This excerpt was taken from the issue of the WorldWatch Paper, number 39, "Microelectronics at Worl<: Productivity and Jobs in World Economy" written by Colin Norman.
World Watch Institute, 1776 Massachussetts Avenue, N.W., Washington, D.C. 20036, USA.
In 1946, the world's first electronic computer was switched on at the Moore School of engineering in Pennsylvania. It was an impressive machine. Called ENIAC (Electronic Numerical Integrator And Calculator), it occupied a large room, contained 18,000 vacuum tubes, and consumed enough power to drive a locomotive. Today, a computer with equivalent capabilities fits into a pocket, costs less than $100, and runs on flashlight batteries. Such are the dimensions of the microelectronic revolution.
As electronic components have shrunk both in size and cost, they have become pervasive. "It is not an exaggeration to say that most of the technological achievements of the past decade have depended on microelectronics," asserts Robert Noyce, a leading figure in the industry. He points out that microelectronic devices have played a central role in the development of space technologies, missiles, calculators, digital watches, word processors, and industrial control equipment; yet he suggests that we have seen only the tip of the iceberg. 'The microelectronics revolution," says Noyce, "is far from having run its course."
The effects of the microelectronic revolution will be markedly different from those of the Industrial Revolution, however. The development of industrial technology largely enhanced human physical capabilities, enabling people to harness more energy, process and shape materials more easily, travel faster, and so on. But the development of microelectronics extends mental capabilities, for it increases the ability to process, store, and communicate information, arxi it enables electronic "intelligence" to be incorporated into a broad range of products and processes
It has taken just 30 years for the microelectronic revolution to develop, and only in the last five years has it begun to take off. It started in 19-47-the year after ENIAC made itsdebutwith the development of the transistor. The basic building block of all modern electronics, the transistor swiftly rendered obsolete much of the circuity on which electronic equipment had been based. Transistors consist of semiconductor material —usually silicon—to which minute quantities of impurities such as phosphorus or borun have been added in discrete regions. The impurities alter the electrical properties of the semiconductor, causing it to conduct electricity when it is subjected to sufficiently large voltages. Transistor usually function like tiny electronic switches, shuttling electrons around electrical circuits.
Since transistors are smaller and much less power-hungry than the vacuum tubes they replaced, goods such as radios, television sets, and computers became more compact during the fifties and they required relatively little power to operate. For better or worse the transistor radio emerged. But there was a limit tos this trend: individual transistors, like the bulky components they replaced, still had to be wired together, and thus a piece of equipment with many thousands of electrical parts remained a pretty hefty contraption. This posed serious constraints for military and space hardware, where size and weight are particularly critical, consequently, the U.S. Government
and private companies began to pour money into further miniaturization efforts.
These investments paid off in 1959 with the development of the integrated circuit, the centerpiece of microelectronic technology. In the late fifties, transistors were produced in batches in thin slices of silicon that were later cut into sections, each of which contained a single transistor. Scientists at Texas Instruments and at Fairchild Camera and Instrument Company had the bright idea of wiring the transistors together into a complete circuit while they were still in the slice of silicon. The result was the first integrated circuit. The first experimental devices were made by connecting the transistors together with wires, but later models had tiny aluminum conductors
deposited directly on the semiconductor surface.
As the design and construction of integrated circuits improved, the number of transistors and other electronic components, such as resistors and diodes, that could be placed on a single piece of silicon increased exponentially, doubling roughly every year from the early sixties to the late seventies. By 1980, the most densely packed circuits contained close to 100,000 components on a silicon chip measuring just five millimeters across, and the aluminum conductors linking them together were about 30 times thinner than a human hairIn three decades, a roomful of vacuum tubes, wires, and othercomponents has been reduced to the size of a cornflake. And the process is not over yet. Although it becomes increasingly difficult to design and imprint the circuits as the density of the components increases, chip manufacturers are confident that by 1990 they will be able to produce integrated circuits containing at least one million components. Already, chips containing some 250.000 components are being designed.
Putting thousands of components on a silicon chip is, to put it mildly, a difficult process. First, the layout of the components and the pattern of connections between them are designed with the help of a computer. The circuit design is then drawn by the computer on a series of photographic plates, which are used to print a reduced image of the design on glass plates, called masks. These masks form the heart of the process of imprinting the circuit design on silicon chips; like photographic negatives, they can be used time and again to produce multiple copies of a particular integrated circuit
Imprinting the transistors and other components requires microscopic areas of the silicon chip to be "doped" with impurities, and a maze of connections more complex than the street plan of a large city to be laid down between them. It is done in an automated process in which thousands of identical circuits are constructed simultaneously. The basis of the process is a complicated photoengraving technique that uses the masks to determine the pattern in which the impurities are embedded and to establish the maze of aluminum connections.
