Introduction

Many of the things we use today, apart from computers themselves, are computer-controlled. And you would have to go a long way before you came across an electronic device which did not include one transistor.

Indeed, all the mobile telephones, pocket-sized media devices, and other small gadgets that bleep that make our life worthwhile depend upon semiconductor devices. Even mankind's relentless urge to better and more expensive digital watches relies upon them.

A microprocessor, such as in a computer or an MP3 player, for example, may contain MILLIONS of transistors. Even washing machines contain thousands. They rule our lives more than we imagine.

So what is a semiconductor?

The most common semiconductors are Silicon and Germanium. They are called semiconductors because their conductance fits in between the conductors and insulators:

• Mica, an insulator, has a conductance of 5 x 10¹² ohm/cm
• Silicon, a semiconductor, has a conductance of 5 x 10⁴ ohm/cm
• Silver, a conductor, has a conductance of 5 x 10^-5 ohm/cm

One peculiarity of semiconductors, however, is that the conductance of the material goes up as the material gets hotter, rather than decreasing.

Why are they so useful?

So far, it would appear that semiconductors only have a few specialised uses. But they have one more trick up their metaphorical sleeve, and that is doping. The word "doping" in this instance is used in a similar way to in colloquial speech; it refers to the introduction of a foreign chemical into a body so that the body behaves very oddly afterwards.

Since Silicon has a valency of 4, it forms a lattice structure. If the silicon is pure then it will form a perfect lattice. To dope the silicon, you add something with either 5 (e.g. Phosphorus) or 3 (e.g. Boron) electrons in its outer shell, producing either N or P-type silicon. In either of these there are excess charge carriers: in N-type there are extra electrons, in P-type excess holes¹, so both individually are better conductors than pure undoped silicon. This also explains the increase in conductivity with heat, since in semiconductors the benefits of having more charge carriers knocked out of the electron shells of atoms far outweigh the problems caused by the nuclei of the atoms themselves vibrating more and getting in the way.

The P-N Junction

To make all the multitude of devices, however, another concept is needed; that of the P-N junction. If P-type and N-type are created on the same wafer², a P-N junction is formed where they touch. As soon as they do touch, some of the holes from one side cancel out some of the electrons from the other. This has the effect of charging both sides slightly, producing a potential difference which prevents any more charge carriers cancelling out.

Once this junction has been made, it has one unique property: an electric current can only go one way through it. This is because, when the N-type silicon is made negative and the P-type positive, the charge carriers are each attracted by the opposite charge across the potential difference in the middle, so a current flows. However, if the connections are made the other way, that is to say the N-type is made positive and the P-type negative, the charge carriers are attracted back into their own halves of the junction, and so no current flows.

Diodes

A diode is probably the simplest semiconductor device. It consists of a single P-N junction, thus acting as a "one-way system" for DC circuits. It can also be used to rectify AC. They are used extensively as discrete components.

There are, however, lots of variations on this theme, of which some of the most popular follow:

• Light Emitting Diode (LED)
  
  If the substances which dope the semiconductor material in the P-N junction are changed, it is possible to produce P-N junctions that glow when conducting. These have become increasingly used as indicator lights, since they generate little heat and do not burn out in the way that hot-filament bulbs do.
• Zener Diodes
  
  A Zener diode can actually conduct in both directions. In the forward direction, it behaves like a normal diode. In the "backwards" direction, it will not let any current through until the voltage approaches the "Zener Voltage" (specified for the component), when a whole lot of current is allowed through. It is wise to connect a resistor in series with the diode so that it does not melt. The reason these are useful, however, is for a range of supply voltages, the voltage across the diode will remain constant. Therefore, they can be used to ensure that the voltage in a part of the circuit can not rise above a certain level.
• Photodiodes
  
  As in a normal semiconductor, heat can increase the conductivity by increasing the flow of charge carriers, so can light. By careful choice of dopants (and by placing the diode in a transparent case), light can be made to affect the conductivity.
  

The field effect transistor

The transistor is the next level up as regards complexity. There are two types of transistor: bipolar and Field Effect. Discrete transistors are normally bipolar; ICs tend to use FETs. FETs consist of a block of P-type silicon with two nubbins of N-type on the top with a gap between. Normally, no current will flow between the N-types since there are two P-N junctions between them, back-to-back, and one of them will always be reverse biased. However, if you apply a positive charge to the middle of the transistor, between the two nubbins, then it will repel the holes in that area so that a current can flow between the two.

And that was a whistlestop tour through the exciting world of semiconductors. I realise it has been very simplified (it makes no attempt to explain reverse bias breakdown, for example) but I hope it is just enough to get people interested. :).

Any questions?