From the silicon's crystal structure to discuss how to make doped semiconductors and the mechanics.
To see how we can make silicon a useful electronic material, we
will have to go back to its crystal structure. Suppose somehow(and we will talk about how this is done later) we could substitute a few atoms of phosphorus for
some of the silicon atoms.
A silicon crystal "doped" with phosphorus If you sneak a look at the periodic table, you will see that
phosphorus is a group V element, as compared with silicon whichis a group IV element. What this means is the phosphorus atom
has
five outer or
valence electrons, instead of the four which silicon has. In a lattice
composed mainly of silicon, the extra electron associated withthe phosphorus atom has no "mating" electron with which it can
complete a shell, and so is left loosely dangling to thephosphorus atom, with relatively low binding energy. In fact,
with the addition of just a little thermal energy (from thenatural or latent heat of the crystal lattice) this electron can
break free and be left to wander around the silicon atom freely.In our "energy band" picture, we have something like what we see
in
. The phosphorus atoms are
represented by the added cups with P's on them. They are newallowed energy levels which are formed within the "band gap"
near the bottom of the first empty band. They are located closeenough to the empty (or "conduction") band, so that the
electrons which they contain are easily excited up into theconduction band. There, they are free to move about and
contribute to the electrical conductivity of the sample. Notealso, however, that since the electron has left the vicinity of
the phosphorus atom, there is now one more proton than there areelectrons at the atom, and hence it has a net positive charge of
1
. We have represented this by
putting a little "+" sign in each P-cup. Note that thispositive charge is fixed at the site of the phosphorous atom
called a
donor since it "donates" an electron up
into the conduction band, and is not free to move about in thecrystal.
Silicon doped with phosphorus How many phosphorus atoms do we need to significantly change the
resistance of our silicon? Suppose we wanted our 1 mm by 1 mmsquare sample to have a resistance of one ohm as opposed to more
than 60 MΩ. Turning the resistance equation around we get
And hence (If we continue to assume an electron mobility of
Now adding more than
phosphorus atoms per cubic centimeter might seem like
a lot of phosphorus, until you realize that there are almost
silicon atoms in a cubic centimeter and hence only one
in every 1.6 million silicon atoms has to be changed into aphosphorus one to reduce the resistance of the sample from
several 10s of MΩdown to only oneΩ. This is the real
power of semiconductors. You can make dramatic changes in theirelectrical properties by the addition of only minute amounts of
impurities. This process is called "
doping " the semiconductor.
It is also one of the great challenges of the semiconductormanufacturing industry, for it is necessary to maintain
fantastic levels of control of the impurities in the material inorder to predict and control their electrical properties.