Page 31 - A Practical Guide from Design Planning to Manufacturing

The Evolution of the Microprocessor 7

to use a metal plate placed near the surface of a semiconductor, but not
touching, to control the flow current within the semiconductor. He
believed the electric field from the plate could create a channel of charge
carriers within the semiconductor, allowing it to conduct dramatically
more current than possible when the field was not present. The idea
seemed sound, but when Shockley built an apparatus to test his theory,
it didn’t work. Even with 1000 V on the metal plate, he could not measure
any amplification at all.
Shockley gave this puzzle to two of the researchers on his team. John
Bardeen was one of very few men in the world at the time with a thor-
ough grasp of the physics of semiconductors, and Walter Brattain was
an experienced experimentalist with a reputation as a man who could
build anything. Together these two began experiments to determine
why Shockley’s “field-effect” device did not work. They came to believe
that the problem was charges trapped at the surface of the semicon-
ductor. If the semiconductor naturally formed a slight excess of electrons
at its surface and these charges were unable to move under the influ-
ence of an electric field, they would prevent an electric field from pen-
etrating the surface and creating amplification. Bardeen and Brattain
began looking for a way to neutralize these surface charges. They tried
placing the semiconductor in various liquids and got some improve-
ment but never with enough amplification at a high enough frequency
to be useful.
Bardeen and Brattain had a piece of germanium (another column 4
semiconductor) prepared with a thin oxide layer on top. They hoped the
oxide would somehow neutralize the surface charges. Brattain care-
fully cleaned the crystal and began testing. At first his results made no
sense at all. The device responded as if the oxide layer wasn’t even
there. Then to his horror, Brattain realized that in cleaning the crystal
he had accidentally removed the oxide layer. Germanium oxide is solu-
ble in water. Without the oxide layer what he had created was a number
of cat whisker diodes on the same piece of semiconductor. Frustrated,
he continued experimenting with the device anyway and was surprised
to find some amplification between the diodes. Perhaps neutralizing
the surface charges wasn’t necessary at all.
They decided to try two cat whisker diodes touching the same semi-
conductor crystal at almost the same point. In December 1947, they
wrapped a piece of gold foil around a triangular piece of plastic and
then used a razor blade to make a cut in the gold at the point of the tri-
angle. By touching the point of the triangle to a piece of germanium crys-
tal, they created two cat whisker diodes. They discovered that electrons
emitted into the semiconductor by one diode were collected by the other.
The voltage on the crystal base could increase or suppress this effect
allowing their device to amplify the signal at the base. Because a control