by Jon Gertner
To wander into the stately entrance of Bell Labs’ Murray Hill building in mid-November 1947—into the new building’s airy art deco lobby, up a short flight of wide stairs, past the guard station, up eight half flights of granite stairs bounded by a smooth, continuously curving wooden handrail, turning off the fourth-floor landing to walk down the impossibly long, brightly lit, polished red terrazzo hallway, and then on toward Walter Brattain’s laboratory, room 1E455—was to wander into a realm of opaque chatter and experimentation. Brattain’s lab was furnished with workbenches weighted with bulky calibration equipment, all of it adorned with myriad dials and meters. The circulating air was often infused with a sharp burn of smoke from cigars, cigarettes, and soldering irons. The men there spoke of their concerns in another language: “dipoles” and “space-charge layers” and “hysteresis” on the surface of the small semiconductor crystals they bent over to work on. But in large part their quest had come down to how to clean or prepare the surface of their materials—they were at that point using many varieties of silicon—so that either the electrons, or the holes within, or both, would move freely when a current passed through the silicon.
On November 17, Brattain and an electrochemist in the solid-state group, Robert Gibney, explored whether applying an electrolyte—a solution that conducts electricity—in a particular manner would help cut through the surface states barrier. It did. Shockley would later identify this development as a breakthrough and the beginning of what he called “the magic month.” In time, the events of the following weeks would indeed be viewed by some of the men in terms resembling enchantment—the team’s slow, methodical success effecting the appearance of preordained destiny. For men of science, it was an odd conclusion to draw. Yet Walter Brattain would in time admit he had “a mystical feeling” that what he ultimately discovered had been waiting for him.2
Within a week of the experiment with electrolytes, Brattain recalled, “Bardeen walked into my office one morning” and suggested a “geometry” for building a solid-state amplifier. “I said, ‘Let’s go to the laboratory and do it.’” It involved a drop of electrolyte fluid, and a “point-contact” wire piercing the drop and touching the surface of the semiconductor slice. With Bardeen sometimes looking over his shoulder, Brattain and a lab assistant began building a rough prototype—a slab of silicon with a metal point pushed down into it. After some early tests of running a current through the point and into the slab setup, the contraption indicated a slight gain in power. This seemed promising. That night—a Friday, November 21—Brattain told the members of his carpool that he had taken part in the most important experiment in his life. Over the weekend, though, he had second thoughts, worrying that if he and Bardeen were indeed onto something it was of such importance that it must not become gossip. The next day he rode with the group—Monday—he asked them to forget what he had told them about his semiconductor work. “I had to swear them to secrecy,” he said.3
Bardeen, too, seemed to think they were close; he spent the weekend making notations in his notebook outlining what might be happening within the silicon crystal. The two men let Shockley know about the recent developments. Then Brattain came down with the flu.
THE DEVICE THAT BRATTAIN was fiddling with in his lab by mid-December was a small and inelegant thing: a tiny slice of semiconducting material, about one-fourth the size of a penny, lying flat on a metal base. At Bardeen’s suggestion they had switched from using silicon to n-type germanium. A wire was connected to the base. What looked like a tiny arrowhead—it was actually a small, triangular piece of plastic—pushed down into the top face of the germanium slice. But in fact the plastic arrowhead wasn’t touching the germanium. Brattain had wrapped the edge of the arrowhead in a thin gold foil—“some of the prewar gold that I still had around,” he said later—and by doing so had created a gold, V-shaped wire. With a razor he had then cut a slice in the golden V, precisely at the sharp end of the V, so that there was an almost imperceptible gap. Rather than a V-shaped wire, in other words, he had thus created two separate wires—“points,” as they were called at the time—that were pushing into the germanium. Bardeen had told Brattain that a narrow space between the two gold points was essential. He estimated that they could create an amplifier if the points were spaced only a few thousandths of an inch apart.
