Conductors, insulators and semiconductors
Every diode, transistor and integrated circuit on an aircraft is made from a semiconductor — a material whose ability to conduct lies between that of a metal and an insulator, and, more usefully, can be controlled. To see why, start with the outer shell of the atom. The electrons in that shell, the valence electrons, decide how a material behaves electrically.
- Conductors (copper, silver, aluminium) have one to three loosely held valence electrons. At room temperature huge numbers of them drift freely through the metal.
- Insulators (glass, mica, ceramics, most plastics) have full or nearly full valence shells; their electrons are tightly bound and very few are free.
- Semiconductors have four valence electrons. The important ones are silicon (by far the most common), germanium (the material of early transistors) and compound semiconductors such as gallium arsenide, gallium nitride and silicon carbide, used for LEDs, microwave devices and power electronics.
The energy-band picture explains the difference. Free electrons sit in the conduction band; bound electrons sit in the valence band. Between them is the band gap — energy an electron must gain before it can move. Silicon's gap is about 1.1 electron-volts and germanium's about 0.67 eV; an insulator's gap is more than about 5 eV.
Heating a pure semiconductor gives more electrons enough energy to cross the gap, so it conducts better: its resistance falls as temperature rises — a negative temperature coefficient, the opposite of a metal. NTC thermistors use exactly this effect.
The silicon crystal: electrons and holes
In a silicon crystal each atom shares one of its four valence electrons with each of four neighbours. Each shared pair is a covalent bond, and the bonds lock the atoms into a regular lattice. With every valence electron held in a bond, pure silicon at absolute zero is an insulator.
At room temperature, thermal energy breaks a few bonds. Each freed electron moves into the conduction band and leaves behind a gap in its bond, called a hole. A neighbouring bound electron can jump into the hole, which moves the hole the other way. Holes therefore behave like positive charge carriers. In pure (intrinsic) silicon, electrons and holes are always created in pairs, so their numbers are equal. They also keep meeting and recombining.
When a voltage is applied, the free electrons drift towards the positive terminal and the holes drift towards the negative terminal. The total current is the sum of the two. Intrinsic silicon has very few carriers, though, so it is a poor conductor — not much use until it is doped.
Doping: N-type and P-type material
Doping means adding a tiny, controlled amount of a chosen impurity — typically a few atoms per million silicon atoms or less — to the crystal. It changes conductivity enormously, and it decides which kind of carrier dominates.
| N-type | P-type | |
|---|---|---|
| Impurity | Pentavalent (5 valence electrons): phosphorus, arsenic, antimony | Trivalent (3 valence electrons): boron, gallium, indium, aluminium |
| Name of impurity | Donor — it donates a free electron | Acceptor — it accepts an electron into its hole |
| Majority carriers | Free electrons | Holes |
| Minority carriers | Holes (made by heat) | Electrons (made by heat) |
| Fixed ion left behind | Positive donor ion | Negative acceptor ion |
Each donor atom brings its own extra proton along with the extra electron, so a piece of N-type (or P-type) material is electrically neutral overall. "N" and "P" name the majority carrier, not a charge on the material. Heavier doping gives more carriers and therefore lower resistivity.
The PN junction
A PN junction is made by doping one part of a single crystal P-type and the neighbouring part N-type; two separate pieces pressed together would not work. The moment the junction forms, free electrons near it diffuse from the N side into the P side, and holes diffuse the other way, and they recombine. That leaves a thin layer on each side that has lost its free carriers but still holds its fixed ions — negative acceptor ions on the P side and positive donor ions on the N side. This layer is the depletion region, and it behaves like an insulator.
The fixed ions set up an electric field across the depletion region, pointing from the N side to the P side. The field opposes any further diffusion. The potential difference it represents is the barrier potential: about 0.6–0.7 V for silicon and 0.2–0.3 V for germanium at 25 °C. It cannot be measured with a voltmeter across an unconnected junction, but it sets the voltage needed to make the junction conduct. It falls by about 2 mV for every °C rise in temperature.
Forward bias
Connect the P side (anode) to the positive terminal and the N side (cathode) to the negative. The applied voltage pushes holes and electrons towards the junction, the depletion layer narrows, and once the applied voltage exceeds the barrier potential, carriers flood across. Current then rises very steeply for only a small further increase in voltage. A silicon junction carrying normal current drops roughly 0.6–0.8 V. Its own resistance is low, so an external series resistance (the load, or a current-limiting resistor) must set the current, or the junction overheats.
Reverse bias
Reverse the battery — P to negative, N to positive — and the majority carriers are pulled away from the junction. The depletion layer widens, the barrier rises, and almost no current flows. What does flow is the reverse saturation (leakage) current, carried by the few minority carriers that heat keeps producing. It is nanoamps for silicon and microamps for germanium. It hardly changes with reverse voltage, but it rises sharply with temperature, roughly doubling every 10 °C.
Because the widening depletion layer is an insulator between two conducting regions, a reverse-biased junction also behaves as a small capacitor. Its capacitance decreases as the reverse voltage increases, because the "plates" move apart. The varactor diode in Chapter 3 is built to use this effect.
Reverse breakdown
If the reverse voltage is raised far enough, the junction suddenly conducts heavily. This is breakdown, and it happens in one of two ways:
- Avalanche breakdown — in lightly doped junctions at higher voltages, minority carriers are accelerated so hard that they knock electrons out of bonds, and those free still more: a chain reaction.
- Zener breakdown — in heavily doped junctions the depletion layer is very thin, so even a few volts create a field strong enough to pull electrons straight out of their bonds. It dominates below about 5 V.
Breakdown itself does no damage, but the large current at a high voltage means high power. In an ordinary diode that power destroys the junction unless the circuit limits the current. Zener diodes are designed to work in breakdown on purpose (Chapter 3).
Temperature and semiconductor devices
- More heat → more electron-hole pairs → more leakage current and a lower barrier potential (about −2 mV/°C).
- Every semiconductor has a maximum junction temperature — roughly 150–200 °C for silicon and only 75–100 °C for germanium. That, together with germanium's higher leakage, is why silicon replaced germanium almost everywhere.
- In power devices, rising temperature raises the current, which raises the temperature further. Left unchecked this thermal runaway destroys the device; heat sinks, cooling air and correct circuit design prevent it.
Heat is the main enemy of semiconductors. That is why equipment racks have cooling fans and extract air, why a blocked avionics cooling filter or a failed fan raises a caution, and why power semiconductors are bolted to heat sinks with thermal compound. A cooler semiconductor lasts longer.