Shockley Semiconductor Laboratory, the primary lab of the Shockley Transistor Company, was the first company to work on silicon semiconductor devices in what came to be known as Silicon Valley. It was purchased by Clevite in 1960, and officially closed shortly after being sold to ITT in 1968. Engineers leaving the company stayed in the area; engineers leaving these companies did the same, and soon an entire industry built up in the San Francisco Bay Area.
Contents |
William Shockley had studied his undergraduate degree at Caltech and moved east to complete his PhD at MIT. He graduated in 1936 and immediately started work at Bell Labs. Through the 1930s and 40s he worked on electron devices, and increasingly with semiconductor materials. This led to the 1947 creation of the first transistor, in partnership with John Bardeen, Walter Brattain and others. Through the early 1950s a series of events led to Shockley becoming increasingly upset with Bell's management, and especially what he saw as a slighting when Bell promoted Bardeen and Brattain's names ahead of his own on the transistor's patent. However, others that worked with him suggested the reason for these issues was Shockley's abrasive management style, and it was this reason that he was constantly passed over for promotion within the company. These issues came to a head in 1953. and he took a sabbatical and returned to Caltech as a visiting professor.
Here Shockley struck up a friendship with Arnold Orville Beckman, who had invented the pH meter in 1934. By this time Shockley had become convinced that the natural capabilities of silicon meant it would eventually replace germanium as the primary material for transistor construction. Texas Instruments had recently started production of silicon transistors (in 1954), and Shockley thought he could do one better. Beckman agreed to back Shockley's efforts in this area, under the umbrella of his company, Beckman Instruments. However, Shockley's mother was aging and often ill, and he decided to live closer to her house in Palo Alto.
The Shockley Semiconductor Laboratory opened for business in a small commercial lot in nearby Mountain View in 1956. Initially he tried to hire some of his former workers from Bell Labs, but none of them wanted to leave the east coast, then the center of most high-tech research. Instead, he assembled a team of young scientists and engineers and set about designing a new type of crystal-growth system that could produce single-crystal silicon boules, at that time a difficult prospect given silicon's high melting point.
While work on the transistors continued, Shockley hit upon the idea of using a four-layer device (transistors are three) that would have the novel quality of locking into the "on" or "off" state with no further control inputs. Similar circuits required several transistors, typically three, so for large switching networks the new diodes would greatly reduce complexity.[1][2] The four-layer diode is now called the Shockley diode.
Shockley became convinced that the new device would be just as important as the transistor, and kept the entire project secret, even within the company. This led to increasingly paranoid behavior; in one famed incident he was convinced that a secretary's cut finger was a plot to injure him and ordered lie detector tests on everyone in the company. This was combined with Shockley's vacillating management of the projects; some times he felt that getting the basic transistors into immediate production was paramount, and would de-emphasize the Shockley diode project in order to make the "perfect" production system. This upset many of the employees, and mini-rebellions became commonplace.[3]
Eventually a group of the youngest employees went over Shockley's head to Arnold Beckman, demanding that someone be brought in to replace Shockley in the management role. Beckman initially appeared to agree with their demands, but over time a series of decisions where Beckman sided with Shockley made it clear that nothing had actually changed. Fed up, the group turned to Sherman Fairchild's Fairchild Camera and Instrument, an Eastern U.S. company with considerable military contracts. In 1957, Fairchild Semiconductor was started with plans on making silicon transistors. Shockley denoted them the "traitorous eight":[4][5]Julius Blank, Victor Grinich, Jean Hoerni, Eugene Kleiner, Jay Last, Gordon Moore, Robert Noyce, and Sheldon Roberts.
The eight would later leave Fairchild to start companies of their own, among them Intel, Advanced Micro Devices and others. Over a period of 20 years, 65 different companies were started by 1st or 2nd generation teams that traced their origins in the valley to Shockley Semiconductor.
Shockley never managed to make the four-layer diode a commercial success, in spite of eventually working out the technical details and entering production in the 1960s. The introduction of integrated circuits allowed the multiple transistors needed to produce a switch to be placed on a single "chip", thereby nullifying the parts-count advantage of Shockley's design. However, the company did have a number of other successful projects, including the first strong theoretical study of solar cells, developing the seminal Shockley-Queisser limit that places an upper limit of 30% efficiency on basic silicon solar cells.
