Useless No More: Gordon K. Teal, Germanium, and Single-Crystal Transistors

By David C. Brock

By 1952 it seemed certain that the advantages of the transistor—reduced size, enhanced performance, and increased reliability—spelled the demise of the vacuum tube as the central component in the electronics world. These virtues of the transistor were due to the unique use of materials in its formation. As William Shockley noted in that year’s Proceedings of the IRE: “Transistor electronics exists because of the controlled presence of imperfections in otherwise nearly perfect crystals.” An accomplished physicist at Bell Telephone Laboratories, Shockley had recently invented an important new breed of transistor that fundamentally relied on single crystals of germanium. The key actor in the development of these crystals was Gordon K. Teal, a chemist captivated by germanium and determined to conquer its “uselessness.”

Germanium is a semiconductor, with the same crystal structure as diamond and electrical properties somewhere between those of conductive metals and nonconductive insulators. It was discovered in 1886 by Clemens Winkler, who named the grayish-white, lustrous substance after his homeland, but Dmitri Mendeleev had predicted its existence between silicon and tin in the periodic table more than a dozen years earlier.

Teal fell under the spell of germanium while earning his Ph.D. in chemistry at Brown University during the 1920s. As he put it in 1976, germanium “was [then] a material studied only for its scientific interest; its complete uselessness fascinated and challenged me. My concentration on this shiny metallic-appearing material during my graduate school days resulted in a continuing personal sentimental attachment for germanium, which, to me, at least, was and is an exotic element.”

In 1930 Teal left his university studies to join Bell Labs, which in the 15 years preceding World War II was at the forefront of R&D in vacuum-tube electronics. After a brief stint in the chemical research department, Teal was redeployed to the electro-optical department, which was working hard on new television systems. Teal’s mandate was to form a variety of materials into novel structures for use in electronic television components.

Teal was not the only chemist in Bell Labs’ extended organization whose research focused on materials in advanced electronic systems. Around 1935 Russell Ohl and several coworkers at a branch facility in Holmdel, New Jersey, were involved in projects on radio systems utilizing wavelengths much shorter than those of mainstream systems. They encountered a stumbling block: vacuum tubes used for conventional radio did not perform well with the new wavelengths. Ohl’s solution was to resurrect an antiquated radio technology—the crystal rectifier, the temperamental receiving element used in early radio sets. Investigating a range of crystalline materials, Ohl unsurprisingly turned his attentions to the chemical element silicon, a common ingredient in many of the old crystal “point-contact” rectifiers—devices that rely on the contact between a metal wire and a crystal to convert alternating into direct current, thus allowing them to receive radio transmissions.

Silicon is similar in appearance to germanium, but whereas Teal found germanium “exotic,” silicon is arguably prosaic. It is the second most abundant element on earth, contributing over a quarter of the weight of the planet’s crust; oxygen, the most abundant, contributes just under half. Silicon rarely reveals itself in its elemental form, instead appearing in mixtures, of which quartz and silica (sand) are familiar examples.

At the end of the 1930s Ohl, convinced that using chemically pure silicon crystals in point-contact devices would cure the erratic performance of earlier rectifiers, was working on purifying silicon material—melting and cooling it to isolate impurities. In early 1940 he observed a new and puzzling behavior in one of his polycrystalline silicon samples: the electrical conductivity of the silicon piece changed when he shone light on it. The silicon sample was itself acting as a rectifier, and a photosensitive one at that. Ohl and his colleagues determined that this rectifying action was caused by the junction of two chemically distinct regions within the polycrystalline silicon sliver. On one side of the junction lay silicon with traces of boron, which was deficient in available electrons; it was soon dubbed “P-type” (for positive) silicon.

On the other side lay silicon with traces of phosphorus, or “N-type” (negative) silicon, which had a surfeit of electrons. Ohl had discovered the P-N junction. The Bell Labs leadership was enthralled. Ohl’s P-N junction opened up the possibility for new photosensitive devices and rectifying diodes in which the material was the device.

