Inside a diode

You can skip this page and still wire a diode correctly for the rest of your life. It is here because the one-way behaviour looks like magic until you have seen where it comes from, and because the same junction turns up again inside every transistor, every LED and every chip on the board.

Two kinds of doped silicon

Pure silicon is a poor conductor: every electron it has is busy holding an atom to its neighbours, and there is almost nothing free to carry a current. Mixing in a trace of another element changes that, and there are two useful ways to do it.

• Add a pinch of phosphorus and every phosphorus atom brings a spare electron that no bond needs. That is n-type — n for negative, the charge of the thing that is free to move.
• Add a pinch of boron instead and every boron atom leaves a bond one electron short. The gap is called a hole, and it moves: a neighbouring electron steps into it, which leaves a gap where that electron was. That is p-type.

Both are still electrically neutral. A p-type crystal is not positively charged; it just has gaps where electrons can move about.

What happens when you join them

Press the button and watch the seam.

The last frame is the whole diode. The carriers that crossed left behind charged atoms that cannot move, negative on the p side and positive on the n side, and those fixed charges make a voltage that opposes any further crossing. About 0.7 V of it in silicon. The gap they sit in has nothing free left in it at all, which is why it is called the depletion region.

At this point the junction has stalled itself. Nothing else happens until you put a battery across it.

Which is what this does

Drag it forwards past about 0.7 V and the battery cancels the built-in voltage: the depletion region collapses, carriers pour across, and current flows. That is forward bias.

Drag it backwards and the battery pulls the carriers away from the join instead. The depletion region gets wider, the barrier gets taller, and the harder you push the more thoroughly it is shut. That is reverse bias, and it is why a diode does not just block backwards — it blocks backwards better the harder you try.

Three details that will come up later

Leakage is not zero. A few carriers always have enough thermal energy to get across. That is the nanoamps or microamps you see on the reverse side of the figure, and it roughly doubles for every 10 °C the part warms up. It only matters when you are counting microamps on a battery, and then it matters a lot.

There is a reverse voltage it cannot survive. Push far enough backwards and the field inside the depletion region gets strong enough to rip carriers out of their bonds. Current goes from nothing to everything, and unless something else limits it, the part is destroyed. Datasheets call the limit the peak inverse voltage, or PIV.

Not every diode is made this way. A Schottky diode is silicon against metal rather than p against n. There is a barrier and it is one-way, but it is a lower one — around 0.3 V instead of 0.7 — and there is no stored charge to clear out when it switches off, which makes it faster. It pays for that with more leakage and a much lower reverse rating, and both of those trade-offs get a page in power and protection.