The internal functioning of a transistor

This text is a short explanation about the internal functioning of a bipolar NPN transistor.



1. How a transistor works
2. Some details about the manufacture





1. How a transistor works

Look at this schematic drawing of the inside of a transistor. The whole text will be about that drawing:



Look at following details:

Here is how a NPN transistor works:

  1. An electric tension of say 9 Volt is put between collector and emitter. Positive on the collector (positive wire of the battery is put on collector, negative wire on emitter). The NP diode collector-base prevents the electrons from crossing. It is blocking. Because of that voltage of 9 Volt the no-man's-land between collector and base increases even more: electrons and holes both redraw further apart. There is now a voltage of 9.7 Volt between collector and base, inside the no-man's-land.

  2. A voltage of 0.7 V is put between base and emitter. Positive on the base (positive wire of the battery on the base, negative wire on the emitter). That voltage compensates the voltage difference of 0.7 V that was there between base and emitter. The base-emitter no-man's-land disappears because holes and electrons are pushed towards each other. The holes entering the emitter do quickly recombinate with the numerous electrons of the emitter and disapear. On the other hand the electrons coming from the emitter spread into the base like a drop of color inside a glass of water. They do not encounter much holes.

    One could think a voltage of 0.01 V would be sufficient for the fastest electrons. That's right, but such fast electrons are very scarce. It is only from 0.5 V on that a minimum of flow is attained. Only some transistors or diodes do show a clear current for base-emitter voltages lower than 0.5 V, down to 0 V.

  3. Once into the base, amongst the holes, the electrons swirm around. Unavoidably some of them arrive into the base-collector no-man's-land. They are then propulsated by the 9 V field inside that no-man's-land towards the collector. They enter the collector and merge with the electrons of the collector.

    During the crossing of the no-man's-land, undergoing the electric field of 9 V, the electrons receive each an energy of 9 eV they will evacuate under the form of heat. Inside a transistor, the collector is thus the part that heats up most.

  4. The collector having robbed away electrons that where moving inside the base, the base becomes a little bit more positive and thus attracts new electrons from the emitter towards the base. And so on: indefinitely electrons will be pumped by the collector and will be replaced by electrons from the emitter.

  5. Sometimes an electron inside the base does encounter a hole. It will recombinate and "disapear". To compensate, the battery between the base and emitter will have to furnish one electron current to make one replacement electron come from the emitter and one replacement hole be produced where the wire touches the base.

The number of electrons that make the travel towards the collector before one recombinates inside the base is the amplification coefficient of the transistor. In other words: the base-emitter battery will have to furnish 1 electron while 20 electrons travel from emitter to collector, those 20 electrons being furnished by the collector-emitter battery.

The purpose of the emitter-base current is thus to start the emitter-collector current flow, then to compensate the loss of electrons inside the base. Should the emitter-base battery not compensate those losses, then the emitter-collector current will decrease each time an electron recombinates/disapears. Down to zero. This makes it can take a while before the collector current stops whenever the base battery is removed. That phenomenon can easily be seen on an oscilloscope. If you want the collector current stop immediately, then a negative voltage must be put on the base, in order to repell the electrons out of the base.

Why is the collector thick? Because that way when an important voltage is put between collector and emitter, say 50 V, the electrons of the collector will be able to redraw on a long distance, creating a wide empty zone that will not be crossed by an electric spark that would ruin the functioning of the transistor or even destroy it.



2. Some details about the manufacture

The metal wire-semiconductor contacts are made by pushing metal pieces against the semiconductor, then heating the whole up to high temperature. Metal and semiconductor melt a little bit together. Once cooled down, the resulting glueing mixture is very coarse but highly reliable.

The contacts between the n+, p and n semiconducting zones have not to be made, because they've always existed. One could think the N, P and N zones of a transistor are separate blocks that are pushed against each other. That's not the case: a transistor is a sole monocrystalic block of silicium wherein different zones are created by putting different sorts of ions. Somehow like a Belgian flag made out of one sole piece of cloth, one third being painted in black, the second third in yellow and the third third in red. The color is very important, it is what makes the flag be a flag, yet physically it is only a few miligrams of pigments on a continuous single rectangle of tissue. There is one piece of tissue, not three. (One reason therefore is that if three pieces of crystal where pushed against each other, the two contact zones between them would not be a continuous crystal. Those ruptures into the crystalic continuity would create a lot of unwanted holes and electrons that would disturb everything.)

The reason why the emitter is a n+ semiconductor is the following one:

If an aluminium wire is pushed directly against the p semiconductor of the base, and the whole is heat up, aluminium atoms will penetrate into that p semiconductor and create a very strong local positive doping, p+. Should one try to inject electrons inside the p semiconductor trough the aluminium contact, that will fail: because the electrons must first pass trough that heavily p+ doped zone, full of holes, where they will unavoidably be trapped by recombination with those holes. (Instead of injecting electrons into the base, it will be the base that will deliver holes to the heavily doped contact zone.)

The aluminium / p zone is thus an unavoidable electron trap. The current will go trough without any problem, but it will not be made out of electrons: it will be made out of holes travelling in the opposite direction.

Therefore one solders the aluminium wire on a n+ semiconductor located just near the p semiconductor. That n+ semiconductor will have no problems to allow electrons go into the p zone.

Yet there remains a problem: if when one solders an aluminium contact on a n semiconductor it makes that semiconductor become locally p-doped very heavily, how can we avoid that it erases the local n doping and creates a parasitic diode that will disallow any electron movement? Answer: before approaching the aluminium contact, one dopes heavily, negatively, the surface of the semiconductor, in order it ceases to be semiconducting. It is so full of n ions that it becomes simply a conducting material, just like metal. Thereafter there can no more be erasing by p aluminium ions, because the surface is now a conductor and no more a semiconductor. The added p ions simply make the zone even more conducting.

The electrons passing from emitter wire towards the base will thus cross successively:

  1. aluminium

  2. conducting silicium full of p and n ions

  3. conducting silicium full of n ions

  4. semiconducting n+ silicium

  5. semiconducting p silicium (the base)




    Eric Brasseur  -  1994       [ Homepage ]