Sunday, October 12, 2008

Doping of Semiconductors: n-type and p-type

Doping of semiconductors is achieved by introducing atoms with more or less electrons than the parent element. Doping is substitutional, the dopant atoms directly replace the original atoms. Suprisingly low levels of dopant are required, only 1 atom in 109 of the parent atoms.

Looking at silicon; if phosphorous atoms are introduced into a silicon crystal then extra electrons will be available (one for each dopant atom introduced as P has one extra valence electron). The dopant atoms form a set of energy levels that lie in the band gap between the valence and conduction bands, but close to the conduction band. The electrons in these levels cannot move directly as there is not enough of them to form a continuous band. However, the levels themselves can act as donor levels because the electrons have enough thermal energy to get up into the conduction band where they can move freely.

Such semiconductors are known as n-type semiconductors, representing the negative charge carriers or electrons.

What if, instead of doping with phosphorous, we doped silicon with an element with one less valence electron such as gallium. Now for every dopant atom there is an electron missing, and the atoms form a narrow, empty band consisting of acceptor levels which lie just above the valence band. Electrons from the valence band may have enough thermal energy to be promoted into the acceptor levels, which are discrete levels if the concentration of gallium atoms is small. Therefore, electrons in the acceptor levels cannot contribute to the conductivity of the material. However, the positive holes in the valence band left behind by the promoted electrons are able to move.

These type of semiconductors are known as a p-type semiconductors, representing the positive holes.

There are two fundamental differences between extrinsic and intrinsic semiconductors:

1) At standard temperatures extrinsic semiconductors tend to have significantly greater conductivities than comparable intrinsic ones.

2) The conductivity of an extrinsic semiconductor can easily and accurately be controlled simply by controlling the amount of dopant which is introduced. Therefore materials can be manufactured to exact specifications of conductivity.


Controlled valancy semiconductors:

Some transition metal compounds can be conductors due to the presence of an element in more than one oxidation state. NiO is a very good example. On oxidation the compound goes black and becomes a relatively good conductor. Some of the Ni2+ ions have been oxidised to Ni3+ and some Ni2+ ions diffuse out to maintain charge balance leaving cation holes.

The reason for the conduction is the ability of electrons to transfer from Ni2+ to Ni3+ ions. This basically allows the Ni3+ ions to move and black NiO is therefore a p-type semiconductor. Slightly different to the p-type discussed earlier this type is known as a hopping semiconductor because the transfer process is thermally controlled and therefore highly dependent on temperature.

This makes controlling the conductivity a tricky process. Therefore controlled valancy semiconductors rely on control of the concentration of, in this case, Ni3+ ions by controlled addition of a dopant (such as lithium). Instead of having NiO, you now have Li+xNi2+1-2xNi3+xO, hence, the concentration of Li+ ions controls the conductivity.
The P-N junction:

This occurs in the situation where a crystal has been doped such that half of it is n-type and the other half is p-type. The two halves have different Fermi levels (n-type’s is higher) and electrons flow from the n-type section to the p-type section to try and equalize the electron concentrations (fig. 9). This creates a positive charge on the n-type region and a negative charge on the p-type region which leads to an electric field pushing electrons back to the n-type region. Eventually a balance is reached (fig. 10).

If you apply an external potential difference to make the p-type region positive and the n-type region negative a continuous current can now flow. Electrons enter from the n-type electrode, travel through the conduction band of the n-type region, drop into the valence band of the p-type region, continue through the positive holes and then leave at the other electrode. Cannot flow the other way with a relatively low voltage as the electrons are unable to jump up to the n-type conduction band.


The Mott-Hubbard gap and breakdown of the band model

The simple band approach says that a compound with partially filled d orbitals should form a metallic solid. However, this is not the case; a large number of halides, oxides and compounds with less electronegative ligands form non-metallic solids, which have magnetic and spectroscopic properties associated with partially filled levels.

A good example is NiO. Pure NiO is green and shows d-d transitions associated with octahedral ligand-field splitting of the 3d orbitals, as is the case with an isolated ion, eg Ni(H2O)62+ (fig. 11)

The magnetic properties reveal two unpaired electrons on each Ni2+ ion. There is a strong interaction between neighbouring ions which leads to antiferromagnetic ordering of the spins at lower temperatures. In compounds of this type the d orbitals appear localized as apposed to forming a conduction band like in metallic compounds. This non-metallic behaviour stems from strong electron repulsion between electrons in d orbitals. The band model relies on a simple approximation to deal with electron repulsion which isn’t good enough for many transition metal compounds.


http://www.chem.ox.ac.uk/vrchemistry/solid/Page12.htm