-
1. How a Bad Idea Became Good (to Some): THE EMERGENCE OF THE MOS TRANSISTOR, 1945 – 1963
- Johns Hopkins University Press
- Chapter
- Additional Information
12 ONE How a Bad Idea Became Good (to Some) T H E E M E R G E N C E O F T H E MOS T R A N S I S T O R , 1 9 4 5 – 1 9 6 3 The story of the origins of the MOS transistor is the story of how a bad idea became good (to some). Even though the MOS transistor showed striking resemblances to structures Bell Labs scientists had proposed in their investigations that led to the invention of the transistor in 1947, the formal announcement of the MOS transistor’s existence as a potentially useful technology did not occur until 1963. In the intervening years semiconductor technology had matured, but it had done so technically by repudiating the principles behind Bell’s early semiconductor work, which had tried to construct a semiconductor amplifying device that operated on the surface of the semiconductor. Instead, Bell and the industry it created followed another path, which was blazed by William Shockley. Shockley invented the bipolar junction transistor, whose genius was that it bypassed the seemingly intractable problems of the semiconductor surface—problems that had plagued Bell’s initial work. By the late 1950s, his transistor and variations on it dominated semiconductor technology . Bell Labs’ concentrated efforts led to advances that made Shockley’s transistor faster, more reliable, and able to handle higher power levels—some of the key measures by which engineers judged a transistor. When researchers first described what is now known as the MOS transistor 13 How a Bad Idea Became Good (to Some) at a professional conference, it provoked very little interest from other semiconductor researchers. This response was hardly surprising, for in significant ways the MOS transistor was a return to the bad old days when the semiconductor surface controlled the performance of a transistor. Furthermore, by conventional measures, the MOS transistor seemed to lack any offsetting advantages. The MOS transistor was substantially slower than the bipolar transistor and much less reliable, making it unusable in the most critical bipolar transistor applications. Two new developments made something like an MOS transistor seem promising to some in the early 1960s, when it had seemed pathological before. First, the use of silicon dioxide to passivate silicon surfaces gave hope that the problems of the semiconductor surface could be mitigated. Second, the advent of integrated circuits brought a new measure by which to judge transistors. Speed was still important, and the MOS transistor was slow, but the difficulties of trying to build multiple components on a single chip made the simplicity of the MOS transistor attractive to some. They believed that if the problems of the semiconductor surface could be solved, it would be possible to put many more MOS transistors than bipolar transistors on a chip. This vision was not compelling to everyone. The MOS transistor was the work of outsiders. The researchers who conceived of it and worked on it at Bell Labs, RCA, and Fairchild were by and large newcomers to semiconductor technology with little stake in Shockley’s bipolar junction transistor. Although the MOS transistor was initially conceived at Bell Labs, the organizational commitment to the older bipolar transistor was too great for the new technology to take root, and management quickly halted work on the MOS transistor, dismissing it as unpromising. However, the expansive nature of the R&D environment at RCA and Fairchild provided a space where the new technology could flourish. By 1963, even though the MOS transistor was still not ready to be sold as a product and many doubted that it would ever be successful, RCA, Fairchild, and other companies had seen its promise clearly enough to devote substantial R&D resources to it. [54.157.61.194] Project MUSE (2024-03-19 13:45 GMT) 14 T O T H E D I G I TA L A G E Bell Labs and the Bipolar Junction Transistor: Commitment to a Technology Bell Labs invented the transistor in 1947, and by 1955 its development effort had progressed to the point where it had made a major commitment to one particular type of transistor technology, the silicon diffused bipolar junction transistor. This commitment made sense at a number of different levels. Economically it was a good fit with the needs of the Bell System, Bell Labs’ prime customer, as it met the requirements of many different systems and offered low costs. Technically it had a number of favorable attributes. It avoided the worst...