Brattain then connected the top ends of each of the points, located at the top corners of the arrowhead, to separate wires, each of which led to separate batteries. He was forming a simple circuit. The men had been experimenting for the past two weeks with various setups—they had tried different metals (n-types and p-types of germanium and silicon), different types of points (with different distances between them), sundry chemical preparations, and a variety of electrolyte solutions. Each configuration had yielded new insights and small degrees of amplification. And by the morning of December 16 it had come to this. Brattain connected it up. “I found if I wiggled it just right,” he explained later, so that both points of the gold had contact with the germanium slab, he could get a remarkable effect. What was happening within the device was exceedingly complex—a subatomic movement of electrons and holes brought about by the impurities in the metal and the current introduced through the gold points. At the time, even Brattain and Bardeen didn’t understand it. But the net amplification was the important fact. Years later, William Shockley would comb through Brattain’s old notebooks in an effort to tell the story of how this amplifier had been born. It was on December 16, Shockley concluded, that Brattain’s entries demonstrated the real achievement: a very significant power gain.
Over the next few days, Brattain and Bardeen refined their device in preparation for a demonstration to the Bell Labs management. It was scheduled for the afternoon of December 23, 1947. Mervin Kelly was not invited. A believer in granting a degree of autonomy to researchers, he had not asked about, and had not been kept apprised of, Bardeen and Brattain’s work. What’s more, there was a tendency at Bell Labs to confine important developments to middle management for a purgatorial period, lest word of a breakthrough reach upper management too soon. The concern was that research that appeared to be important could turn out, upon closer inspection, to be nothing of the sort. Thus the practice was for a supervisor to move any big news up a step—a week or two at a time, in Brattain’s recollection—only after he was convinced of its importance. The worst scenario would be telling Kelly without being sure.4
On Christmas Eve 1947, Brattain wrote an account of the previous day’s events in his notebook. Amid the lined pages, he made a quick drawing of the amplifier he had built. He drew another diagram of the device after he and some colleagues had arranged for a voice signal to pass through the circuit that ran through this ugly little amplifier. Brattain wrote, “This circuit was actually spoken over and by switching the device in and out a distinct gain in speech level could be heard.” There was no noticeable change in quality. The power gain was eighteen times or more. “Various people witnessed this test and listened,” Brattain added. In the room that day were Bardeen, Brattain, and Shockley; Brattain’s lab partner Pearson; R. B. Gibney, the chemist who had helped; and H. R. Moore, a circuit expert who had also pitched in. Harvey Fletcher, the head of physical research—who along with Mervin Kelly had come from Millikan’s labs so many years before and was finishing his career now at the Labs—was also present. As with all important entries in the scientists’ notebooks, Brattain’s entry ended with a signature and verification by third parties: Read & understood by G. L. Pearson Dec. 24, 1947 and H. R. Moore Dec. 24, 1947.
Another man in the room that afternoon was Ralph Bown, the vice president of research. Bown was an affable and well-regarded manager who had been promoted to research chief when Kelly moved up to become the Labs’ executive vice president. Bown was more open and more approachable than Kelly. At Murray Hill a large arcade connected the two main buildings. Bown would often hold court in that arcade, and people would come by to chat with him.5 At Bardeen and Brattai
n’s demonstration, Bown was the most important and skeptical of the attendees. He reacted to what he saw with a challenge. “The acid test of an amplifier,” Bown declared, “is whether it can be made to oscillate.”6 In other words, if it is believed that a device can truly produce more power than it takes in—the very definition of an amplifier—then there is a way to check its authenticity. It is done by changing the circuit so that the wires are arranged in a certain manner (the output is “fed back” into the input). What then gets produced is a consistent wavering—that is, oscillating—signal, like a sine wave. In communications systems, oscillating elements are fundamental: They form the basis for everything from a telephone’s dial tone to the broadcasting of radio waves. At Bown’s request, Brattain and his colleagues soon made their amplifier do exactly that. And at that point the research boss was satisfied. In Bown’s personal account of the episode, he admitted that he was immediately convinced the device was not only real but important. It was, he privately suspected, more than an amplifier, more than a switch, more than a replacement for the vacuum tube. But on Christmas Eve, with the year almost over and with a powerful snowstorm bearing down on the East Coast, knowing precisely what it portended, or indeed how it all fit together, would require some serious consideration, as well as some time.
