittable through state-of-the-art optical fibers, have been the most well developed. It also appears that the communications community has been most fortunate in that the semiconductor systems required, although far more difficult to deal with than Si, are relatively benign compared to those that operate at significantly shorter or longer wavelengths. Two developments that are now more than 15 years old initiated the extensive studies of compound semiconductor epitaxial structures that are now so important. These were the introduction of the heterostructure concept in the study of injection lasers and light emitting diodes (1), and the realization that lattice matched isoelectronic epitaxial layers, consisting in part of solid solutions between the binary compounds, would permit such structures to be grown. The development of epitaxial technologies for the growth of heterostructures started with the growth of GaAs/AlxGal-.As light emitters at about 900 nm. With this semiconductor system it is possible to take advantage of the lattice match that exists between GaAs and AlxGai-.As for all values of x. The initially used liquid phase epitaxy method for this system is being largely replaced by molecular beam epitaxy (MBE) (2) and Metal Organic Chemical Vapor Deposition (MOCVD) (3). These epitaxy methods have permitted the growth of an enormous variety of heterostructures with dimensional precision such that both layer thickness and interface abruptness can be controlled to a monolayer. The advent of optical fibers with low dispersion at 1300nm and low loss at 1550nm spurred the development of epitaxial methods for other III-V systems, most particularly Galnl-..AsYPI-y/InP and Allnl-.As/Gaxlnl-.As/InP, with direct bandgaps that included those wavelengths. Epitaxy in both systems may be achieved with monolayer precision, the former by several varieties of Gas Source MBE (GSMBE) (4,5,6) and the latter by conventional MBE (7), in spite of the need precisely to control the composition to maintain the lattice match with the substrate. These epitaxial methods have permitted a bounteous crop of lasers, light emitting diodes, optical detectors and transistors. This proliferation of III-V structures and devices has been largely restricted to wavelengths shorter than about 1700nm, most likely because, for many of the systems, the formation of extensive solid solutions between the smaller bandgap III-V compounds is inhibited by miscibility gaps (8) at compositions where there is a lattice match to a convenient binary III-V compound. This unfortunate situation has so f
Data Loading...
