In their article “The Promise and Challenge of Solid-State Lighting” (Physics Today, December 2001, page 42) Arpad Bergh, George Craford, Anil Duggal, and Roland Haitz show a graph (page 43) of performance of light-emitting diodes as a function of time. The graph obscures the early history by suggesting that GaP:Zn, O devices first appeared in 1968, whereas by that time a number of manufacturers had them on the market. The breakthrough in gallium phosphide came in 1962 when Jerzy Starkiewicz and I discovered that good red emission required both zinc and oxygen to be present, and we mapped the appropriate concentration ranges. 1 That work allowed for the development of devices (we called them “crystal lamps”) with sufficient reproducibility for manufacture. They were described in March 1962 in a journal that was restricted at the time, and a more complete description was given in the open literature later that year. 2
Sample devices were sent to potential military users in May and June 1962. The response was enthusiastic, so toward the end of the year, a proper production line was set up in Phil Gurnell’s semiconductor device group in the Services Electronics Research Laboratory in Baldock, England. One application was a film marker for Royal Air Force reconnaissance planes: An array of emitters put digital information about flight parameters onto each frame of the film. Obviously the array had to be small so that the marker used up very little of the frame. For that, the encapsulated lamps were 1/32 inch (just under 1 mm) in diameter and were mounted within the thickness of a printed-circuit board. That application was secret at the time but was later made public.