Ohl’s work had implications for Teal’s work on photosensitive materials for television and undoubtedly resonated with Teal’s attachment to silicon’s cousin on the periodic table, germanium. Teal now saw that germanium’s uselessness might be an illusion—if silicon could form a rectifier, so might germanium. He hurriedly created some films of purified germanium material (deposited pyrolytically from germanium hydride gas) and personally brought them to Ohl’s Holmdel crew. However, the branch laboratory continued to focus on silicon rather than Teal’s import.

When Bell Labs shifted its priorities to accommodate important new war-related
R&D like radar, its commitment to television research was drastically reduced. Teal returned to the chemical research department. In early 1942 he began to improve his method for making polycrystalline germanium, yielding P- or N-type material to use in pointcontact rectifiers. But he soon went further and created a germanium rectifier itself. His system—a complex jungle of glass tubes, gas lines, flow meters, valves, heaters, and reaction vessels—relied on controlled mixing in a reaction chamber of germanium chloride gas and a chloride gas mixed, or “doped,” with the desired impurity (e.g., boron). The mixed gases decomposed as they passed over heated elements and deposited layers of a germanium-impurity alloy on a metal base at the bottom of the chamber. The surface of the alloy was then etched and a metal-wire point contact was applied, yielding the final germanium rectifier.

Teal’s system hardly made a splash at Bell Labs, which had set its course firmly on silicon rectifiers. Disappointed, he switched his focus, adapting his germanium approach to incorporate silicon. Soon he was producing silicon-germanium resistors, depositing film on ceramic substrates. These had highly desirable resistance and temperature characteristics, but Teal’s research still failed to make waves at Bell Labs.

Ironically, while Teal was out of work for an extended stretch (probably with chemical pneumonia linked to his inhalation of noxious substances in a laboratory mishap), the MIT Radiation Laboratory prevailed on Bell Labs to undertake a crash program for germanium rectifiers, and Bell gave that work to others. On his return, Teal was assigned to other, non-germanium research for radar components.

Immediately after the war, however, he returned to work on semiconductors, managing the materials aspect of a large-scale project that used a silicon compound to create a new electronic component known as a varistor, a variable resistance element for use as a static and crackle buster in telephone handsets. Elsewhere in Bell Labs a semiconductor research group was formed under William Shockley, whose mandate was to conduct fundamental studies on the electronic behavior of silicon and germanium—with an eye to possible new devices based on that behavior.

In December 1947 Walter Brattain and John Bardeen of Shockley’s group opened up a new era in electronics with their invention of the first working solid-state amplifier using germanium: the point-contact transistor, formed by two closely spaced point contacts atop a piece of a germanium-tin alloy. Shockley, Brattain, and Bardeen won the Nobel Prize for physics in 1956 for this work.

Germanium was certainly no longer useless. It was the material—the alloy of germanium and tin—that allowed the point-contact transistor to exhibit remarkable levels of power and voltage gain. Teal dashed off a number of memoranda to his Bell Labs superiors, now proposing production of single, near-perfect crystals of highly purified germanium. He reasoned that these crystals would give researchers a medium in which the fundamental operation of the transistor could be determined, basic knowledge that would lead to improved devices. Teal’s superiors were aware that imperfections in crystal structure and the presence of chemical impurities were both factors that determined the electrical behavior of semiconductor materials, but they rejected his proposed single-crystal material as needlessly expensive. They felt that if a relatively uniform crystal sample was required, it could be removed from a larger mass of readily available high-purity polycrystalline germanium.

