The transistor and the computer. The computer's miniaturization

The first wide applications of semiconductor electronics were in telephone industry and then in transistor radios. Computers began to use transistors in 1955. In that year IBM⁴ company marketed a-computer in which 1250 valves had been replaced by 2220 transistors, reducing the power consumption of the computer by ninety-five per cent.

The invention almost at the same time of the stored-program digital computer⁵ provided a large potential market for the transis-

' By the time — К тому времени когда

-' came to be called — стало называться з .

as early as — еще

IBM = International Business Machines (фирма в США)

^c' the stored-program digital computer —цифровой компьютер с хранимой программой

tor. The reason is that digital systems require very large numbers of transistor circuits. The synergy between a new component and a new appli­cation generated an explo­sive growth of both transistor and computer. The historical fact is thyt early efforts to miniaturize electronic com­puters were not motivated by computer engineers. Various satellite and mis­sile programs called for complex electronic systems to be installed in equipment in which size, weight and power requirements were severely constrained. So the effort to miniaturize was promoted by space and mi­litary agencies.

Figure 1. The memory board of the ru'w powerful computer produced by IBM can easily pass through the eye of a needle

Planar technology

The next step in microelectronics developments, the integrated circuit, was facilitated by the emergence of planar technology. In 1958 the first transistor was produced by the planar technique.

The planar technique consists of a sequence of three process­es—oxidation, photo-etching and diffusion. A wafer of silicon is oxidized and then coated with a photosensitive polymer, a photore­sist. A suitable miniature pattern is laid onto the resist and then exposed to light. The resist is vulnerable to some chemicals when exposed.' The exposed part of the surface can therefore be etched through the resist and through the silicon oxide underneath it. The resist is.then washed off and the desired "p" and "n" im­purities are allowed to diffuse² into the exposed parts of the silicon wafer, while the other parts are protected by the remaining oxide layer. This process can be repeated at will,³ and in this way intricate patterns of diverse conducting layers can be built up.

Initially the process was used just to produce a large number of transistors on a silicon wafer. The wafer was then sliced into individual chips.

' when exposed - будучи засвеченным

² are allowed to diffuse—получают возможность диффундировать

^!< at will — hi” желанию

A chip is a small square piece of pure silicon, layers of which have been etched away and doped with impurities so as to form a transistor and in future many alternate transistors, insulators and conductors which together make up the pattern of a complete circuit equivalent to thousands of transistors. Each chip contained one transistor and these were cased in a shell and eventually wired into electronic circuits. The great advantages of the process were that it improved manufacturing yield' and produced reliable products relatively cheaply. For example, the average price of a silicon transistor dropped seven times in six years.