Semiconductor diode luminescence and lasers. A perspective

The perspective covers mostly work in which I was involved at MIT. In 1958, we switched from working on germanium to GaAs. I visited Professor Welker, the expert on GaAs, and he affirmed our decision. We were the first to report p-n junction diodes in GaAs. We used forward-biased luminescence as a diagnostic to determine why the electrical characteristics of diffused diodes were different from those of alloy diodes. We obtained and reported in July 1962 the high-efficiency production of luminescence from the diffused diodes. With this efficiency it was obvious to us and others that a diode laser was feasible. We demonstrated TV transmission over a 50-km path with the diode luminescent source. In October 1962, we had the diode laser. Subsequently, laser action was demonstrated in diodes for many of the III-V direct-gap semiconductors, including in 1964 a longitudinal laser (lasing in the direction of the current) in InSb. In a bootleg operation, laser action was demonstrated in lead salt (PbTe) diodes. The wavelength of the laser emission from the lead salts was tuned further into the infrared by hydrostatic pressure and by going to mixed crystals of PbSn salts. Lead salt lasers were used in infrared spectroscopy and in pollution detection. Returning to GaAs, to demonstrate higher power, individual diode lasers were combined coherently in an external cavity containing a spatial filter, and also with each laser fiber coupled into this cavity. We describe a device research philosophy of working on more projects that expected, failing on some, bootlegging projects if necessary, and asking the question of the results: why. Two failures are described.