I remembered hearing about how, before the transistor was invented in
1947, people were actually using the principle back in the 1920s or
Recently, I came across a web site giving some details.
is also of interest.
In the 1920s, a Russian, O. Lossev, constructed a sort of tunnel diode
using point contact with zincite (ZnO, Zinc Oxide). And before that,
radio operators on board ships often used two crystals for their
radios - a galena crystal (PbS, lead sulfide) in the conventional way,
and a carborundum crystal (SiC, silicon carbide) used as what appears
to be, from the description, a bipolar transistor.
This doesn't detract from Bardeen, Brittain, and Shockley's invention
of the transistor, since if a transistor were something that had to be
delicately adjusted with a "cat's whisker", devices using many
transistors would be impossible. Only when a transistor could be
manufactured, and bought off the shelf, like a vacuum tube, could it
be envisaged as replacing the vacuum tube.
But it's interesting that silicon carbide, perhaps the material used
in the very first "transistors", is being considered today for
electronics. It has the advantage of being a very rugged material that
can operate at high temperatures. For the moment, however, it is
difficult to make wafers that are free of defects. (Tunnel diodes, as
discrete devices, had very low yields, which I see has been noted as
the reason they're not seen on ICs for fast computing.)
Silicon carbide has a larger band gap than silicon - and lead sulfide
a smaller band gap than germanium. Running a chip at a lower voltage
has advantages for power consumption and density. Since lead sulfide
is used for some IR vidicon tubes, I suppose there are companies
making lead sulfide wafers.
I see that wafers _are_ being made out of a number of materials these
days. Silicon carbide is used for blue lasers, lead sulfide for IR
detectors, and many other semiconductors have their own specialized