THE SEARCH SEMICONDUCTOR FOR MATERIALS

By ALBERT TRAINOR

In this article, the first of two, the author describes the rapid growth of materials technology caused by the development of semiconductor devices and outlines the new applications which have been found for the established semiconductor materials. Next week he will deal with new materials for conventional applications, elements, binary and ternary compounds, organic semiconductors and the effect of new developments on circuit design.

Reproduced from Electronics Weekly (No. 245 - 12th May 1965) with kind permission of the author and the publishers – The Certificated Engineer October 1965.

When Bardeen and Brattain discovered the transistor in 1948 they initiated what must be the most rapid industrial growth ever to occur. In only 16 years industry of great diversity has emerged and Its products have had an incalculable effect upon everyday life. 

In considering this mushrooming technology, however, the corresponding growth of a complex materials technology is often overlooked. 

Successful mass production of devices depends on the large scale manufacture of materials and single crystals to standards of purity and perfection hitherto unimagined. 

This in its turn has demanded research and development activity on a proportionate scale; first in the chemical industry, to prepare extremely pure starting materials such as germanium, germanium dioxide, gallium and indium, and secondly, within the semi-conductor industry itself in the field of further purification and the preparation of crystals of the required materials having good physical and electrical properties. 

Trends 

These articles are concerned not so much with the technology itself, but with its products and their uses and with possible future trends. 

Copper oxide and selenium are semiconductors that will be familiar to engineers as the active components of many power rectifiers. As such they have, of course, been used for much longer than the semi-conductor devices that we know today. 

Another material of ancient lineage, by semi-conductor standards, is lead sulphide, which enjoyed brief popularity in the days of the cat's whisker radio. It is now commonly employed in infra-red photocells. 

At rough estimate germanium and silicon form, the basis of at least 99 per cent of the semiconductor devices market. 

Germanium is of interest as being the "eka-silicon" of Mendeleeff, who predicted the properties of this then-unknown element as a test of the accuracy and usefulness of his new periodic table in 1869. 

Isolated 

The element was subsequently isolated in small quantities but remained extremely rare. However, since 1948 world production has risen until it is now measured in tons. 

The second group of materials, comprising the sulphides, selenides, and tellurides of lead and cadmium, are used in more specialised applications as photo-conductive materials in the visible and infra-red regions of the spectrum. Germanium and silicon also find limited use in this area. 

Gallium arsenide and indium antimonide are known as group IIl/V compounds since they are prepared from equal molar quantities of elements from these groups in the periodic table. They are the first commercial fruits of the search for new materials. 

Gallium arsenide finds limited use in fast switching diodes with switching times of the order of 100 picoseconds. 

Detector 

Indium antimonide is used as a photoconductive detector or as a photodiode operating in the far-infrared (3 to 7-micron region). Recently it has been used as a detector in the 30 microns to 8 mm wavelength region, thus bridging the gap between the far-infrared and microwave frequencies. 

Bismuth telluride has a field almost to itself-that of refrigeration utilising Peltier cooling. 

The appearance on the market of planar transistors, solid-state integrated circuits, etc., is due to advances in device technology. Nevertheless, they have brought with them n w materials, new problems and new techniques. 

Many planar transistors and most solid-state circuits have been made possible by an advance in materials technology known as epitaxy. 

In general, the term epitaxial growth refers to the overgrowth of a crystalline substance on a single crystal substrate in such a way that the orientation of the grown layer is related to, and controlled by, the crystal orientation of the substrate. 

Crystal 

In the semiconductor industry, the term is at present used in the specialised sense of the growth of a thin layer or film of a substance (usually silicon) on a single crystal substrate of the same material, the layer having the same orientation as the substrate. 

Such layers are commonly five to 20 microns thick and are grown from the gas phase by chemical reaction, the thickness, resistivity and type (N or P) being accurately controlled during processing. 

Such a structure allows devices to be fabricated in the thin semiconductor layer, using the substrate as a low-resistance back contact, effectively bringing the back collector contact of the device in close proximity to the actual junction.

This provides much lower collector series resistance than would otherwise be the case and at the same time a wafer of semiconductor of a thickness suitable for handling in the normal manner. 

Such material has, however, even greater potential than this and ever more complex structures are being investigated in research and development laboratories. 

Junction 

The growth of single and multiple P/N junctions is possible, thus taking junction fabrication out of the device manufacturing area. This is already having the effect of bringing the device designer and the materials technologist even closer together so that within a few years the two may be virtually indistinguishable. 

Of more interest to this article, however, is the possibility of devices operating on different principles to those of the conventional transistor. 

One such device, which has already been realised, is the metal-oxide-semiconductor transistor (MOST). Though called a transistor the MOST operate on an entirely different principle-it is a majority carrier device. 

It has an exceptionally high input impedance, and thus has many possible new applications, and may replace the thermionic valve in many types of circuit. 

Another device that should benefit from improved materials is the Esaki, or tunnel diode. Interest in this device centres around its potentially high operating frequency, which may be thousands of megacycles. 

Present materials and techniques have produced successful devices for microwave applications, although a really practical tunnel diode for switching purposes is still sought. 

The light-emitting is another recent development. When minority carriers are injected into a semi-conductor they recombine in one of several ways with majority carriers. 

In one form of recombination, two carriers of opposite type are annihilated with the emission of the energy of a wavelength equivalent to the difference in the energy level of the carriers. With a suitable choice of material the energy is emitted in the visible or near infra-red and, in principle, this gives a highly efficient method of converting electrical energy into light. 

Because of its band structure, gallium arsenide may be used to make diodes emitting light in the near infra-red, and other higher energy gap materials may find application in the production of visible light sources in the future. 

It is worth noting that gallium arsenide has also been used as a semiconductor laser producing coherent light in the near-infrared. 

All of these devices may benefit from the development of new materials suited to their special needs.