Compound Semiconductors

Abstract

The invention of the transistor (1, 2) in late 1947 and its later embodiment in junction form (3) marked the beginning of a profound change in electronics. Not only did the receiving tube basically become obsolete, but many new device functions emerged, and ultimately unbelievable size reductions and enormous packing densities resulted. As is well known, it became possible to build an entire functional system in a single conducting block of material. Also, new forms of power devices, based on the transistor effect, became possible.With the invention of the transistor the study of electrical conduction in solids took on new importance, and similarly the growth and study of semiconductor crystals, at first, Ge (4). Since Ge is a much more tractable material than other elemental semiconductors, it was natural that the transistor began with Ge, and that in turn the transistor (and p-n junctions) motivated the growth and study of single crystal Ge of great purity (4, 5). In fact, Ge and its successor Si, because of the transistor, have been purified to such high. levels and have been grown with such perfection that Si now even serves as the standard for determination of Avogadro's number (6).The original transistor pioneers were quite aware that beyond their initial experimental material (Ge), Si offered a large potential (7). The real emergence of Si as a transistor material probably came in late 1954 and early 1955 when, at Bell Laboratories, Sidiffused-impurity devices were first constructed. Based on his considerable understanding of switching devices and the need for low off-stage leakage current, John Moll above all others urged the development of diffused-impurity Si devices; for this purpose, in his own group Au-Sb and A1 metallization of Si were developed (8), as well as certain forms of transistor devices (9). The most important form of diffused-impurity p-n junctions came from Carl Frosch (and L. Derick) (1O), including oxide masking and with it the teaching to leave the oxide on Si for protection of the device (11). These developments were sufficient for Moll and his colleagues to be able to convince Jack Morton (spring of 1955) to switch transistor development from Ge to Si. The rest is history, and now where Si "can do the job," there is little interest on the part of device researchers to attempt to displace it with another material. No other material has been as extensively developed, nor offers a natural oxide affording so many processing and device advantages.As useful as Ge and, more so, Si have proven to be in semiconductor devices, they possess also certain limitations. Their bandgaps are indirect and are fixed 4 487C 488C

Thermal annealing of liquid phase epitaxial gallium arsenide

“…Their band gaps are indirect and are fixed at Eg(Ge) ≈ 0.68 eV and Eg(Si) ≈ 1.1 eV. Hence, they are not useful as light emitters, transferred electron devices (Gunn oscillators) or efficient solar energy converters (Holonyak et al (1978)). When constructing room temperature operating radiation detectors, high band gaps are required such as Eg(GaAs) ≈ 1.41 eV to reduce thermally generated leakage currents and high purity with very low carrier concentration is needed to create large depletion volumes in these devices.…”