This intricate piece of technological wizardry has slashed the cost of electronic circuitry. The price of a given amount of computer memory, for example, dropped by a factor of about 50 during the seventies—at a time when the price of just about everything else rose by leaps and bounds. And equally sharp reductions in the price of the electronic "brains" of consumer goods such as calculators and electronic watches have transformed them from luxury products to everyday items.
The chief reason for these plummeting costs is that once the integrated circuits have been designed and etched onto the masks, they can be mass-produced. As with automobiles and other assembly-line products, the immense capital costs of setting up a manufacturing plant are spread over a large number of items. Moreover, the actual production process costs about the same no matter how many components are packed on each chip. This means that as more and more transistors are crammed into each circuit, the cost of each transistor drops. In 1960, an individual transistor cost about $10; today, a transistor in an integrated circuit costs a fraction of 1c.
This combination of shrinking dimensions and declining costs would alone be sufficient to ensure an expanding market for integrated circuits. But the circuits have yet another valuable attribute: they perform their electronic operations extremely rapidly. As more and more components are packed on a chip, the distance between them shrinks and the time it takes for electrons to shuttle from one component to another is reduced. Thus, in general, the more densely packed an integrated circuit is, the faster it functions, this is especially important in the performance of computers, which carry out millions of electronic operations in the processing and manipulation of data.
Most integrated circuits are designed for specific tasks, such as operating a digital watch or storing information in a computer memory in the form of pulses of electrons. In 1971, however, the American microelectronics company Intel brought out a radically different, more flexible type of integrated circuit that vastly extended the range of applications of microelectronic technology. Intel essentially put the entire central processing unit of a computer—the complex circuitry that processes information and carries out computations— on a silicon chip. The resulting integrated circuit, known as a microprocessor, can be programmed like a computer to carry out a broad range of functions.
A microprocessor is essentially a device for processing information that is fed to it through a keyboard, a wire from an instrument, or some other input mechanism. Once the information is converted to an electronic binary language (with only two possible words, zero and one, which are determined by whether the transistors are switched "on" or "off"), it is processed according to instructions stored in the microprocessor's own memory or in other, attached memory circuits. These instructions, called programs, cause the microprocessor to perform a series of logical steps in which pulses of electrons are moved through its circuits, stored, and moved again at extremely high speeds—a process a little like the high-speed shuffling of beads on an abacus. When these steps are completed, the microprocessor activates an electrical circuit that is connected to an output device, such as a screen that displays the results of a calculation or a switch that opens or closes a valve.
Even the early microprocessors boasted more computing power than ENIAC, and their capabilities have since increased many times over. The central processing units of powerful computers that would have cost thousands of dollars a few years ago are thus being mass-produced for a few dollars apiece. A startling achievement in its own right, this development means thatthe computer's ability to process information and carry out instructions can now be incorporated relatively cheaply into a variety of machines, ranging from cruise missiles to microwave ovens.
These tiny electronic devices did not spring suddenly from the laboratory bench and begin to change society, however. As with any new technology, the development of microelectronics has been pulled along by economic and political forces. During the sixties, the U.S. military and space programs provided the driving force, accounting for most of the integrated circuits produced in the United States. This burgeoning military demand provided a stable market for the small, innovative microelectronics companies that spearheaded the technological development, and it helped launch the industry on its high-growth trajectory.^ It also changed the nature of many weapons. Microelectronic controls now constitute the brains of guided missiles, "smart" bombs, electronic sensors, and other ingredients of modern warfare, and they have played a central role in the development of military space systems.
During the seventies, the focus shifted toward civilian applications, and commercial incentives are now pushing the development of the technology. The manufacture of integrated circuits has turned into a $10-billion-a-year industry, and sales are growing at the phenomenal rate of 30 percent per year. Military programs account for about $1 billion worth of microelectronic devices, and the rest of the expenditure is widely disseminated throughout the economy.*
The economic impact of microelectronic devices is far greater than their sales figures suggest, however. "I like to think of semiconductor technology and its evolution as being the crude oil of the electronics industry," says Jerry Sanders, president of the U.S. semiconductor company Advanced Micro Devices. Electronics, he notes, is "a $100 billion industry now, arxl headed for perhaps $800 billion by the late 1980's-all grounded on semiconductor components."'
The electronics industry thus faces a period of growth that will propel it into the front ranks of the world's major industries. The microelectronic revolution, by broadening the applications of electronic devices and slashing their costs, has sown the seeds of this expansion. And as microelectronic devices have become more pervasive, the nature of many products is changing rapidly.
