Ricochet is the best place on the internet to discuss the issues of the day, either through commenting on posts or writing your own for our active and dynamic community in a fully moderated environment. In addition, the Ricochet Audio Network offers over 50 original podcasts with new episodes released every day.
We all love blasting things with ions, and most of us could spend all day shooting at wafers, but eventually someone is going to ask you to build something useful. What am I doing with all this mess of silicon anyhow? Here’s where we see the use of all that stuff. What do you suppose happens when you put a p-doped chunk of silicon next to an n-doped chunk of silicon?
Diodes to Kill For
Recall that n-doped silicon is called that because it’s got a negative charge carrier. Or don’t; it’s not like I’m paying you to. There’s phosphorus or some such providing one extra electron that nobody really wants. And p-doped silicon has a positive charge carrier; one of the atoms is an electron short of a full orbital (sounds like the sort of things chemists whisper behind each other’s back). These charge carriers (the electron and the hole where an electron ought to be) tend to diffuse around in the crystal.
Take a look at that cookie sheet again. On the left, we’ve got an antimony atom (double chocolate chunk if you’re keeping track) providing an extra electron. In the obviously-photoshopped section on the right, you’ve got boron that’s missing one. Supposing the extra electron diffuses into the missing one; what happens? Suddenly all the atoms have eight electrons and everyone’s happy. All along the junction between the crystals electrons are seeking out holes like frustrated golfers finally making birdie.
There’s a reason though that boron didn’t have that electron to bring to the party (lousy moocher). Boron comes with five protons in its nucleus, and when it’s neutrally charged it has three electrons in its outer shell. It doesn’t have the protons to a fourth electron. Relative to the electrons the protons are pretty well stuck in place (they diffuse like we covered before, but not nearly as quickly as the electrons or holes will.) As these holes and electrons merge you uncover more and more effect from the fixed charges. The p-doped side of the diode will start to generate a negative electric field, and the n-doped section will generate a positive field. The field opposes the motion of further electrons, and the whole process grinds to a stop. The metaphorical golfers get frustrated again.
What you’re left with is a crystal that has an insulating layer in the middle; all those electrons are perfectly happy right where they are thank you very much. Let’s fix them. Let’s apply a voltage across the thing. Hook up your negative electrode to the p-doped silicon. The negative electrode provides extra electrons to the part of the crystal that has extra holes. What happens? The electrons fall into the holes and it gets even more insulating. Huh; that didn’t work.
Switch the electrodes around. Stick the negative electron up next to the n-doped silicon. All the fresh electrons jostle the existing electrons (and the crystal’s natural electric field) like an influx of subway passengers, and the electrons move towards the back of the car, or the positive side of the crystal.
Conversely, on the p side of the crystal, there’s a demand for electrons at the positive electrode. It starts shunting electrons that way, which narrows the depletion zone (that insulating layer I mentioned earlier.) If the voltage is high enough there’s no profit in shunting electrons into locally produced holes, it’s all in the export market. The depletion zone goes away and the diode starts conducting electricity.
What we have then is a chunk of silicon that will conduct electricity in one direction but not in the other direction. (You can get it to run backward if you apply enough voltage, but it isn’t much good for anything afterward. Learned that one entirely by accident in the physics lab). There are plenty of interesting things you can do with just that, but wait; there’s more!
From Diodes to Transistors
What happens if you stick three crystals together? Either P-N-P or N-P-N; NNP or PPN just gives you a longer diode. Let’s talk about an NPN transistor; which I’ll represent with this Neapolitan ice cream sandwich
The chocolate and strawberry sections are doped with boron, making them n-type crystals. The vanilla section is doped with antimony, making it a p-type crystal. Okay, you stick your electrode in either end of the sandwich, you won’t get a current going across. You can think of it as two diodes facing opposite directions; either way you choose you chose wrong.
In the example photo, we’ve got the negative probe stuck into the chocolate and the positive into the strawberry. The chocolate-vanilla diode is perfectly willing to let electrons pass, but the vanilla-strawberry diode won’t. Okay, it’s a multimeter and not actually something that provides power; I’m looking to make an example not fry my ice cream. Quibbles aside, hook up a positive electrode from another source to the vanilla (p-doped) silicon.
Electrons can flow from the chocolate to the vanilla to the third electrode. However, because you’ve now got electrons moving in that vanilla section they can also move across the depletion zone (that area where all the holes are filled) and into the strawberry section. If you take away the voltage from the third electrode acting on the middle section then the electrons no longer have an excuse to get into the vanilla and from there mosey into the strawberry. The current across the whole thing switches off.
So What’s it to Me?
Practically what this gets you, you’ve got a switch that you can turn on and off with electricity. It doesn’t seem like much but it opens a world of possibilities. Bell Labs developed it to amplify a phone signal (a weak signal on the middle part of the transistor from New York City will toggle a strong local current in Philadelphia, and can be relayed on to points beyond. Coast-to-coast calling.)
But the real interesting thing you can do with this is logic. Line up a whole mass of transistors and set them to turning themselves on and off. You can do math with that. With clever math, you can produce absolutely everything else a computer is capable of. Join us next week when we do some simple transistoring in “NPN Junction What’s Your Function?” or “They Said There Would Be No Math.”
This is part four of my ongoing series on building a computer, the Russian Novel way. You may find previous parts here: 1 (silicon) 2 (crystallography) 3 (doping) or all of them under the tag How to Build a Computer. This week’s post has been brought to you by Fyodor Dostoevsky. If there’s one man who knows excellent extremely long form writing it’s Fyodor Dostoevsky.