Teal still hoped to initiate a single-crystal program, and his opportunity arrived in fall 1948, when he learned that John Little, a coworker and mechanical engineer, had been assigned to devise a way to cut germanium into uniform slices so as to reduce waste in manufacture. Teal told Little that he could create a rod of germanium with a small, relatively uniform diameter by fashioning it as a single crystal of pure germanium. Teal worked with Little to design and build a new machine, a crystal puller, a device that used seed crystals to draw pure crystals from a molten pool of material. To maintain crystal perfection, Teal and Little had to control a number of variables: the temperature of the melt, the cooling rate of the pulled crystal, and the rate at which the crystal was pulled by the seed crystal. To meet these requirements, their crystal puller employed a melt of ultrapure germanium in a graphite crucible, heated using radio-frequency induction coils. The movement of the seed crystal was governed by an adjustable precision motor. The temperature of the interface between the pulled crystal and the liquid melt was controlled by jets of hydrogen, which also determined the diameter of the pulled crystal. To preserve the chemical consistency of the single crystal, most of the apparatus operated under a bell jar filled with a continually refreshed atmosphere that counterbalanced the inevitable impurities in the original germanium melt.

Impressed by Teal’s success, Jack Morton, head of Bell Labs’ transistor development, allowed him to continue the work in its own right. However, Teal retained his full-time job on the varistor program and had access to his crystal-forming facilities only at night. Pulling germanium single crystals from 4:30 p.m. to 2:00 a.m., he labored through the first half of 1949 to produce the material and spread it around Bell Labs.

In the latter half of 1949 Teal gave batches of his new material to Morgan Sparks, a fellow chemist and a member of Shockley’s group, who in turn gave some of it to a physicist in the group, J. R. “Dick” Haynes. Haynes had earlier collaborated with Shockley on a “crucial experiment” in measuring the electrical properties of germanium samples, work intended to settle definitively what was the basis for transistor action. While Bardeen and Brattain emphasized phenomena at the surface of the germanium as primary to the workings of the transistor, Shockley was convinced that a phenomenon called “minority carrier injection”—involving the bulk of the germanium, not just the surface—was central. Haynes and Shockley’s experiments, which measured “minority carrier lifetimes” in relatively uniform pieces of crystal from a polycrystalline mass of germanium, demonstrated the reality of minority carrier injection and quickly convinced Bardeen and Brattain.

When Sparks gave Haynes samples of Teal’s new single-crystal germanium in 1949, Haynes acted without delay to repeat his crucial experiment using the new material. The result was bracing. Teal’s material showed minority carrier lifetimes 10 times greater than those in the most uniform material harvested from polycrystalline ingots; it thus promised much improved devices and a clearer understanding of transistor action. Shockley was converted to single-crystal material, and the entire research side of Bell Labs quickly followed suit.

Over the next five years Teal’s methods for making single-crystal germanium—and later single-crystal silicon—served as the foundation for an extraordinary flurry of transistor “firsts” at Bell Labs. Teal collaborated with Sparks in using the crystal-pulling method to create the first of a new breed of transistors, the junction transistor, which had been proposed by Shockley. In 1955 another Bell Labs chemist, Morris Tannenbaum, using silicon made by Teal’s method, made the first double-diffused silicon transistor, a device central to the development of the semiconductor industry for a decade.

Teal left Bell Labs in 1952 for Texas Instruments, where he was instrumental in bringing to the market the first silicon transistor based on crystal-growing expertise. In the year of Teal’s departure for his native Texas, Shockley unequivocally connected the successes of Bell Labs in the new electronics to the object of Teal’s passions: “For the last few years, practically all advances at Bell Telephone Laboratories in transistor electronics and transistor physics have been based on the availability of single-crystal material.” Germanium was useless no more.

For Further Reading

Fite, Robert C. “Germanium, a Secondary Metal of Primary Importance.” Scientific Monthly 78 (1954), 15–18.

Goldstein, Andrew. “Finding the Right Material: Gordon Teal as Inventor and Manager.” In Sparks of Genius: Portraits of Electrical Engineering Excellence, ed. Frederick Nebeker (New York: IEEE Press, 1994), pp. 97–99.

Lécuyer, Christophe; David C. Brock. “The Materiality of Microelectronics.” In manuscript.

Riordan, Michael; Lillian Hoddeson. Crystal Fire: The Birth of the Information Age. New York: W.W. Norton, 1997.

Teal, Gordon K. “Single Crystals of Germanium and Silicon—Basic to the Transistor and Integrated Circuit.” IEEE Transactions on Electron Devices ED-23.7 (July 1976), 621–639.