One conspicuous result of this flood of microelectronic technology is a growing array of new consumer goods. In the space of just a decade, the manufacture of pocket calculators, digital watches, electronic games, and children's toys has mushroomed mushroomed into a $4-billion-a-year business. And new electronic gadgets, such as language translators and calculators built into wristwatches, are entering the market in quick succession.*
Established products are also being transformed as microprocessor and other integrated circuits replace mechanical and electromechanical components. Some washing machines, for example, are now equipped with a microprocessor that controls the sequence of wash cycles and water temperatures according to instructions entered through a calculator-style keyboard; about 60 percent of the microwave ovens currently sold in the United States are equipped with similar microelectronic timing devices.^ More sophisticated controls are being incorporated into a range of machine tools, gukling the actions of cutters, welders, drills, and stamping machines.
There are several advantages to incorporating microprocessors into household as well as industrial products. Microelectronic controls are more reliable and flexible than their mechanical counterparts. They can be used to regulate complex cycles lasting from a few seconds to several hours, and they can quickly be reprogrammed to enable the machines they control to perform different tasks. The reduction in the number of moving parts that results from the replacement of mechanical components with electronic controls can, moreover, be significant: a sewing machine marketed in the United States in the late seventies uses a single integrated circuit to control the sequence and pattern of stitches, in place of some 350 cams, gears, and other parts.'
The automobile industry is also turning to microelectronics in its headlong rush to improve fuel economy arxl reduce exhaust emissions. In the next few years, virtually every new car made in the United States will have a microprocessor under the hood to control fuel intake or ignition timing. This application alone will account for millions of integrated circuits each year.
The biggest single user of microelectronic devices will, however, be the computer industry itself. The development of cheap and powerful microprocessors and integrated memory circuits has spawned a broad range of small, flexible computers that can be programmed for a variety of tasks. Just a fews years ago, the cheapest computers on the market cost hundreds of thousands of dollars and they were big, powerful machines. Now home computers the size of a typewriter can be bought for less than $1,000, and powerful business machines for less than $10,000. These developments have brought computing power to the fingertips of a rapidly growing number of people, and they have opened the way for an expansion of electronic record keeping, information processing, and data gathering. They have also blurred the distinction between computers and other office machinery, for computer intelligence can now be built into typewriters, copies, and a host of other pieces of office equipment.
Finally, one of the most important but least conspicuous uses of microelectronics in the coming yeras will be in the improvement of telecommunications systems. The last great change in telecommunications networks took place only a few decades ago, when manual telephone exchanges were replaced by electromechanical switching stations. But those systems are already out of date, and they are now being replaced by faster more reliable all-electronic exchanges and satellite linkages,
both of which rely heavily on microelectronics.
Most industrial countries are now embarking on the enormously expensive task of upgrading their telecommunications systems. As a result, telephone networks not only will be able to handle more calls, but they will also be better equipped to relay electronic messages between computers, word processors, and other intelligent machines. In other words, the improved telecommunications systems will provide a vital link that will permit growing numbers of computerized machines to "talk" to each other. This pending merger between telecommunications, computing, and information processing is likely to be the most far-reaching consequence of the microelectronic revolution, for it will greatly extend humanity's capacity to process and transmit information.
Developments in microelectronics have thus resulted in startling changes in the dimensions, cost, performance, and reliability of electronic components. As the technology is pushed even further, with hundreds of thousands of components jammed on a silicon chip, the power of microelectronic devices and their range of applications will be enhanced even more.
NOTES
1. Robert N. Noyce, "Microelectronics," Scientific American, September 1977
2. A general description and history of computers and microelectronics is given in Christopher Evans, The Micro Millennium (New York: Viking Press, 1980); more detailed descriptions of microelectronic technologies are contained in several articles in the September 1977 issue of Scientific American and in an excellent series of articles by Arthur L. Robinson in the issues of Science of May 2, May 9, May 30, June 13, and July 11, 1980. The description of computer and microelectronic technology in this and
following paragraphs is taken from these sources.
3. The impact of demand from military and space programs on the development of microelectronics is discussed in Charles River Associates, Innovation, Competition, and Government Policy in the Semiconductor Industry (Boston: 1980).
4. Estimates of sales and growth rates are derived from figures in Semiconductor Industry Association, 1979 Yearbook and Directory (Cupertino, Calif: 1979), Charles River Associates, Innovation, Competition and Government f^licy, and Electronic Industries Association, Electronic Market Data Book 1980 (Washington, D.C.: 1980).
5. Sanders quote is taken from testimony before the U.S. International Trade Commission, May 30, 1979, as reported in Semiconductors ndustry Association, 1979 Yearbook and Directory.
6. Estimate of sales of electronic consumer goods is from "Microelectronics Survey." The Economist, March 1, 1980. It consists of $1 billion for toys and games, $1.5 billion for calculators, and $ 1.7 billion for watches
7. Ibid.
8. Sewing machine application is described in Robert T. Lund et al.. Microprocessor Applications: Cases and Observations (Cambridge, Mass.: Center for Policy Alternatives, 1979).