By 1948, Morton had achieved a reputation as an imaginative engineer who knew how to get sophisticated devices into production. One day in June, Kelly tersely summoned him to his office. Such a one-on-one encounter was not to be taken lightly, for Kelly—who had risen to executive vice president—had a legendary temper. A friend of Morton’s later described the meeting as follows:
“Morton,” Kelly began, “You do know about the work we’ve been doing on the transistor?”
“Yes, sir,” Morton replied a bit hesitantly. He’d heard through the grapevine about the transistor, the secret new solid-state amplifier invented the previous December by Bardeen and experimental physicist Walter Brattain, but he was unsure whether to admit it. “At least I know it’s pretty important.”
“Morton, I’m going to be away for the next four weeks,” Kelly proceeded imperiously. “When I get back, I would like to see a report from you on the transistor. I want you to tell me how to develop it commercially.”
Kelly didn’t know it, but Morton had been running scared of him ever since being hired, and he was now terrified at the prospect of failing on such an enormous assignment. For the next three weeks Morton talked to the scientists and engineers working on transistors but made little progress. Then, in the last week, he somehow pulled it all together and had a 46-page report waiting on the boss’s desk when he returned. Kelly read it, approved it, and immediately put Morton in charge of transistor development.
Bardeen and Brattain’s point-contact transistor was a crude, fragile device consisting of two closely spaced metal points jabbed into a germanium sliver. That it worked at all was a minor miracle. Already a pilot production line at the Labs was turning out hundreds of prototypes every week for further experimentation, measurement, and testing. But the transistors were extremely noisy, variable, and unreliable. “In the very early days, the performance of a transistor was apt to change if someone slammed a door,” Morton was quoted as saying in a 1953 article in Fortune.
In the fall of 1948, he gathered about a dozen engineers and a similar number of technicians for a meeting to launch his transistor development team [see photos, “Team Leader”]. According to two of his lieutenants, Eugene Anderson and Robert Ryder, Morton already seemed to anticipate that their work would make history: “We shall change the world,” he prophesied. “In what manner I do not know, but change it we will.”
The team attacked the noise and reliability problems on several fronts, and by mid-1949 they had two improved versions ready for production, one to amplify signals and the other for switching applications. Western Electric began manufacturing the transistors in 1950, and Bell System engineers soon found uses for them. The first commercial application was in a tone generator used in toll-call signaling.
Morton also had the foresight to pursue what became known as “fundamental development” of basic manufacturing processes and technologies that could have across-the-board implications for the new semiconductor industry. A prime example: when the chemist Gordon Teal could not convince his own department head to let him grow large single crystals of germanium for use in making transistors, he appealed to Morton for support. Coming from electron tube manufacturing, Morton understood that transistor action demanded a near-perfect medium—like the vacuum in a tube—and so he readily came up with the pittance required to buy or build the necessary crystal-growing equipment. Dollar for dollar, it was probably the best investment in Bell Labs’ history.
Morton knew how to get research ideas and designs out of the Labs and into manufacturing at Western Electric, too. Here the groundwork had been laid by his mentor, Kelly, who during the war had recognized the obstacles to transferring technology from an isolated central lab. He felt it was crucial to have people familiar with the latest scientific and technological advances working right on the shop floor, especially when fabricating high-tech components. And so, after 1945 he established a system of branch labs at several Western Electric plants, consisting of teams of Bell Labs employees focused on production engineering and acting as liaison with their colleagues back in Murray Hill.
Morton fine-tuned this approach at the new Western Electric plant in Allentown, Pa., which produced electronic devices and components for the Bell System. He set up a semiconductor development group there and put Anderson in charge. A tube engineer who had been with Morton’s Murray Hill transistor team from the outset, Anderson had a good grasp of the necessary solid-state physics and semiconductor technology.
So Morton’s development team was poised to move quickly in mid-1951 when Bell Labs announced the successful fabrication of the junction transistor, much more rugged and practical than the delicate point-contact device invented by Bardeen and Brattain. Conceived by Shockley and fashioned by chemist Morgan Sparks using Teal’s crystal-growing apparatus, this three-layer germanium sandwich had a much simpler structure than the point-contact transistor and far outperformed it. That the junction transistor would be the preferred path to commercialization was immediately obvious. And Morton’s group led the way, getting the device into production within a year.
In the beginning, transistors were made of germanium, not silicon. Although germanium’s lower melting point makes it far easier to purify, transistors crafted from it are sensitive to temperature changes. And they make lousy switches: tiny leakage currents continue to flow even when the devices are nominally off. Silicon doesn’t have these problems, but it’s a lot more difficult to work with. In the early 1950s, only a few farsighted researchers like Shockley recognized that silicon was the semiconductor material of the future.
In 1954 Morris Tanenbaum fabricated the first silicon transistor at Bell Labs [see “The Lost History of the Transistor,” IEEE Spectrum, May 2004]. Later that year AT&T executives decided to pursue the first electronic switching system—known as ESSâ”1—based on semiconductor devices rather than electromechanical crossbar switches. A trial run was set to begin in 1958.
Morton faced a crucial decision: whether to employ the (by then) well-established germanium technology or bet the house on silicon, which still had a long way to go in development and was thus far riskier. Both Sparks and Tanenbaum, who worked on the research side of the fence and didn’t have to worry about manufacturing devices of extreme reliability, now say in hindsight that the choice to go with silicon was obvious. At the time, however, it was anything but.
In March 1955 Tanenbaum improved on his earlier invention by diffusing impurities into the silicon. This process allowed him to fashion a narrow base layer—the “meat” in the semiconductor sandwich—only about a micrometer thick. The device, which came to be known as a diffused-base transistor, could amplify and switch signals above 100 megahertz, into the range of FM radio and television. Best of all, such a high switching speed, about 10 times that of previous silicon transistors, meant that it could be used for electronic switching.
When he heard the news, Morton was in Europe. He immediately canceled his travel plans and rushed back to Allentown. “On a snowy, miserable day,” recalled Anderson, Morton decreed that ”it was to be in silicon as a material and diffusion as a technology that future transistor and diode development would move in the Bell System.”
His bold decision proved correct. Bell Labs researchers soon resolved the difficulties with purifying silicon and growing crystals of it. They then discovered how to make a glassy, protective oxide layer on the silicon surface that could be used to pattern the impurity diffusions. Fairchild Semiconductor Corp., in Mountain View, Calif., led by Robert Noyce, would adapt these silicon technologies to produce the first commercial microchips in 1961. Western Electric, in turn, used Fairchild’s patented planar process to make diffused-base silicon transistors for the Bell System’s ESS-1, which began to show up in phone exchanges in the early 1960s [see photo, “From Lab to Factory”]. Kelly’s dream of electronic switching finally became reality, thanks in part to Morton’s courage and vision.
