Moore’s law and the silicon revolution - The Computing Universe: A Journey through a Revolution - page 122

thermal kinetic energy corresponding to vibrations about their positions in the
crystal lattice. As the conduction electrons move through the metal they can
gain or lose energy in collisions with the lattice ions and with each other. Thus,
instead of the conduction electrons filling up only the lowest energy levels in
the band, some electrons will be thermally excited to higher levels in the band and
even to higher bands. This has the effect of leaving some empty energy levels in
the bottom of the lowest band.
For an insulator like carbon, the lowest energy band is full and there is a
large energy gap to the next band ( Fig. 7.3b ). In this case, almost no electrons
are able to gain enough energy from collisions to jump into the empty, higher
energy band. When a voltage is applied to the material, there are therefore no
empty levels close by for the electrons to be able to move to and gain energy,
so the material acts as an insulator. The energy bands in a semiconductor are
shown in Figure 7.3c . These materials have a similar band structure to an insu-
lator with the bottom band filled, but the energy gap to the next band of energy
levels is much smaller. At ordinary temperatures, some electrons are excited by
thermal collisions into the upper conduction band. When a voltage is applied,
the electrons in the upper band have plenty of empty states to move to and
allow the electrons to gain energy. There will also be some empty states in the
lower band that allow conduction. Thus semiconductors will conduct currents
fairly easily, but their conductivity will depend strongly on temperature, in
contrast to metals and insulators.
Two Nobel Prizes: The transistor and
the integrated circuit
Pure semiconductors are not in themselves of great practical importance.
In metals almost every atom contributes one or more conduction electrons but
in semiconductors only one atom in about a thousand million contributes an
electron to conduct electricity. This apparent drawback has the great advan-
tage that the conduction properties of semiconductors can be easily modified
by introducing tiny amounts of additives called impurity atoms - at the level of
around one atom in a million. Germanium and silicon both have four valence
electrons , electrons in the atom's outermost shell that can be easily transferred
to or shared with other atoms. The valence electrons fill up most of the states in
the valence band , which lies below the almost-empty conduction band . If we intro-
duce an impurity such as phosphorous, which has five valence electrons, into
the pure semiconductor only four of these electrons are needed to maintain
the crystal lattice structure. As a result there will be an electron left over that
can easily be detached from the phosphorous atom and contribute to the con-
ductivity. Similarly, if we introduce an impurity atom such as boron, with only
three valence electrons, there will be one electron missing in the bonds that
hold the lattice together. The missing electron creates a site that can capture
electrons from filled states in the valence band, leaving empty states and allow-
ing some conduction. These two situations are represented on the energy level
diagram shown in Figure 7.4 . The process of adding impurity atoms is called
doping . Semiconductors that have been doped with phosphorus are called n-type
semiconductors . The phosphorus atoms give rise to electron donor states just