The "Friendly Effect" In Early Transistors.
By Robert Ryder (c) SMECC
|Some of the early junction transistors, around 1952, showed a peculiar and unexpected behavior. Mounted in a test rack, as we walked past, the transistor currents would change. If you waved your hand at the transistor, the meters would wave back! Therefore, the effect was immediately dubbed the "Friendly Effect."
No such effect was known in pre-transistor semiconductor devices, such as the silicon point-contact microwave diodes widely used as "mixers" in microwave radars of World War II - and also for years afterwards. The diodes were customarily exposed to ambient air, without any surface protection, and without noticeable time variation in their properties.
Of course, those diodes had very poor-quality current-voltage characteristics, as measured on a low-frequency curve tracer. Front-to-back current ratios hardly exceeded 10 to 1, and reverse breakdown voltages of the order of only two or three volts were usual. A small change in reverse current would not have been noticed. Nevertheless, those diodes were preeminent for use in microwave radar receivers, because they were very much faster than any vacuum-tube mixers, and would operate at previously unheard-of high frequencies, like 10,000 megahertz (MHz). Together with the British invented, and Bell Labs-developed and Western Electric-produced, magnetron vacuum tubes which could produce such frequencies, the advantage of the Allied radars over the Germans (who were blind to such frequencies), had an important effect on the course of World War II.
We had high hopes, therefore, in the early 50ís, for transistors as microwave devices. But it took almost twenty years to achieve the very outstanding performance of ultralow-noise microwave field-effect transistor amplifiers (FETís).
The first point-contact transistors, even though much better rectifiers than the microwave diodes, still had rather poor reverse-biased resistance, compared to vacuum-tube amplifiers of comparable performance. The collector points drew rather high "leakage" currents in the reverse direction, and were so prone to burn-out that great care had to be used in turning them on. Consequently, we concentrated on measuring their amplifying properties, rather than risking the loss of the experimental units by putting high voltages on the collectors. Typically they might be biased at 25 or even 50 volts, with currents of 5 or 10 milliamperes. The fact that one could obtain amplification of the order of 20 decibels (db) with such low biases was obviously going to be a great advantage of semiconductor devices over tubes.
With the advent of high-purity semiconductors, and the new junction transistors (1951), a number of things began to become clear. To make transistors withstand voltages of 100 volts and up, material of unprecedented high purity would be needed, in the form of single-crystals of the semiconductor. This formidable requirement was quickly met by our materials people, especially by Bill Pfann, who developed zone-refining and zone-leveling to be able to make single-crystal germanium with impurity concentrations of less than one part per billion. Pfann could also "back dope" such pure materials to the order of one part per million, controlled to plus or minus a few per cent!
At a time when "U S P" chemicals normally contained impurities of the order of a few parts per thousand, the new transistor materials were literally millions of times purer than standard research reagents -- an astounding achievement. It was such materials of super-high purity which showed the "Friendly Effect" -- unprecedented, because that kind of material had never before been available.
The first commercial junction transistors were an immediate and colossal commercial success in portable AM radios. They were produced by the millions, and were used all over the world, after 1954. They brought the news of modern civilization even to illiterate natives in darkest Africa, and to the islands of Indonesia. But the success did not extend to use in the telephone company plant!
The new portable radios would operate on so little current that the ubiquitous pilot lights -- normally used to show the radio was turned on -- immediately were relegated to the waste basket. Typically, the pilot light might draw 175 milliamperes (ma) of current, while the entire complement of the radio receiver transistors could be operated for no more than 20 milliamperes. Clearly there was no point in wasting battery life on a pilot light.
But the entire telephone plant was operated at ten times higher power levels, of the order of a major fraction of a watt. Indeed, ringing the telephone bell -- long a routine function -- required the plant to transmit some 10 or 20 watts in short pulses. The early alloy transistors were too small in power capability to give adequate power performance in the telephone plant. They also petered out in frequency response at some two megahertz, more than a thousand times slower than the microwave diodes, and thus were not useful at microwaves either. Clearly, much work would be needed to make transistors capable of displacing vacuum tubes in the telephone system.