Bardeen and Brattain’s device, Bown simply noted, was “a basically new thing in the world.”
THE NEW THING NEEDED a new patent. In early 1948, probably a month after the invention, Mervin Kelly was informed of the breakthrough.7 At that point, Bardeen and Brattain began working with the Bell Labs lawyers on assembling an application. The two men—Brattain especially—worried that other scientists in the United States or abroad might patent a similar device first. One possibility was the research team at Purdue University, which had been experimenting with semiconductors all through the war. Any Bell scientist knew about the spooky and coincidental nature of important inventions. The origins of their entire company—Alexander Bell’s race to the patent office to beat Elisha Gray and become the recognized inventor of the telephone—was the textbook case.
The new thing needed a new name, too. A notice was circulated to thirty-one people on the Bell Labs staff, executives as well as members of the solid-state team. “On the subject of a generic name to be applied to this class of devices,” the memo explained, “the committee is unable to make [a] unanimous recommendation.” So a ballot was attached with some possible names. “Triode” seemed to have a natural appeal, since the new device had three main elements (the two points and the base), much like a vacuum tube had three main elements (the cathode, anode, and grid). Vacuum tubes, in fact, were often called triodes. The recipients were asked to number, in order of preference, the possibilities:
__Semiconductor Triode
__Surface States Triode
__Crystal Triode
__Solid Triode
__Iotatron
__Transistor
____(Other Suggestion)
Bell Labs engineers had become fond of the suffix “-istor”: Small devices known as varistors and thermistors had already become essential components in the phone system’s circuitry. “Transistor,” the memo noted, was “an abbreviated combination of the words ‘transconductance’ or ‘transfer,’ and ‘varistor.’”8 To the company brass, the other names had some winning aspects, too. Iotatron, for instance, “satisfactorily conveys the sense of a minute element.” Semiconductor triode, the memo noted, was a “fairly good name,” if a bit unwieldy. But when the ballot results came in, transistor was the clear winner.
This so-called transistor needed, finally, to be better understood. “The most gifted electronic engineers from all parts of our laboratory,” Bown, the head of research, recalled, “were brought into a concerted plan to study the device from all angles and to use it to do the various things an amplifier should be able to do.” It was true, as well as obvious, that those minute impurities within the germanium had helped to amplify a voice signal, just as a vacuum tube might. But how the holes, and electrons, moved within the semiconductor slab when a voltage was applied needed far more study.
In early May 1948 the transistor was officially designated a classified Bell Telephone Laboratories technology—“BTL Confidential.” To those with knowledge of the device, Ralph Bown sent around a lengthy protocol to instill an envelope of security.9 The transistor work, and anything relating to it, was given a code name: Surface States Phenomena. Yet there were doubts about how long the Labs could maintain secrecy—or even how long it should. Almost from the start, Labs executives agreed that they should show the device to the military before any public debut but should try to resist any orders to contain the device as a military secret. On the other hand, the executives—Kelly, Bown, et al.—doubted they could keep the transistor rights to themselves once the device became public knowledge. For one thing, AT&T maintained its monopoly at the government’s pleasure, and with the understanding that its scientific work was in the public’s interest. An audacious move to capitalize on the transistor, should it turn out to be hugely valuable, could well invite government regulators to reexamine the company’s civic-mindedness and antitrust status. What’s more, sharing the technology with competitors in the electronics field might be a positive development. AT&T could earn licensing fees from the patent. And if Bell Labs could gain a head start of a few months, it could take a lead over the competition and reap further rewards as a host of outside engineers and scientists worked to improve its functionality. There was in this strategy “a modicum of self interest,” according to Bown. “Who is in a better position than the originator to recognize and profit from further advances?”
