# Why do power diodes have a p+ n- n+ construction and why not p+ p- n+?

I've been learning about power diodes and how they differ from low power diodes with the addition of a lightly doped n-type layer.
This n-type layer improves the breakdown voltage rating of the device, and improves conduction in forward bias due to the high number of injected carriers from the heavily dopes regions.
Will a power diode work the same if this n- layer is replaced with a lightly doped p-type layer? If it does, why is an n- layer preferred? Or, if it doesn't, why?

## 1 Answer

Electron mobility is approximately twice that of hole mobility, so using the electrons as majority carriers means you get:

• For fixed size, twice the performance or...

• For fixed performance, half the size.

• +1 Used to be almost three times hole mobility in silicon (not germanium), when I was studying this in 1980. I remember the old figures of 1300 vs 500 for silicon and 3800 vs 1800 for germanium. But measurements may have been refined since the olden days, I suppose. (Room temp of $300\:\text{K}$.) – jonk Sep 18 '18 at 12:01
• @jonk Mobility is a function of dopant concentration. Your numbers are accurate for low dopant concentrations, but mobility drops substantially, and the ratio changes to 2:1 at the higher concentrations that would be used in a diode. – Matt Sep 18 '18 at 15:08
• @Matt Thanks. I remember that mobility was some power of T (temperature) and also depended upon the electric field intensity. But I hadn't recalled it depending on the dopant concentration. Certainly conductivity is, of course. But I guess I need to read up, again. Do you have a reference I might look over? – jonk Sep 18 '18 at 15:20
• @jonk The Bart book addresses mobility here ecee.colorado.edu/~bart/book/book/chapter2/ch2_7.htm or "The Physics of Semiconductor Devices" by Simon Sze is an excellent book. – Matt Sep 18 '18 at 15:22
• @Matt Thanks Matt. That helps a lot. The lattice phonon model is invoked there, too. I'm familiar with it, so that's also a nice segue. I also believe I see that electron mobility declines quite rapidly at high dopant levels and that the ratio can be even less than 2 at high enough levels (where mobility is rather low, overall.) Appreciated. – jonk Sep 18 '18 at 15:29