Nevertheless, the leaders of Bell Telephone Laboratories -- especially the legendary Mervin Kelly, who had much personal experience in the development of pioneering high frequency vacuum tubes -- had enough faith in the prospects of the new transistor device, and in the project leaders, especially in Jack Morton, that they decreed, as early as 1954, that there would be no more development of low-power receiving tubes at Bell laboratories for use in the telephone plant. They expected that by 1959 all new telephone systems would rely on transistors.
Such faith was not based on personal confidence only. It was soundly based on scientific understanding, even though the necessary performance had not yet been achieved. I cannot imagine any political body taking such a gargantuan risk, especially after seeing what the U.S. government has done to destroy the U. S. nuclear power industry. Unfortunately, U.S. law is not based on science -- it is based on suspicion of science, and lack of understanding of scientists.
As you can see, the performance gap was tremendous in 1952; yet the gamble was pushed hard by Morton, with almost unbelievable-success. Most of the performance gap was erased by the invention of the silicon diffused-base transistor, announced in 1955; and entirely new levels of performance were attained by the Telstar communications satellites in 1961. Yet Congress in 1963 refused to permit AT&T to operate a satellite communications system, on the ground that they would prefer to wait for some competitor to do the job. If AT&T can do it, why canít others? In this way, the U.S. Congress gave away, for nothing, the ten-year lead in high technology that the skill and understanding of Bell Labs had established over all the rest of the world.
By 1961, Bell Laboratories had made feasible Telstar satellites containing electronics with only one vacuum tube -- the transmitting amplifier, a specially-designed traveling-wave tube; and powered by radiation-hardened solar cells, a by-product of the transistor development. The subsequent success of the U.S. space program depended entirely upon transistors. The fact that the Russians had no good transistors, or computers, for many years meant space supremacy for U.S. communications.
By-products of the transistor -- solar cells for power supply, particle detectors, "integrated" circuits, and eventually lasers, charge-coupled devices (CCDís) and ultra-low-noise microwave amplifiers, eventually took over most of the leading roles in the space program. But it took a lot of vision on the part of our leaders to see such possibilities as early as 1954.
The "Friendly Effect" appeared as a result of the use of semiconductors of unprecedented high purity. It appeared that such transistors operated best in an atmosphere including oxygen and about 50% relative humidity, and moderate temperature -- very much like the atmosphere good for sustaining life in humans! But such human-like variability as the Friendly Effect was quite intolerable in electronic apparatus. Within about ten years, as a result of the development of protective coatings of silicon oxide and silicon nitride, the Friendly Effect was completely eliminated. That fascinating story of how it was done is now well-known to scientists, but it entailed many discoveries of the previously unknown properties of semiconductors. The Friendly Effect was one of the early and unexpected properties which was involved in discovering how to make todayís powerful communications and control systems, as well as the nowadays ubiquitous and indispensable electronic computers, and other "smart" machines.
To sum up, I conclude that this story demonstrates once again that science is the most powerful method yet discovered for obtaining knowledge. But it also demonstrates the need for knowledgeable administrators who can understand when and how to make use of the new knowledge. Unfortunately, the supply of the necessary scientific understanding is very inadequate, among the lawyers who design and operate the U.S. government.
|THE AUTHOR: ROBERT M. RYDER graduated from Yale in 1937 with a B.S. degree in physics, followed three years later by a Ph.D. degree in physics from the same university.
He joined Bell Telephone Laboratories in July, 1940, to work on microwave amplifier circuits, and during most of the war was a member of a group engaged in studying the signal-to-noise performance of radars. In 1945 he transferred to the Electronic Development Department to work on microwave oscillator and amplifier tubes for radar and radio relay applications. In 1948 he joined the group engaged in the development of Transistors. He spent many years following the work on transistors, involving himself in diodes that were used to generate microwave signals.
Mr. Ryder is now retired from Bell Telephone Laboratories and lives with his wife in New Jersey, where his interests are bridge, GO, and writing about science and electronics.