At 9 a.m. on May 26, 1948, Bardeen, Brattain, Shockley, and a number of Bell staffers held a meeting on the transistor in Ralph Bown’s spacious Murray Hill office. A noticeable degree of paranoia was setting in. The men talked about the possibility that others, perhaps at Purdue, were “skirting close to this subject in their work” and would announce it first.10 There was agreement that they didn’t want to maintain secrecy any longer than it would take to secure their patents, which were almost ready for application. The men decided to continue apace with the patent application and that Bardeen and Brattain would quickly write a letter to the journal Physical Review describing their device—in effect staking out their ground in the scientific community. It was also decided that the Labs’ public relations department would begin writing a press release and discuss ideas for a public presentation.11
By mid-June 1948, the patent application on the transistor had been filed and a press conference on the new device had been scheduled for the end of the month at the large auditorium in Bell Labs’ West Street offices in Manhattan. Crafting the news release had been something of a nightmare, with at least a half dozen cooks (Shockley, Bardeen, Bown, and Brattain included) adding to the broth. “It has been rewritten at least N times, where N > 6,” Shockley noted.12 The transistor was not much larger than the tip of a shoelace, the news release said, and more than a hundred of them could fit easily in an outstretched hand. Yet it was quite complicated. The explanation of the device took up seven typed pages.13 Bardeen and Brattain’s letter to the Physical Review that announced their breakthrough, meanwhile, was impenetrable to all but an accomplished solid-state physicist. If anyone really wanted to know what the scientists had accomplished over the past few years, they would need a world-class understanding of metallurgy, quantum physics, and electrical engineering.14
WHEN HE WAS A TEENAGER in the 1920s, William Shockley took a high school mathematics exam that haunted him for an entire weekend. “It reflects a personality aspect which may be common to a number of persons who work hard at technical endeavors,” he later explained. It wasn’t that Shockley felt he had failed his math test. “I was especially concerned about my relative grade to that of another student—one who was not at all remarkable,” he said. “All weekend I worried over whether I had done better than he.”15
As
Bardeen and Brattain closed in on the transistor in the autumn of 1947, Shockley, their supervisor, had become increasingly interested in their work, sometimes offering suggestions and periodically receiving updates from the men. When the breakthrough came in December, Shockley would admit to a complex set of reactions—“I must confess to a little disappointment that I hadn’t been more personally involved in it,” he later admitted.16 On another occasion he conceded, “My elation with the group’s success was balanced by not being one of the inventors.”17 Perhaps far more than Bardeen and Brattain, Shockley understood the implications of the new device; he had been dreaming of its possible existence since Kelly had visited his office, so many years before, to talk about his idea of an electronic switch.
Walter Brattain would later recall that shortly after the December 23 transistor demonstration for the Bell Labs executives, Shockley called both Bardeen and Brattain into his office, separately, to say, “Sometimes the people who do the work don’t get the credit for it.”18 In Brattain’s telling, Shockley seemed confident that he himself could write a patent that covered his field effect idea and that would overshadow subsequent work done by his two colleagues. This would prove to be impossible for a number of reasons. First, many people at the Labs already knew that Bardeen and Brattain had built the first transistor together, and their lab notebooks could verify it. But in a patent search on Shockley’s field effect, it also appeared that an inventor named Julius Lillienfield had come upon a similar idea two decades before. There was little evidence that Lillienfield had made a working model of his device, or that the device outlined in his patent would work. What’s more, there was no likelihood that he had any theoretical understanding of semiconductors, let alone a knowledge of the movements of holes and electrons at the subatomic level. Still, legally speaking, Lillienfield had been there first. And—though they used a different approach—so had Bardeen and Brattain.