Morgan Sparks
Morgan Sparks was an American scientist and engineer at Bell Labs who, with Gordon Teal, built the first grown-junction transistor in 1950 and who later became president of Sandia Laboratories. Born in 1916, he lived to 2008, and his long career runs from the wartime and postwar laboratory work that made the junction transistor real through the management of one of the United States' principal national-security engineering laboratories.
His place in transistor history rests on turning Shockley's junction theory into a working device. Shockley had proposed in 1948 that a sandwich of n-type, p-type and n-type (or p-n-p) semiconductor, with a very thin middle layer, should amplify, and that it should do so more cleanly and with less noise than the point-contact transistor demonstrated in December 1947. The difficulty was making such a structure: layers had to be uniform, thin and free of grain boundaries. Sparks worked in the Bell Labs group with Teal, whose crystal puller supplied single-crystal germanium, and the two of them showed that dopants added to the melt during pulling could create the layers inside a single growing crystal. In 1950 they produced working n-p-n grown-junction transistors from such crystals, and the result was written up in a 1951 paper by Shockley, Sparks and Teal, p-n junction transistors, as proof that the theory could be built.
The division of labour between Sparks and Teal is part of why both names appear together. Teal and John Little provided the crystal-growing apparatus and the single-crystal germanium, while Sparks, with a talent for device fabrication, took the grown crystals and made the transistors: cutting bars, locating the thin p layer, attaching fine contacts and measuring the results. That pairing of a materials chemist and a device builder is typical of how Bell Labs turned science into hardware, and it is also why the first junction transistor does not have a single inventor in the sense that the point-contact device, credited to Bardeen and Brattain with Shockley's theoretical role, is often described. The same method was soon moved to Western Electric for production and to licensees after the 1952 symposium.
After the transistor he rose through Bell Labs management and then moved to the laboratory's national-security cousin. Sparks went on to lead device and electronics work at Bell Labs, a period in which the laboratory developed diffused-base transistors and, in 1954, the silicon solar cell. In 1972 he became president of Sandia Laboratories in Albuquerque, New Mexico, which was operated for the Atomic Energy Commission and its successors by a Western Electric subsidiary and later AT&T, and he held the post until 1981. In that role he oversaw a large engineering organisation that designed non-nuclear components for nuclear weapons, among other programmes, a long way from the quiet crystal puller of 1950 yet reliant on the same culture of rigorous, systematic engineering.
It is a common mistake to credit Sparks with the first transistor or to call his device a point-contact one. The first transistor, the point-contact device of December 1947, belonged to Bardeen and Brattain. The grown-junction transistor was the second major step, and its significance lies in showing that a junction device could be built reproducibly and that its behaviour matched theory. A second pitfall is to assume that grown junctions remained the standard method; they gave way within a few years to alloy-junction, surface-barrier and then diffused and planar processes, which allowed far thinner base regions and much higher frequencies. The grown junction is best seen as the proof-of-principle that unlocked the field rather than the end of its technology.
Modern relevance is easy to miss because the process itself has vanished from production. Contemporary transistors are made on silicon wafers by lithography, diffusion or implantation and thin-film deposition, not by changing the melt as a crystal is pulled. What survived is the idea that the same material can be turned into regions of different conductivity types by controlling impurities, which underlies every diode and transistor, and the habit of pairing materials scientists with device engineers that Sparks and Teal exemplified.
Read Morgan Sparks as the device builder who took Teal's crystals and Shockley's theory and showed on a laboratory bench in 1950 that a junction transistor was not merely a good idea but a working part, before going on to run one of the nation's largest engineering laboratories.