What are semiconductors?  

Explanation

A semiconductor behaves as an insulator at very low temperatures, and has an appreciable electrical conductivity at room temperature although much lower conductivity than a conductor. Commonly used semiconducting materials are silicon, germanium, gallium arsenide, indium phosphide, and mercury cadmium telluride.

A semiconductor can be distinguished from a conductor by the fact that, at absolute zero, the uppermost filled electron energy band is completely filled in a semiconductor, but only partially filled in a conductor.

The distinction between a semiconductor and an insulator is slightly more arbitrary. A semiconductor has a band gap which is small enough such that its conduction band is appreciably thermally populated with electrons at room temperature, whilst an insulator has a band gap which is too wide for there to be appreciable thermal electrons in its conduction band at room temperature.

How semiconductors work

Operationally, transistors and vacuum tubes have similar functions; they both control the flow of electrical current.

An absolutely pure silicon crystal is also an insulator, but when an impurity e.g. arsenic is added (called doping) in quantities minute enough not to completely disrupt the regularity of the crystal lattice, it donates free electrons and enables conduction. This is because arsenic atoms have five electrons in their outer shells while silicon atoms have only four. Conduction is possible because a mobile carrier of charge has been introduced, in this case creating n-type silicon ("n" for negative: the electron has a negative charge).

Alternatively, silicon can be doped with boron to make p-type silicon, which also conducts. Because boron has only three electrons in its outer shell, another kind of charge carrier, called a "hole", is formed in the silicon crystal lattice.

In a vacuum tube, on the other hand, the charge carriers (electrons) are emitted by thermionic emission from a cathode heated by a wire filament. Therefore, vacuum tubes cannot generate holes (positive charge carriers).

Note that charge carriers of the same polarity repel one another so that, in the absence of any force, they are distributed evenly throughout the semiconductor material. However, in an unpowered bipolar transistor (or junction diode) the charge carriers tend to migrate towards a P/N junction, being attracted by their opposite charge carriers on the other side of the junction.

Increasing the doping level increases the semiconductor conductivity, providing that the crystal lattice, overall, remains intact. In a bipolar transistor the emitter has a higher doping level than the base. The ratio of emitter/base doping levels is one of the main factors that dictates the junction transistor's current gain.

The level of doping is extremely low: in the order of parts per one hundred million, and this is the key to semiconductor operation. In metals, the carrier population is extremely high: one charge-carrier per atom. In metals, in order to convert a significant volume of the material into an insulator, the charge carriers must be swept out by applying a voltage. In metals this value of voltage is astronomical: more than enough to destroy the metal before it converts to an insulator. But in lightly-doped semiconductors there is only one mobile charge carrier per million or more atoms. The level of voltage required to sweep so few charge-carriers out of a significant volume of the material is easily reached. In other words, the electricity in metals is incompressible, like a liquid, while in semiconductors it behaves as a compressible gas. Doped semiconductors can be rapidly changed into insulators, while metals cannot.

The above explains conduction in a semiconductor by charge carriers, either electrons or holes, but the essence of bipolar transistor action is the way that electrons and holes seemingly make a prohibited leap across the insulating depletion zone in the reverse-biased base/collector junction under control of the base/emitter voltage. Although a transistor may seem like two interconnected diodes, a bipolar transistor cannot be made simply by connecting two discrete junction diodes together. To produce bipolar transistor action they need to be fabricated on the same crystal, and physically sharing a common and extremely thin base region.

Fundamental semiconductor physics

Band structure of a semiconductor

Band structure of a semiconductor showing a full valence band and an empty conduction band. The Fermi level lies within the bandgap